add all 2025 transcripts

Agents in the database (Interview)_transcript.txt

@@ -0,0 +1,441 @@
+**Adam Stacoviak:** So friends, we're here with Ajay Kulkarni, a new friend of mine from Tiger Data, previously Timescale. We've had this relationship, we work with you as a sponsor, and I've been a fan, obviously, of time-series data, and I had ideas for you all, I've been working with Isabel behind the scenes... And then it finally came back to this moment here, where you have agentic Postgres, which is just super-interesting to me... So I thought we would dive deep into who you are, Ajay, what your journey might be, and how you've come to love building databases. So let's start there.
+
+**Ajay Kulkarni:** Yeah, yeah. I'll try and give you the short version --
+
+**Adam Stacoviak:** I assume you love building databases.
+
+**Ajay Kulkarni:** You know, I love building.
+
+**Adam Stacoviak:** Yeah, okay.
+
+**Ajay Kulkarni:** I love building, and I love solving problems. Now, databases are interesting, but it would not be the only thing I've done in my career. I've been in love with technology since I was a kid, and I remember using the internet for the first time in high school, in 1995, 1996, and thinking "I don't know what this is, but this is fun."
+
+**Adam Stacoviak:** Yeah.
+
+**Ajay Kulkarni:** And I was a pre-med at that point, applying to colleges, and I switched to computer science, went to MIT... And I guess the rest, yeah, is history. MIT is where I met my co-founder, by the way... So that's how we know each other for 28 years.
+
+**Adam Stacoviak:** That's wild, to know a co-founder of 28 years. I mean, the history, and the level of trust and maybe somewhat antitrust... I don't know if that's the case, but --
+
+**Ajay Kulkarni:** It's not antitrust, but you know what it is? It's --
+
+**Adam Stacoviak:** Vulnerability is what I mean by that.
+
+**Ajay Kulkarni:** When he and I started working together at this point, probably 10 years ago, I remember thinking "I know what I like about Mike, and I know what I don't like about him. And so I want to work with this guy", you know?
+
+**Adam Stacoviak:** That's interesting.
+
+**Ajay Kulkarni:** And I'm sure he felt similarly. I think it's a little bit of like a -- I'm not sure I want to call it a spousal relationship, but almost, you know? You're like "Hey, I like you for who you are, the whole package, even though sometimes that package annoys me... And I'm sure you'd say the same thing about me."
+
+**Adam Stacoviak:** Well, Mike is not here, so we can't speak for him, but I'm assuming he might say something like that...
+
+**Ajay Kulkarni:** \[laughs\]
+
+**Adam Stacoviak:** So CEO of Tiger Data... You were under a different name, a different moniker before... I do want to go there, but I kind of want to zoom back out a second, just to kind of identify who you are, but more so how you got here. You mentioned MIT, a long road, 28 years, knowing Mike, met him at college, university... What was the journey from there to go into, I suppose, your career, and what are some of the things you've done that you really feel have defined or identified who you are today?
+
+**Ajay Kulkarni:** Yeah, I think I've tried to always follow the thing that kind of tugs my heartstrings... To be like "Hey, I don't know why, but this is interesting", and really to listen to that. That's why I ended up at MIT... My high school guidance counselor tried to dissuade me from going, and I was like "Nah, I wanna go. I'm gonna go, you know?"
+
+**Adam Stacoviak:** "I'm gonna do this."
+
+**Ajay Kulkarni:** And it was hard, but it was great. I graduated during the dotcom collapse. So at that point tech was not hiring, which just sounds crazy at this point, you know... And I ended up on Wall Street as a bond analyst. But very quickly, after a couple of years, I ended up back in Palo Alto, working for a startup, and I've been working at startups since.
+
+**Adam Stacoviak:** What year was that, roughly? That back to Palo Alto.
+
+**Ajay Kulkarni:** Oh, back to Palo Alto? 2004.
+
+**Adam Stacoviak:** Okay. I know exactly what I was doing that year. I was so wayward then... That was the beginning of my developer career, I would say. Not that this is my story, but just so you kind of understand the cloth there, the timing... That was the same year my daughter was born. She is now 21, so that's 21 years ago for you, almost 22 as of this January, if that was the case for you too...
+
+**Ajay Kulkarni:** \[08:05\] Congrats, that's incredible.
+
+**Adam Stacoviak:** Thank you. I was in Canada at the time. Oddly enough, I'm an American, I'm United States... Served our military... I don't say that as like a nationalist necessarily \[unintelligible 00:08:17.14\]
+
+**Ajay Kulkarni:** Where did you serve?
+
+**Adam Stacoviak:** In the Army. But just to say that I'm not Canadian. Not that I'm against Canadians... I'm just not Canadian, so I want to be clear about that. Some people think because I lived in Canada for a bit, they're like "Are you Canadian?" I'm like "No, no. I don't say aboot."
+
+**Ajay Kulkarni:** Probably because you're so friendly. People are like "Oh, you must be Canadian..."
+
+**Adam Stacoviak:** Well, yeah, maybe that's it. But I was fresh... So I didn't go to college, or to MIT, or to a computer science school, or get a degree in that... I learned by messing with GeoCities that same year. And that's so funny to even think about... I started playing with WordPress, got a job, and almost got fired from a company called IT Weapons. Still around, in Canada, itweapons.com. One of the most formative moments in my career - I got hired into a sales role I could not do well, because their sales style was so different from what I was used to. And I was terrible at that job. And I knew I was so close to getting fired... And I took it upon myself to be like "Nah, man... You're not getting rid of me." Almost like you were with your guidance counselor, "I'm gonna do this." And so I defined the business development role for the company, and started to add value beyond what they thought that I could add... And they never really told me directly that I was minutes away from getting fired, but I knew it, I knew it in my soul... But I also saw the turnaround, and the benefit there, and it was pretty awesome. So anyways...
+
+**Ajay Kulkarni:** I know -- I like that story, because I feel like that is something you don't know you have until it's tested.
+
+**Adam Stacoviak:** Yeah.
+
+**Ajay Kulkarni:** Which is the, "Hey--"
+
+**Adam Stacoviak:** I didn't even know I had it, Ajay. I had a wife and a brand new daughter coming in those moments, so that timeframe for me... I could not not survive. I had to find a way.
+
+**Ajay Kulkarni:** I think it's some combination of grit, but also just like an FU mentality... \[laughs\] It's like "Oh, no, no, no --"
+
+**Adam Stacoviak:** Yeah, I was like "Listen, I belong here. I love this company... I'm gonna do something about it, I'm gonna figure it out." So I've always been good at connecting the value dots, I would say. You can call it sales if you want to, and I think that's the easy, in quotes, word you can use... But I really think I have this uncanny ability to -- and I've just grown into it in my career, and I've leaned into it... Where I've been able to connect the dots between things that just aren't normally connected, and express and redefine and help shape that value in those connected dots.
+
+**Ajay Kulkarni:** That's great. I like that story.
+
+**Adam Stacoviak:** Let's go back to you, though... This is not about me, but I thought we'd at least encapsulate what was happening in 2004 between our lives... So - much different perspectives. But similar, I would say we're -- I'm not a co-founder or a CEO of a tech startup, but certainly a tech brand... Changelog.com has been around, we're an institution at this point; 18 years, Ajay. We started in 2009.
+
+**Ajay Kulkarni:** I mean, in the tech industry, that's like a hundred years or so.
+
+**Adam Stacoviak:** That is. We're a Wikipedia basically, but not really. I'm just kidding. Wikipedia is the eighth, ninth, tenth \[unintelligible 00:11:29.20\] one of those two things, or three things...
+
+**Ajay Kulkarni:** Something like that, yeah.
+
+**Adam Stacoviak:** So...
+
+**Ajay Kulkarni:** I'll tell you what's been the theme in my career, is that - when I was on Wall Street, it was a really interesting, an eye-opening experience... Because I was around some really smart people. I was a quant, around people who all had PhDs, some had multiple PhDs, really hardworking... It was a meritocracy. It wasn't like a lot of, you know, FaceTime. I mean, you would work, but it was... And I was paid really well, but I didn't like it. I didn't like it. And I was like "You know what? I feel like this is not who I am." And I felt like I was someone who needed to build, and needed to create. And I actually -- I remember going soul-searching, like "Hey, should I go into consumer packaged goods? Should I go into--"
+
+\[12:16\] At one point I thought about starting a record label, because I'm like "I like music, and I like business, so maybe I'll start a record label." I remember buying a book about it, and doing research... And at some point I started a chocolate company... And I kind of realized - and it's kind of obvious in hindsight, that "Dude, if you want to build something to make an impact, tech is the best way."
+
+**Adam Stacoviak:** Yeah.
+
+**Ajay Kulkarni:** I mean, even more so now, but even back then... And so I think that's what kind of brought me to tech, and I think that's what keeps fueling me, is like... Yeah, just the ability to have a positive impact in the world, and on other people. And it's not always fun, it's not always easy, but I think that's probably the main thing that's been driving me, at least since that insight.
+
+**Adam Stacoviak:** Hmm. Can you expose some scars, some hard years, some bad choices that you can not reflect on, that are formative to who you are today? Kind of like what I did with you a little bit.
+
+**Ajay Kulkarni:** Yeah... So you mean like life choices?
+
+**Adam Stacoviak:** Yeah, where are the forks in the road that you're like "Man, that's the moment where I learned this hard lesson", or that's the moment -- I know you mentioned Mike, and 28 years, so maybe that's one of those forks, but... Just those moments where you look back now -- and it's kind of funny, because in the moment sometimes we can't see the forest for the trees, right? And we think like this choice or this thing happening is the end of the world, or is the biggest choice ever. And it may be in that moment, but when you get past it, you sort of get past it with hindsight... And that's why it is truly 20/20, because you look back with so much clarity. What were some of those moments for you when it was unclear, a choice you had to make, but now it's like, it's clearly defined who you are today.
+
+**Ajay Kulkarni:** As a mistake or a good choice?
+
+**Adam Stacoviak:** You pick. Whichever direction you wanna go.
+
+**Ajay Kulkarni:** I feel like mistakes are more interesting.
+
+**Adam Stacoviak:** Scars, bloody knuckles... You know, sometimes mistakes can be either/or, sometimes the positive choice we make can be either/or, a scar or a bloody knuckle.
+
+**Ajay Kulkarni:** I mean, look, I think you probably know this - entrepreneurship is a series of mistakes. And the key is just like not letting the company die, despite your mistakes. There's many I've made over the years: bad hires, bad strategic decisions... I'll tell you one actually that started even before then... I used to work at this company - and I won't mention \[unintelligible 00:14:42.17\] because I'm not trying to throw anyone under the bus... But I used to work at this company where I was in charge of an engineering team, and somehow I thought I was on track to get more responsibility, and somehow there was this internal uproar, and I realized people didn't like me. And it was an uproar, and I ended up getting sidelined. And that hurt for a couple of reasons. Number one, I'm kind of a sensitive guy... Yeah, I'm an entrepreneur, but I'm like "Look, dude, if you genuinely don't like me", I'd be like "Why?" I'm curious, because I'm like "Did I do anything to offend you?" But also, I was just caught off guard. And what I realized is that part of that was my fault. Part of that was I think I came across more dictator-like than I intended, but part was, I would say, maybe indirectly my fault, where there were some decisions that were made where I took the heat for it, where it wasn't actually a decision that I made.
+
+\[15:58\] I guess, I don't know, maybe -- I guess, long story, I think I changed my style, but I also realized that you have to look out for yourself, and your manager, whoever, that they will not necessarily look out for you. But also, conversely, I've tried to be the type of manager who looks out for people. I don't know if that's an interesting story, Adam, but like... \[laughs\]
+
+**Adam Stacoviak:** Well, so let's zoom into maybe how are you changed as a result? What did you do to -- where I'm trying to go to is how did you get here? ...in terms of not just where you're at in this moment, but like the person you are leading the company you are. What are some of the things that were in your past, or in your choice line, that has now helped Ajay today be Ajay to lead?
+
+**Ajay Kulkarni:** Yeah, I would say that I've made -- I'm not sure you want to call them mistakes, or just, I've had learning experiences over the years... And I think that I've formed who I am today. I think the key theme - and this is not all of them, but I think one key theme is I have learned how to trust myself. I remember early on I would make some hires, for example, in sales, that didn't gut feel right to me... But I was new to enterprise sales, I was new to the database business. And in hindsight, my gut was right. My gut was right that "Hey, the general --" I like to say "In a fast-moving industry, expertise is a liability." So in the database world, as the databases were moving from enterprise sales and on-prem to like cloud and PLG, the people who could sell this would be totally the wrong people for this. And I kind of felt that, but I couldn't articulate it... And I made a few hires that I had to let go, and that's my fault. That's on me. But I think that's one thing that taught me that "Hey, you got this far for a reason. Listen to your gut. And if you can't articulate it, then try to take the time to articulate it."
+
+I think more recently... Look, I have a conversational style that myself -- both Mike and I, we talk about this... It's called cooperative overlapping. It essentially means that we interrupt a lot. But we interrupt because that's how we talk. Like, if you interrupt me, I feed on that. And one thing we realize is that -- well, at one point there was some people on my executive team who were like "We need to interrupt less." And I remember thinking -- my first thought was "You're right, we should interrupt less." And my second thought was "No, this is who we are..." \[laughs\] Like, we got this far for who we are, you know? And at a company either you ask the founders to change, or you ask the whole company to change, you know?
+
+**Adam Stacoviak:** Yeah.
+
+**Ajay Kulkarni:** And what I realized is that there are people in the company who liked who we are. They liked that we interrupted, you know? And so I got to the point, I was like "Yeah, you know what--" You know, we talk about California mentality versus New York mentality... New York mentality is "Hey, I'll be kind to you, but I may not be nice. I'll be gruff on the street, but I'll help you." And that's just the New York way. And if you're in New York, you have to realize that, that when people are too busy to talk to you, they're not being rude; they're actually trying to be respectful of your time. So yeah, I think -- I don't know, I feel like over the years not trusting my own instinct has... I mean, you live with the consequences. And so I think I've learned that "Hey, just trust your instinct, because whether you win or lose, it's on you."
+
+**Adam Stacoviak:** \[19:42\] Yeah. One thing I wrote down a while back... This is actually 2012. September, 2012. And it was the question of "Where does a company's culture come from?" And essentially, it's boiled down to what you just said there, which is a founder's values and principles define the company's culture. And so these things, they kind of -- they're top down in a way, but if you don't have, I guess, fortitude in who you are and why you are the way you are to push back when somebody says "Hey, maybe that's not cool", and you're like "I can kind of see that, but at the same time, it's kind of who I am. And here's why I'm delivering my message this way, or why the way I speak this way." And it's less about interrupting, but more about that cooperative -- what did you call it, cooperative overlap? Is that right?
+
+**Ajay Kulkarni:** Cooperative overlapping. I think that's the technical term.
+
+**Adam Stacoviak:** Yeah. And so it does put a name to it to tame it. So it's not interruption, it's meant to be a courteous interrupt to probably provoke deeper conversation, or to provoke more collaboration and involvement, right? Isn't that probably the reason for --
+
+**Ajay Kulkarni:** No, totally, but I think it's also a function of like "Don't try to shore up your weaknesses, try to lean on your strengths." So I am an intense person. When I was younger, I would feel bad about it, because some people were like "Oh... Well, you're intense." But now I'm like "No, it's just who I am." And some people love that about me. And there's some things I can do really well because I am intense, you know? It's who I am. If you don't like that, cool. We don't have to work together.
+
+Similarly, I talk a little fast. I remember at one point I was giving a presentation, and some of my peers were like "Oh, why don't you slow down? You're talking really fast." And I slowed down, and they were like "No, this is worse. It was better when you talked fast, because your energy was there", you know? And I was like "You know what? Yeah, I talk fast, and maybe I might muddle some words sometimes, but that is who I am. That's how my energy comes through." And just lean on that. I often talk -- I talked about it with Mike, it's like, we both have weaknesses, we both have flaws... And in the past, I would have been like "Oh, we have to work on these." And now I've said, "You know what? Don't try to teach a fish to climb a tree, you know?" If you're a fish, swim better, you know? Don't be like "Oh, I can climb a tree."
+
+**Adam Stacoviak:** Just keep swimming, right?
+
+**Ajay Kulkarni:** Just keep swimming, yeah. And so, it's --
+
+**Adam Stacoviak:** Dory's got the best advice.
+
+**Ajay Kulkarni:** Dory... Yeah, that's right. Keep swimming. But it's just like, figure out who you are... Yeah, and lean into it. I mean, people talk about this in athletics, how once upon a time, maybe 50-60 years ago, the idea of the ideal athlete was someone who was just balanced in every possible way. Not too tall, not too short, not too strong, not too weak... And now they realize that - no, you actually want the genetic freaks, you know? That's where you get the alpha, you know? You're someone who's like -- if your wingspan is like...
+
+**Adam Stacoviak:** The Michael Phelps with like massive wingspan. Gosh...
+
+**Ajay Kulkarni:** Yeah, dude. Totally. Michael Phelps is like - what, 6'4", 6'5", I think?
+
+**Adam Stacoviak:** He's huge, yeah.
+
+**Ajay Kulkarni:** And there's this marathon runner who I think is like 5'7". And I remember reading an article that talked about how they wear the same pants. Because the marathon runner is all legs. And Michael Phelps is all torso.
+
+**Adam Stacoviak:** Yeah.
+
+**Ajay Kulkarni:** And you're like "Yeah--", you know? You don't get good at a sport by being balanced in everything. You find the thing where you have the edge, and you lean into that, you know?
+
+**Adam Stacoviak:** Yeah. One thing you mentioned was this -- it seems like you've got some history in enterprise sales, enterprise database, you mentioned PLG, product-led growth... Help me understand some of your backstory when it comes to I guess just databases in general. How did you get into that world? What was that world for you, and how has it changed to now?
+
+**Ajay Kulkarni:** \[23:52\] Yeah, I mean, I've been using databases since the late '90s, I guess, ever since I started getting into computer science. I started using probably MySQL in 2004, Postgres in 2011, I think, thanks to Heroku... So I've always been involved in databases.
+
+I think what's interesting to me is when we started this company, which was then Timescale, we entered a world where the success stories were companies like Hortonworks and Cloudera. Really big on-prem. And that was not my background at all. I'd never been a salesperson, never worked with enterprise salespeople... Now I know what these terms mean, but back then capacity planning, quota coverage, territories - they meant nothing to me. And we slowly had to learn it. But I also had this inkling that "Hey, you know what? I feel like the database industry is changing. It's changing from that old model to something that looks more like SaaS." And SaaS is less about enterprise sales, more about building a great product. And I knew how to do that. And so I think that was part of the journey, it was starting off as a user, getting into an industry not really realizing how the business worked... But then, again, I think following my instinct to be like "Hey, I think this industry is shifting."
+
+One thing I like to say about myself and my co-founders is that we're very good students of the game. And so we were new to open source licensing, and then quickly became open source licensing experts. I would say we were new to PLG and sales, but I think now we're -- I'm not sure if we're experts, but I think we're probably top quartile or whatever for that.
+
+**Adam Stacoviak:** You know, I think about the way databases have changed, and the way the sales of them have changed. Largely, it's as if -- if you're not an open source database, you're not worth your weight in gold, because it's a black box; things can change. Even the term - and this may cut deep to you, potentially - is Postgres-compatible. I think Timescale has never been Postgres-compatible. It's always been Postgres-native...
+
+**Ajay Kulkarni:** That's right.
+
+**Adam Stacoviak:** ...but some out there choose that Postgres-compatible, and that's because they wanna do business differently, they wanna license differently... And I think the PLG model has obviously -- one, I feel as a developer if I can't go and play with your tool, even not so much in a free capacity, but the ability to explore it and learn it, and then trust it... That's the way. That's the way of the developer. You've got a side project, you've got an itch you wanna scratch, you wanna try this different thing... And if you're not going that route, it's kinda hard to really instantiate change in your organization. You may go from one database to another, or you have an idea, and if you can't go and explore it and carve it out for yourself and present it to your team, that's the way of PLG; that's the way of open source.
+
+**Ajay Kulkarni:** Yeah, no, it's fascinating, because I think databases, like software, used to be something where the key decision-maker was the CIO or CTO, and that deal was done in a steakhouse or on a golf course. That's how it used to be. And with the shift to SaaS, and then to shift to cloud for databases, that decision moved to "No, it's a developer sitting at their computer, just making the choice based on some combination of what they read, what their peers told them, and what their own visceral experience was." And I think that's been fascinating for me, because - I mean, I think it's totally changed. I mean, the core job that a database does has not changed, but I think the way you build the business has totally changed from being more sales-led to being more product-led. I think that's been interesting.
+
+**Adam Stacoviak:** Yeah. What was the original challenge you faced to even consider creating Timescale? What was that moment, and when was that moment?
+
+**Ajay Kulkarni:** \[28:06\] We started off as a company building an IoT platform, internet of things. And it's like 2015. Yeah, I mean, back then I'd just spent 10 years in mobile, and I remember thinking "Wow--" Mobile was really exciting, but around 2014 it started to get a little bit boring. Pre and post iPhone it was exciting, but then in 2014 you're like "Okay, I have enough apps." I'm like "What's the next thing?", and IoT felt like the next wave of computing. And so we started off building what we thought the market needed, which was a data platform for IoT devices. And that idea was moderately successful. We tracked over 100,000 devices, we raised the seed round, built a small team... But we needed a database to store all this data. And we were using a time-series database, we were using a relational database... And I remember at one point we wanted to sort the console by like uptime, but then show all the device metadata... And what should have been a simple SQL join ended up being like a two-week engineering sprint, because you had to connect these two siloed systems. And I remember thinking "Oh, this is awful. This sucks." And one of our engineers said "Hey, I could build this on Postgres, but it'll take me a month." And I was like "Okay, cool. You're an optimistic engineer, so it'll probably take you three months." And 12 months later... \[laughs\]
+
+**Adam Stacoviak:** Oh, gosh...
+
+**Ajay Kulkarni:** ...we had this database. And I'm trying to sell this IoT platform... And look, as an entrepreneur you learn to listen to signals of when someone is really engaged. And when someone's like "Oh yeah, that's cool", that's not cool. But when someone's like "Wait, wait, wait, hold on... Can you tell me that again?", you're like "Okay, there's something that's interesting to you."
+
+So here I am, trying to sell this IoT platform, and I'm meeting with the German head architect of this large shipping logistics company... I'm meeting them in Mountain View, and I'm telling him about this IoT platform. And he's like "Look, there's so many IoT platforms out there. What makes yours different?" And I was like "Oh, we built our own database, it does SQL, it scales for time-series..." And then he was like "Wait, wait, hold on... Wait, can you tell me more about that database?" And I think that's when I realized "Wait, this is solving a -- this is actually solving the right problem." And that's how it became a database company.
+
+So yeah, long story short, we kind of scratched our own itch and realized other people have the same itch... But again, we are students, we listen really well to the market, we try not to be dogmatic.... And when someone says "That thing's not interesting, but this thing is", we're like "Hey, we're not here just to build that thing. We're here to solve a problem. This solves a bigger problem. Let's explore that."
+
+**Break**: \[30:50\]
+
+**Adam Stacoviak:** Today you're not called Timescale, though.
+
+**Ajay Kulkarni:** That's right.
+
+**Adam Stacoviak:** So there's a name change of recent, which I think is challenging, but it kind of maybe shows the evolution... What's the evolution of that discovery, that 12-month discovery, scratch your own itch, IoT company, to Timescale to now be Tiger Data?
+
+**Ajay Kulkarni:** We started off thinking we were building a time series database for IoT. That's where we started. And so when we kind of pivoted to become a database company early 2017, we called ourselves Timescale. That seemed like a good name for a time-series database.
+
+**Adam Stacoviak:** I like it.
+
+**Ajay Kulkarni:** It's cool, it's pretty self-explanatory. And we actually saw there was a big demand in time-series that was more than IoT; some in finance, and events, and then soon crypto, and some other areas... But over time -- again, we were going with this journey with the industry. We started off as an enterprise sales motion... At the beginning of the pandemic we went all-in on cloud, stopped selling on-prem, and said "Hey, we're gonna be a cloud company." And yeah, at that point we were running databases for customers, and we got to see a lot of data. And I saw "Wow, there are a lot of companies who are like 20 people spending 50k, 100k, 150K with us." And I remember asking "Hey, what are you doing? What do you use this for?" And I remember one of them was like "Hey, I think of you as a better Postgres. I don't think of you as just a time-series database. I think of you as my main -- you are our main database. You are 50% of our cloud spend. You are our main database, because you are a better Postgres."
+
+And then we realized what we had built was not just a better time-series database, but a better Postgres. And we kind of expanded that better Postgres theme by adding vector support, and better native AI support... And we got to the point where this year we realized "Hey, we keep calling ourselves Timescale, and people view us as a time-series database", but we had already become something more than that. It's as if like amazon.com has started off as books.com. And you're like "Hey, at books.com we also sell CDs." People would be like "Yeah, but you're books.com." "No, hey, we also sell like socks." "Yeah, but you're books.com." And that's what Timescale was. Timescale was like "Yeah, you do AI, but AI with time-series, right?" And we're like "Well, no. AI. You know?"
+
+And I think our first attempt was trying to make Timescale mean more than time-series. Like, you know, when it's time to scale, I don't know... \[laughs\] And I'm not sure if you've ever seen the movie Mean Girls, but there's a line in there that Mike has quoted, where he goes "It's like trying to make fetch happen. Like, it's not gonna happen." Like, we're trying to make this term stick which wasn't sticking. And so we're like "Hey, you know what? We need a new name, because we are Amazon. We're not books.com." But we're not pivoting, we're actually changing the name to properly reflect who we'd already become.
+
+**Adam Stacoviak:** Yeah.
+
+**Ajay Kulkarni:** And Tiger Data was like the perfect name for us, because our mascot's always been the tiger... Internally, we talk about Tiger time, that's our all hands. We talk about state of the Tiger; it's my monthly presentation to the company. New people who joined the company are Tiger cubs... And -- yeah, we've an internal tiger mascot... And so the name Tiger just worked because it was who we already identified with, and to our existing customers it looked like "Yeah, it's your same logo. Cool, I get it." And some new people were a little confused, but we were like "You know what? In 12 months you will not be confused. And this is the right thing to do." So that's why we made that change.
+
+**Adam Stacoviak:** Was it scary to get to the point where you're like "You know what? The only way forward is a name change. Nothing else, just the name."
+
+**Ajay Kulkarni:** You know, it got to the point -- which probably means we'd waited too long... It got to the point where it was painfully obvious to us.
+
+**Adam Stacoviak:** Ah, yeah...
+
+**Ajay Kulkarni:** And I remember -- and I told the team "Look, this is not going to solve any problems, but it is going to remove an anchor that is holding us back." Like, it's still on us to kind of move the ship, you know... But it removes an anchor. I think the key thing is you just have to stick to the decision and be like "Yeah, it's going to be messy. It's going to be maybe a year of people being like "Who are you again?" And we being like "Yeah, we're Tiger Data", you know?
+
+I actually think the name is catching on faster than I expected. I think the name change went as smooth as we could have hoped, and I think people -- TimescaleDB I think still has more brand awareness, but I think Tiger Data is catching up.
+
+**Adam Stacoviak:** Yeah.
+
+**Ajay Kulkarni:** Yeah... Some people made fun of us, but whatever... \[laughs\]
+
+**Adam Stacoviak:** Well, haters hate. That's how it works.
+
+**Ajay Kulkarni:** Haters hate, dude, yeah.
+
+**Adam Stacoviak:** Haters do hate. So our audience knows this, you've been a sponsor for a bit, too... And I mentioned maybe in the pre-call, maybe in the early part of it, Isabelle is someone who works for you, and we've known each other for years, since the MongoDB days... And we would have conversations -- and this is when you were Timescale. And this is what I do whenever I sit down with a brand and I think about how can we help them reach our audience in a way that is informative, educational, and just something that helps them be curious and try it, if it's something that fits in their world, essentially. And I always do this version of an investigation, to some degree. I look at your homepage, I look at your products, what are you doing... And things just didn't pair up, especially when you mentioned with AI, "Yeah, it's time-series, but it's also just Postgres." And it was hard for me, because I would tell Isabelle, "Hey--" I would write her, "You've gotta change that headline on your homepage. It just doesn't work. Something's not fitting here."
+
+\[40:00\] And so when I saw the name change reflecting on your mention of "It seemed natural", essentially... When I saw the new name, the first thing I did was email Isabelle and I was like "Listen, that's an awesome name. We've gotta talk. When can you make time?" kind of thing. Because it had been a few months since I'd talked to her, and that's what happens when you do a rebrand and change your name; you rethink your model, you rethink your brand, and you come back out with a new plan. And that's what I saw. And so the moment - just to kind of reiterate, the moment I saw this new brand, Tiger Data, I was like "Mm-hm, that makes total sense. The Tiger's there, the logo's awesome... It makes sense."
+
+**Ajay Kulkarni:** I'm glad you felt that way, because I feel like people who knew the company mostly felt that way, which was like -- like, we didn't even change our logo. Our logo stayed the same. We just changed the name. And people were like "Yeah, no, they get it. And by the way, tigers are cool!"
+
+**Adam Stacoviak:** Tigers are cool.
+
+**Ajay Kulkarni:** They're cool, they're fun...
+
+**Adam Stacoviak:** What I liked too, when you did this, was the logo animated from Timescale to Tiger Data. Logo stayed the same, the mark that you have... I thought that was a nice little subtle touch.
+
+**Ajay Kulkarni:** That was our marketing team, yeah.
+
+**Adam Stacoviak:** I also -- I'm not a... I think a lot of people like dark mode, and I think your previous site was all dark mode... And you may even offer a dark mode version of your site... I don't know, because I see the stark white version of it... It doesn't bother me. Like, the yellow and the white and the black... Whoever was in charge of that process did a great job... Thinking through the core of who you are, and how do you come back \[unintelligible 00:41:44.11\] not being completely different, but being different at the same time.
+
+**Ajay Kulkarni:** Yeah. I mean, it was a team effort, but the marketing team drove that. Kudos to them... But I think it definitely gives you some thought on "Hey, does the name really matter?" And I think where I'm landing right now is that the name may not help you, but it can hurt you. I mean, there are two schools of thought. One school of thought is you name yourself books.com, the other school of thought is you name yourself Amazon, which could sell books. And I liked the clarity of Timescale, because it was like, "Hey, time-series." But in hindsight, it was limiting. While Tiger - dude, Tiger could be anything.
+
+**Adam Stacoviak:** Tigers are cool.
+
+**Ajay Kulkarni:** Tigers are cool.
+
+**Adam Stacoviak:** So one of the things that you have now, I think, is this burgeoning idea of agents in our databases, right?
+
+**Ajay Kulkarni:** Yes.
+
+**Adam Stacoviak:** And I'm at this point where -- there was this arc of acceptance, I would describe it, of "Hey, come on now... AI hype, I'm over it. Too much", this and that. And I think the game changed when Claude Code changed the game. That's where I really think it happened. I had been a user of ChatGPT, like many people... \[unintelligible 00:43:07.17\] the API to do something in ChatGPT and copy it out somewhere else, whether it was written material, whether it was an idea, whether it was a framework in terms of a thought framework... Or maybe even a Bash script, because AI is pretty good at Bash scripting. And then you kind of get to this other side where you're like "You know what? Wow, agentic is really revolutionary." And I think Claude Code really changed that game from the new browser or the new destination point for developers... It is still in the IDE, and that's still taking place today. But I think you're seeing this shift to the CLI that is just truly revolutionary, that now the terminal - it was never not cool, but it's cool again. A lot of folks are hanging out there, Claude Code kind of put that on the map in a way... And then a lot of folks decided to follow that direction. And I imagine that's kind of what you thought too, is like - you're probably doing a lot of agentic coding, you're probably playing with side projects, you're getting curious again... And you're like "Well, the next best thing is to -- how can I just talk to my database? How can I just put an agent in my Postgres?" Talk about that.
+
+**Ajay Kulkarni:** \[44:24\] Yeah. I think -- yes, I had a very similar experience with Claude Code. I remember one of my friends talking about agentic workloads in 2023, and I remember thinking "What are you talking about? Agentic? What does that mean?" And now I use the word "agentic" at least five times a day. And Claude Code was that moment. Claude Code -- I mean, ChatGPT was cool, but it was like a party trick. It was kind of cool, it could edit for me...
+
+**Adam Stacoviak:** For developers it was a party trick. I mean, for everyone else it's pretty dang powerful. But as a developer -- and ChatGPT is not the right interface. Chat yes, but not \[unintelligible 00:44:59.24\]
+
+**Ajay Kulkarni:** I remember building -- because I was trying to talk to a friend. I was showing it to a friend, and he was like "Oh, can we build an app that tracks pushups?" And I'm like "I don't know. Let's try it." And I think 45 minutes later -- Claude had gone out and found the right computer vision library, and other stuff... And I had like a mobile web app that would use computer vision to detect if you're doing a pushup or not, and it didn't get it always right, but it got it right maybe 80% of the time... Which was pretty good for 45 minutes of work. And I remember thinking "Dude, I can build anything. Like, I can build anything now." And I'm with you, I think Claude Code felt like "Wow, this is actually an agent doing work for me, writing code, making decisions... I can steer it, making me more productive..."
+
+I remember going to a social event that night and I couldn't even talk to people, because I was so excited. Like, I went home early. I was like "I need to go back to Claude Code."
+
+**Adam Stacoviak:** "I'm not feeling well, y'all. I've gotta go."
+
+**Ajay Kulkarni:** I swear to God. Well, I'm trying to talk to people about Claude Code, and they're like "Are you okay?" I'm like "No, this thing is amazing. You have to try it."
+
+**Adam Stacoviak:** What kind of party was it? Was it nerds, or was it normal people?
+
+**Ajay Kulkarni:** It was part nerds, part normal people.
+
+**Adam Stacoviak:** Okay. Part normal...
+
+**Ajay Kulkarni:** But even the nerds were like -- I mean, Claude Code had just come out. There was still "Yeah, I use Claude..." I'm like "No, not Claude. Claude Code." You know when I told you about that experience in 1996, using the internet for the first time? That's what this felt like. Number one, it felt like "I don't know what this is, but I want to be a part of this. This is fun." But also it brought me -- and I know a lot of my peers feel this way... It kind of brought out this childish sense of wonder, that I think years of being of an entrepreneur had kind of beaten down... And it kind of brought it back to the surface, to be like "This is fun." This is fun the way technology should be fun, where you're just like "I just want to tinker with this. I don't know what I can build, I just want to build things, explore this new world together."
+
+\[47:28\] And yeah, so I'm really excited... I mean, I think as a company we're excited, because we see this trend, that like 80% of Claude Code is written by Claude... The majority of new software over the next 12 months will be written by AI... We already have customers who tell us that 70% of their code is written by agents. And you just look at this and you say, "Okay, if that's happening, then what else is happening?" Well, that means the surface area of software development needs to evolve for agents. And the surface area of databases. So now databases are serving a new user. They're not serving a human, they're serving a human using an agent. So then you ask yourself, "Okay, then how does a database need to evolve for that?" Well, number one, it's less of a GUI... So you remember databases went from on-prem to SaaS, like cloud... But now you're going from SaaS to, I don't know, MCP, CLI... It's like, the interface is totally different. It's no longer clicking. It's more like commands... And there's some other things, too. Like, you want databases to boot instantly, you want to be able to fork, and create sandboxes quickly, in a safe way, in a cost-effective way... You probably want native search, native memory... Yeah, so I don't know, I'm just excited as like the little kid in me who just loves building things, excited...
+
+**Adam Stacoviak:** For sure.
+
+**Ajay Kulkarni:** ...but also as a company, it's like "Yeah, this feels like a problem someone should solve", and the answer is probably Postgres, and we might as well solve it.
+
+**Adam Stacoviak:** This lyric lands for me in this moment... It's from -- I had to look it up. I don't know this by memory, but it's from Marvin Gaye and Tammy Terrell. "Ain't nothing like the real thing, baby." So when you're in ChatGPT and you're just ideating the future of what your software thing could be, if that's where you're kind of hanging out at, versus that experience of in Claude Code doing the same thing, but it's making the thing. It's making the real thing, and you're seeing the real thing change. And maybe you're doing Git commits, maybe you're doing spec-driven or document-driven development and you're thinking a ton, and you're writing it down, or it's writing it for you, or it's writing the markdown... But the ideas and the vision and the direction is coming from your lived experience, and your wisdom, and places you've been, and valleys and cul-de-sacs you've been down, and you're like "Not going there..." And it may try to take you there, but you're like "No, no, no, hang on. Let's curve that back. Let's go here." But you're seeing the real thing be made. That to me was, I think, the moment where I was like "That's pretty wild. The real thing."
+
+**Ajay Kulkarni:** And it can be like "Okay, cool, commit those changes and push them upstream." And so you're like "Oh, it's actually fitting --" Because when I use Replit or Loveable - they're fine, but I was like "Okay, cool, but I don't build this way. Maybe other people do, I don't." But now I'm like "Yeah, I'd build this way. Yeah, commit the changes, push it upstream. Yeah, cool. Great. Show me the diff. Alright."
+
+**Adam Stacoviak:** The speed you can move at, I think, is kind of wild. I laugh about this, and I don't know how often you laugh about this, but I love when it makes a plan and it's like "Well, this is week one, this is week two, and this is week three." Meanwhile, four hours later, the thing's done. Or an hour later, whatever the number is.
+
+**Ajay Kulkarni:** That is always funny, yeah.
+
+**Adam Stacoviak:** And I like how it manifests its time ranges, but I'm like, "It's kind of funny." It's still making a plan, or it's making a plan with you, and the range of its timeframe is -- maybe it's actually accurate to how it should be if it was done with a team of humans, versus a team of one, with an agent and an idea. And I think that's kind of wild, it's just the --
+
+**Ajay Kulkarni:** Or maybe it's just under-promising and over-delivering, because that's what humans like... \[laughs\]
+
+**Adam Stacoviak:** Yeah, maybe that's it; maybe it's psychology. Something's going on there... Even your mention of -- you know, we had a conversation, you and I, a few weeks back, and you mentioned the same thing, you mentioned this childlike play... And that's kind of the definition of flow. Like, flow state is that -- when you're in that state of flow, that state of play, a lot can happen there. You stay more engaged... It's kind of addicting, in a way.
+
+**Ajay Kulkarni:** I'm a big fan of that idea, the concept of play. When something feels like play, that's when you know you're in it. You're in the flow, you are doing the thing you should be doing.
+
+**Adam Stacoviak:** Yeah.
+
+**Ajay Kulkarni:** And if I ever write a book, I think the title will probably be Play...
+
+**Adam Stacoviak:** Nice.
+
+**Ajay Kulkarni:** \[51:57\] I think that should be the driver. It's like, look for things that feel like play. Because those are things that really kind of fit who you are. And if you find something that feels like play and it also lets you kind of support yourself or your family - yeah, dude, that's the jackpot. But the AI story kind of reminded me of something... When the personal computer came out in the late '70s, people struggled how to describe it. Probably similar to how we struggle describing ChatGPT and Claude today. And Steve Jobs, who was probably one of the best tech marketers of all time - either he came up with this or Apple came up with this, but they would talk about how a computer is like a bicycle for your mind. Which is kind of a weird -- today that feels like a weird analogy... But I get it. Because back then you're like "Hey, this thing--" This is like pre-spreadsheet, right? "Hey, this thing -- we don't know what the key use cases are, but kind of like how a bicycle lets you get from point A to point B faster, this is just making you better." And that's how I feel about Claude Code. And there's all this discussion about "Hey, our agents are gonna replace us..." I mean, maybe some of the things we do, but for most of the things I think it'll be like a bicycle. It'll just help us get from point A to point B faster.
+
+Yeah, I don't know, I think that's part of the joy, is being like "Hey, this thing that used to take me a week, or I couldn't even do before, now and get done in an afternoon, totally throw it away and start again the next day."
+
+**Adam Stacoviak:** Try to juxtapose the timeframe of IoT company, using Postgres, needing time-series... Engineer, developer, employee, whatever, comes to you and says "I could do this in Postgres. I could do it in a month." You're like "Yeah, you're ambitious. You can do it in three months." 12 months later, that whole story... Juxtapose that time, to idea, to CLI, and - we haven't even talked about the CLI yet, but getting to this point of agentic Postgres, and building what it is. How fast did you go from idea to using the thing?
+
+**Ajay Kulkarni:** I mean, a lot faster. \[laughter\]
+
+**Adam Stacoviak:** A week? A day?
+
+**Ajay Kulkarni:** I don't even know, man... I mean, I think this is the theme in the AI land, is "Just ship." You know?
+
+**Adam Stacoviak:** "Just ship"? Okay...
+
+**Ajay Kulkarni:** Just ship. We're a database company, so we have a slightly higher bar than maybe a classic app... But yeah, just ship. Yeah, I mean, there's definitely a shift between -- it used to be hard to build, but now it's much easier to build, which probably means that more of the bottleneck then becomes a distribution and awareness... Which is I think something similar to what the iPhone has --
+
+**Adam Stacoviak:** It's always been the bottleneck though, distribution and awareness; it's just exacerbated now.
+
+**Ajay Kulkarni:** Yeah, but I remember once upon a time, pre-AWS... I mean, you probably remember this too, which is like, to start a company you needed to -- you needed a server rack, you need maybe a couple of Sun Microsystems machines, Oracle ADI database... And just to get started was such a high bar that a lot of people couldn't do it. I was actually gonna say, I think it's similar to what the smartphone did to photography... Once upon a time, being a photographer was really hard; even if you look at before digital cameras, right? And you'd only take so many shots, you'd have to develop it, it would take time... It was hard to get right... You would have to actually apply physical filters and physical lenses to change the thing... And with the smartphone, everyone became a photographer. And what that meant was that in the past, having great photos would set you apart, but now it's like, anyone can create a great photo... But how many followers do you have on Instagram? You know? So it kind of, again, shifted the burden from photography skill to distribution. And I feel like we're gonna go through something similar with software.
+
+**Adam Stacoviak:** \[56:08\] How much do you know about, I suppose, the details of the CLI and the MCP server that is in your products? Because I felt it was pretty -- when I looked behind the scenes of the open source of that, you know, it's elegant, the way you've married a CLI and an MCP server in a single binary.
+
+**Ajay Kulkarni:** That's right.
+
+**Adam Stacoviak:** That's pretty wild. I think the design is -- I had to borrow it, honestly.
+
+**Ajay Kulkarni:** That's good. Great.
+
+**Adam Stacoviak:** I've got a couple of things I'm working on, that are... One is a DNS resolver, and I wanna like talk to the resolver... And so the obvious answer is to have an API. But then you also have to have a CLI that talks to the API. But then I'm like "Well, I'm actually using an agent, so I'm taking a playbook from Ajay, and Tiger Data, and agentic Postgres..." And, well, the same CLI can offer an MCP server running at the same time, and you can talk to it and get back JSON for the agent, or get back CLI response for the human. Like, the design there was so elegant. I wanna know what you know about that design.
+
+**Ajay Kulkarni:** I know the team that built it is really smart... I think they looked at a couple of different options. I think the way it works right now is that the CLI will run a local MCP server, which then speaks to a remote one that we have. And the remote one is where we index -- you know, our MCP will allow you to do things with Postgres and Tiger, but it also lets you search the Postgres documentation, download best practices... And so we put that on its own kind of remote MCP server. I think MCP in general - it's such an interesting area, and I feel like a lot of people hate on it because it seems so simplistic... But I think just the idea that -- you know, people talk about, in the physical world, the idea of a design affordance. You know this concept? It's this idea that when you see something -- like, when you see this mug and you see this thing, you know you can hold this.
+
+**Adam Stacoviak:** Right.
+
+**Ajay Kulkarni:** You know? It's kind of like a door handle. You're like "I know I can turn this." This is one of the reasons why Tesla door handles trip some people off, because they don't see the handle.
+
+**Adam Stacoviak:** They're like "What's going on here?" Yeah, \[unintelligible 00:58:29.22\]
+
+**Ajay Kulkarni:** Actually, I think that's bad design. But an affordance is when the tool teaches you how to use it, because it's designed the right way.
+
+**Adam Stacoviak:** Like games, like Mario. The Super-Mario Brothers, the first game - that was why it was...
+
+**Ajay Kulkarni:** That's right.
+
+**Adam Stacoviak:** Level one was level one. It was teaching you how to play level two and three and four.
+
+**Ajay Kulkarni:** So this is where -- MCP to me is not just an API, it's essentially a way to expose tools for agents. To essentially think about like "Hey, we used to make these handles for human hands, but now we're gonna make them for agent hands." And that is a really growing art... Anthropic, who developed MCP, they're still learning best practices. I think that's really interesting... It's to figure out what proper tool design looks like, how many tools is too many... I think right now tools are mainly a wrapper around APIs, but I think it'd be cool if they were actually doing the job, not just the API.
+
+Yeah, so I think the team's done some good stuff. There's some really other -- some good stuff that we're gonna release soon... But I think MCP design -- I think this idea of designing for a new user, the agent, is I think a really fascinating area.
+
+**Break**: \[59:50\]
+
+**Adam Stacoviak:** Well, we've heard the term dev tools, and now we have agent tools.
+
+**Ajay Kulkarni:** That's right.
+
+**Adam Stacoviak:** That's how I think about things. And I've done a couple toy CLIs just to play with them. One for Proxmox -- so I'm a home labber... I don't know what kind of nerd you are, but I love to home lab. And one of the epicenters of my home lab is some version of a Raspberry Pi somewhere, or an Intel NUC, or maybe a slightly bigger one, which is a NAS, which is usually TrueNAS... And the next best big thing in the network, for me at least, is a Proxmox machine, which is dedicated, where I can do different things, whether it's a service, or tinker with something, or play with a new version of Linux, or just distro hop, or distro play, however you wanna describe it... But as I've explored this world a bit, and I've understood what it's like -- I'd never really built many CLIs before. And I was like "Well, the next thing you wanna do there is really just... You wanna enable, now that we have agent decoding, and it's -- it's almost ubiquitous." So when you build a tool, or you build a new CLI, or you think about a new way to create an API with a CLI that lets you talk to a service, you can assume, or presume pretty good, with pretty decent accuracy, that the person may be using an agent; likely using an agent, or at least has one nearby, and is working towards being okay with that. Some people are still not okay with that.
+
+So I was like, it only makes sense now that when you design these tools, you can't just design the old way, which was "Well, I'm designing this, obviously, to solve a human problem, so it's gonna be a human using it..." Now we have humans that are still there, they're still the human in the loop or on the loop, depending upon the framework, but that human is leveraging one or many agents, or sub-agents, to automate and to do different things in their tasks, whether it's designing the plan, whether it's writing all the code, and they're reading the code... They're still doing the code review; maybe they're doing 20% of the coding, because the agent does 80% of it, or 90% of it, or whatever the number is... Likely, this person has an agent. And so we now have to look at the way we build tooling in a way that's like "Okay, I used to do it this way, human-only, but now the human has a friend, an agent... I've gotta design for both of them", because that human is expecting their agent to have agency on their behalf, as they command it, to do the task.
+
+**Ajay Kulkarni:** A hundred percent.
+
+**Adam Stacoviak:** And the agent speaks a different language, reads a different way, and it interacts with a command line or an API in a different way... And that's where MCP, or even like Claude plugins come into play. How do you feel about that world? How do you feel about agents in our Postgres? ...and now we can actually talk to our database.
+
+**Ajay Kulkarni:** I mean, agents will be in everything, because I think it is someone helping you use the thing that you're using. I mean, you may not have an agent like inside the database itself, but you probably have an agent inside the system that's running the database. That's one thing we're exploring right now on Tiger Cloud, is having agents that help you.
+
+No, but it's totally right. Again, it's not replacing the human. It is the human doing the thing they need to do, but now they're doing it through an agent. And so now you're saying, "Okay the job to be done is still the same, but the thing doing the job is different." And so it's gonna need different things than what the human needed.
+
+**Adam Stacoviak:** Absolutely.
+
+**Ajay Kulkarni:** Yeah. I mean, I think some places we're looking at is like "What are some of the menial tasks in a database that you don't really wanna do?" Okay, cool. Maybe you start there. But also, what are some tasks that you can't do? Maybe an agent can, because agents aren't limited by time and space. They can easily paralyze and run things and shut them down.
+
+**Adam Stacoviak:** \[01:06:06.14\] When I look at what you've done with agentic Postgres, which I think is -- I'm not sure if you're trying to brand that term... I think it is the first time I've heard it, and I wonder if that will just become the thing, and it's a product named like Xerox, or Kleenex... And I won't be using Tiger Data's agentic Postgres, I'll be using somebody else's, because there are more Postgres database providers out there... Yeah, I wonder if that's the case. But when I look under the covers of this thing, and what you've done, it's the first time I can look at something and say "Now I can actually talk to the database. Now I can actually sort of command it to go do something and it goes and does it." It could even be a branch or a fork or whatever terminology you're using to say "Give me a safe area to play in." That's just never been a thing before. And now it kind of is.
+
+**Ajay Kulkarni:** Yeah, I mean, this CLI -- I mean, I think CLIs have been useful for... They've been a thing for a long time. What, 15 years, maybe more?
+
+**Adam Stacoviak:** Yeah.
+
+**Ajay Kulkarni:** I mean, I don't know, if you look at Unix, maybe it's forever, you know?
+
+**Adam Stacoviak:** They're on their way back. I feel like CLIs are like the next frontier, in a way, really. I feel like the next year or so we're gonna have this massive swath... And maybe some will live and some will die, but that's gonna be the delivery mechanism. So yeah, it's just --
+
+**Ajay Kulkarni:** I think the key thing is the CLI is a clear program that agents can use. So okay, so when we built up Tiger CLI - it's written in Go; it's distributed as a single binary. It essentially provides a unified control plane for your database on Tiger Cloud. And so it covers auth, obviously creating, starting, stopping, forking databases, connecting databases... But what's really, I think, nice about it, like you pointed out, is that it has MCP baked in. So you can run essentially a local MCP server through the CLI, that then your Claude Code or whatever can talk to... And now your agent has access to all these basic commands. It has access to starting, stopping, forking databases, connecting the databases... It also proxies another MCP server that we host, where we have indexed the Postgres and Timescale and Tiger documentation, and now we provide semantic search and text search on there. So again, it's not just a CLI MCP server, it's a CLI MCP server that actually makes your agent smarter.
+
+I don't know, I would say -- I'm not sure if this is true, but it probably is... It's probably the first AI-native CLI for Postgres. It's not just -- I think other Postgres CLIs are just like API wrappers, but this is really thinking about bringing the CLI commands into the agent, but then also giving agents the Postgres context it needs, best practices, documentation, so that it's instantly like an expert developer.
+
+**Adam Stacoviak:** Yeah. I mean, I know a lot of developers hang out in the command line, but I feel like to some services the CLI was not so much an afterthought, it was sort of "You have to do it", but not a lot of people would use it as the kind of primary interface. And now that we have agents in the command line, I feel like it's table stakes. If your service doesn't have a fleshed out, thought through, well-defined API, so that you can also have a thought out, well-defined CLI... And then maybe MCP lives and maybe it dies because agents can just sort of map around... I know Claude has plugins and things like that so you can sort of use Markdown to teach it, of all things not TypeScript, not Go, not Rust, but Markdown. Like, words.
+
+\[01:10:06.22\] Maybe MCP - there's talk about "Is it a fad, is it not?" I'm still not sure, because how sure can you be when things change on the weekly, even on the daily...? But just this idea that now the new frontier for a service is a cloud, is an API, and is a CLI at its best... And maybe MCP lives or dies, but for sure, you've got some sort of plugin system that you can teach with a slash command in the CLI that most people are using, if it's OpenCode, if it's Claude Code, if it's Auggie, if it's AMP... Like, you pick your agent...
+
+**Ajay Kulkarni:** That's right.
+
+**Adam Stacoviak:** ...or Codex, even. They're always last on my list, for some reason... But you know, the moment you give them that slash command, now the agent has access to the CLI... If it's got great help docs, so --help on all your commands... Now the new frontier is like this really well fleshed out CLI... I feel like that's where so much attention is going to be going, if not already there.
+
+**Ajay Kulkarni:** No, it's a great point. And I think -- and again, hey, is MCP gonna be the end all protocol? I don't know. But there will be a protocol, and I think that's the point. That's the point you're making, and it's the point I'm trying to make. And why do you need this protocol? Well, number one, you need a way for the agents to interact with your APIs. But number two, you probably want to give them access without giving them the full keys to the kingdom. You wanna give your agents just enough power, but maybe sandboxed in a certain way. Maybe they can do some things, but not others. And I think this is where MCP, your agentic APIs really play a role.
+
+And then the CLI is another way to do that. The CLI is another way to give essentially a tool to the agent that is properly defined in what it can and can't do. Yeah, I think it's really interesting. I think the team's done some really cool -- I think this docs thing... And I'm not sure if I'm beating a dead horse here, but I think what they did with the docs I thought was especially clever, that I haven't seen anyone else do. It's like "Hey, we have an MCP server where we have indexed the Postgres documentation and the Tiger documentation, and we've indexed it for text search and semantic search, so that your agent can effectively get a PhD in Postgres." And we've done the tests. You can create schemas without our MCP, and then with our MCP, and then have Claude compare the results. And with the MCP, it's just a much better schema.
+
+I think that's been really cool, and I think that is -- it's almost a variant of the skills that you see in Claude, but I think that's another element of agents that I think is underexplored, is the fact that with the right markdown, they can become experts in anything.
+
+**Adam Stacoviak:** Did you see the talk from Barry Zhang and - I think his name's Mahesh Murag. They're both from Anthropic... It's the first time I've seen -- well, I guess the first time I thought I would see anybody talk on stage for an hour about markdown. They introduced --
+
+**Ajay Kulkarni:** \[laughs\] I think I saw the tweet about it. It was like a packed room, or something...
+
+**Adam Stacoviak:** Yeah, an hour long talk about using markdown. You just wouldn't expect that. That's not the language I thought would be front and center. Now, I love reading markdown. I personally prefer markdown. I've been an Obsidian user for years now, and so I'm all about flat files, nothing that's sort of obfuscating my prose from my source... And markdown is that. But it's the first time I thought I'd ever see somebody on stage presenting a revolutionary new feature that is now used by so many folks now. I'm starting to use plugins... And it's markdown.
+
+**Ajay Kulkarni:** \[01:14:03.02\] It is wild. In fact, I have a friend working on a markdown editor for vibe coding. Not for note taking, but for like vibe coding. No, it's a new world, man... Maybe we'll find -- that's the weird thing with these agents, is that in some ways they are very similar to us. You don't teach it using machine code, you teach it using text, in markdown.
+
+**Adam Stacoviak:** Words...
+
+**Ajay Kulkarni:** Words.
+
+**Adam Stacoviak:** Words, man... Words.
+
+**Ajay Kulkarni:** Yeah, it is wild.
+
+**Adam Stacoviak:** I do feel that.
+
+**Ajay Kulkarni:** Well, I also think -- with your discussion earlier, I had a thought about the terminal, about how more people are using the terminal, living in the terminal... There's probably a business idea out there - I'm sure someone's working on it - of building the modern terminal.
+
+**Adam Stacoviak:** I think... Well, my favorite - and I'll give them a nod, because they are my favorite - is Warp. I think they've been... Gosh, I recall talking to them almost five years ago, and it was "The future of the terminal", I want to say... "The terminal of the future", I believe, was the title we shipped on that. And Zach Lloyd is -- I think he's a solo founder even, of Warp. And Warp is my daily driver. So that's what I use when I'm in a terminal, I'm using warp.
+
+Now, I know Mitchell Hashimoto and Ghostty is awesome. I love it. I'm so thankful for what they're doing, and especially what Mitchell did with the foundation kind of piece to it, to kind of give Ghostty a home forever... But something about Warp just gets me, and so I just like Warp. So I think they may be ahead of the puck in terms of that.
+
+**Ajay Kulkarni:** Yeah. Well, it sounds like Warp -- did Warp start before this whole AI thing?
+
+**Adam Stacoviak:** Oh, yeah. It's probably two years before, I want to say. At least two to three years before GPT-3 made people think that emails will forever be written by machines. And we're like "Nah, it'll never happen." You know what I mean? And now, "Can you summarize this into an email form for me? I will edit it, and I will remove what I don't like", or whatever... It's not how I write all my emails, by any means, but a few that require attention, or care, or particulars... Why would I write all that out? I can have a thought, and the thought process with the agent, and then it formulates all the thoughts I've just given it into a cohesive form, and I can move along to my podcast with Ajay... Which is actually what I did. I sent an email just before our call, and my agent helped me write it. And I'm cool with that. I read it, I know what the words were in it, because they were all part of a 20-minute back and forth about this or that... And I'm like "Hey, can you package this entire sequence up into an email?" And it did. And I cut out most of it, and edited it to be Adam... That's cool.
+
+\[01:17:14.00\] You can even teach your agents to think like you think. Especially in Claude, with these plugins and stuff like that, and the way you can change the prompt behavior to make it more antagonistic to your ideas, and not be like a yes agent. "Yes, Ajay, that was the most brilliant idea ever. Let's do this." No, no, no. Push back. So you could do a lot of that with these things, and it's just...
+
+**Ajay Kulkarni:** It's a bicycle for your mind. With a mind.
+
+**Adam Stacoviak:** Yeah. And now we have it for our database. I think that's such a wild thing, man. But what's on the horizon for you? I know that we can't uncover things that are maybe releasing soon, or... I'm not asking you to spill the beans, but what's over the horizon? A little bit that's coming up soon, that you can at least tease? What is that? ...for you, for Tiger Data, whatever.
+
+**Ajay Kulkarni:** Yeah. Well, so we started off as Timescale, building a database that scaled Postgres for time-series workloads. And then it became a Postgres that scaled for workloads in general, especially analytical workloads. What we got into with agentic Postgres is now serving not just the scale use case, but the start use case. It's like "Hey, when I am building a new application with my agent, give me a database that just works." And what you'll see coming forward is just us continuing to innovate on those two dimensions, start and scale. On start we're making our MCP even better, and CLI even better, and I'm curious what you think when these changes come out.
+
+We've actually been toying on a side project that is like a separate MCP server that is a little bit more than the database... And this is actually being developed by -- remember that engineer who said it would take a month and it took him 12 months? \[laughs\] It's the same guy doing this one.
+
+**Adam Stacoviak:** Wow.
+
+**Ajay Kulkarni:** But he's moving a lot faster. And we launched a new storage layer called Fluid Storage to enable kind of ephemeral forks... You'll see more stuff around that. Fast forking, safe forking, sandboxes... More around search, and more at scale.
+
+Yeah, and I think -- again, the job we do is "Hey, when you're a developer and you're looking for a database, we give you the database that just works." And it's built on Postgres, so you know you can trust it... But "just works" means something different at start than it does at scale. At start it means "Hey, it just works with my agent, it makes things easy." At scale, it means it's also reliable, and cost-effective, and fast. Yeah, we're like 170 people, and we're pushing on all those directions.
+
+**Adam Stacoviak:** That's a lot of people. I didn't expect that number. I should have asked you that sooner, but yeah, that's a lot of folks. Wow, that's a lot of folks behind you. Pretty wild.
+
+Well, Ajay, it's been a pleasure diving into where you've been, and why you've built what you've built, and how you've built what you've built. It's been fun. Thank you.
+
+**Ajay Kulkarni:** Thank you, Adam. Pleasure. Thanks for having me. Good to see you.

Antirez returns to Redis! (Interview)_transcript.txt

@@ -0,0 +1,401 @@
+**Jerod Santo:** Today we are catching up with Salvatore Sanfilippo. You may know him as Antirez. You may also know him as the creator of Redis. I say catching up because you have not been on the show since 2011.
+
+**Adam Stacoviak:** Gosh...
+
+**Jerod Santo:** That's a long time. Welcome back.
+
+**Salvatore Sanfilippo:** Thank you. Thanks for inviting me.
+
+**Adam Stacoviak:** Let's say forever ago, Jerod. It's the beginning, almost.
+
+**Jerod Santo:** Different times. Those were different times, for sure.
+
+**Adam Stacoviak:** Different times, for sure. Not mainstream tech; underground tech, as you said.
+
+**Salvatore Sanfilippo:** We were younger, for sure.
+
+**Jerod Santo:** We were all much younger. Yeah, like 13 years younger.
+
+**Adam Stacoviak:** We were younger...
+
+**Salvatore Sanfilippo:** Yeah, it was a different environment. There was still the spirit -- even maybe IRC news groups were starting to fade... There was still this spirit of community, what you made, it's cool, let's collaborate about that... It was a lot of underground thing.
+
+I believe that, on the average, programmers are very nice to each other, are very good at communicating with each other. And in general, I believe it was a very nice environment. And in some way, I miss that, but maybe it's just that I am old.
+
+**Jerod Santo:** Rose-colored glasses, we all put them on from time to time, and look back at the old days. Well, we're still talking Redis, though. I mean, 15 years later, and we're still talking Redis. Had it been a few years back with you on the show, we probably would not have been talking about Redis so much. That's part of the story. But can you go back to the beginning for us, and why you built Redis in the first place? First of all, thank you so much for building that piece of software. It's been so influential. I've used it for many years... Dozens, hundreds, thousands, hundreds of thousands of people have also deployed it and it helped improve their lives... So thank you for creating it. But can you go back to the beginning and tell us why, and how, and when, and where?
+
+**Salvatore Sanfilippo:** Actually, thank you for using it, because if you think at Redis, even if it compiles certain ideas in new ways, they are all old ideas. So I believe that software is mainly a platform for people to communicate concepts. And without a community -- Redis was so resilient; it lasted so much time because the community created a mental framework about fixing problems with such tools that are inside Redis. So basically, the community is Redis, in some way.
+
+And the way it started was because -- my first days as a programmer were all about security. I worked in security for a couple of years. Then what happened in security is what is happening right now to software. It started to be cool, it started to be business, a lot of business. And security was no longer research, it was a product. And I grown immediately tired of this industry, so I started to write embedded systems.
+
+\[07:52\] And then I started to get a bigger project, because I was in a very bad situation in my life. I had struggles with my relationships, I had a son that was like three years ago, and I was - I could say depressed, but it's impossible for me to get depressed physiologically, because the kind of attitude I have in general. But I was low in my life, and I needed a project to focus my energies.
+
+So the first weeks I just watched Star Trek Voyager nonstop, and South Park... And then I started to work to an interpreter for the Tcl programming language, for embedded systems. And much of this code was the foundation to make Redis one year later.
+
+So basically, with this interpreter, which was not just a toy interpreter, it was compatible with all the Tcl programming language set, extended it in many ways... It added continuations and other stuff, and was in some way, in some things even faster, and so forth. Finally, it clicked something in my mind about how to design bigger systems. But then I forgot about this, and since I had no money, I remember that in this period in my life, I often looked under in my car to see if maybe one, two euros was around. And then I asked even for money to the girlfriend I had at the time. So I said, "Okay, that's not possible. I need to do something." And so a friend of mine \[unintelligible 00:09:47.11\] went and said "Do you want to collaborate? Let's do something together." And this was the start of Web 2.0 in the United States. But in Italy, there was still -- what was Alta Vista, basically, in the U.S. So these portals, static portals, no user generated content. So we said, "Probably if we just copy what they are doing, we will have an edge."
+
+And we created two services. One was exactly like Delicious, and another one was an improved Digg, much more like Reddit now, basically. So there was complex moderation, there was a Tcl program that did the advanced filtering in order to avoid friends voting themselves. And this became popular in Italy, and so Telecom Italia, which is the biggest ISP here in Italy, a telephone company, called us. They didn't know in any way us. They called and said, "Why you don't join the Telecom Italia network with your services?" \[unintelligible 00:11:01.01\]
+
+I remember when we went in Milan I was dressed so badly that they didn't want to let me at the hotel in. They're like "Are you sure that you have a reservation in this hotel?" I said, "Call this number." They said "Okay, you can enter." Because we were underground people. In Palermo we had Linux systems with cables in order to pay, and we had illegal greens to connect to the telephone to internet... We did this kind of thing, so this was our mood. They were business people, so there was a cultural gap, in some way.
+
+And so we entered the network, and from one day to the other, we had to sustain the traffic of a lot of users, all the Italy connected. And our systems were not ready. So the first thing that I did was in PHP - the system was written in PHP - to write a memoization function that basically whatever you could call inside this function, the return value was stored, serialized into some kind of BTree that I created on the disk.
+
+\[12:22\] And this way we had some kind of cache in order to avoid stressing MySQL too much. And this worked the first weeks, but then we created a new service that was a real-time log analyzer, and the use case was not static caching. You know, till this point I didn't know even that Memcached existed. So I could not use it, just because of the lack of knowledge of existence.
+
+But even if I had Memcached, that was not enough for me, because I needed a cache with data structure where I could add items and remove all the items. So I took Tcl and wrote a prototype, and they tested it, and it worked so well. And they said, "Okay, that's good, but it's not a scripting language thing. I want an actual daemon in the ground, with a low memory footprint." And especially, I realized that I wanted a fork in order to persist on disk with the other child.
+
+And since I in the past had background in security and low-level programming for cryptography, I studied the very in-depth the Stevens books, all the system call of Unix, the implementation of small Unix systems like Minix, provided a few patches to the TCP/IP stack of Linux... So I was a very low-level already, and I realized that with fork - this is an interesting story, because fork, after 15 years, still is the best way to do the Redis thing from the point of view of persistence; this magical thing of copy and write of the kernel pages. So this fundamental idea back then survived to these days.
+
+And so I started to write this Redis thing, and it took me two weeks to have the basic core in order to apply it in our system. And since I had a long story of releasing open source software, \[unintelligible 00:14:34.05\] device drivers for the Canon cameras, web log analyzers, many projects that were in Debian and in other Linux distributions, I said "You know what? If I release Redis, this will not impact in any way our business", because the other application was closed source. So I released it on Hacker News, because Hacker News was very new when I released Redis in Hacker News... There were like 10 upvotes, a few people commented, but basically nobody cared at all. But Ezra Mobius...
+
+**Adam Stacoviak:** Morbius, yeah.
+
+**Salvatore Sanfilippo:** Yeah, that unfortunately now is no longer with us - he was a very smart guy, and I want to take a minute in order to remember him, because when I went the first time in San Francisco, he interviewed me and we did a talk... And basically, he realized \[unintelligible 00:15:39.29\]
+
+**Adam Stacoviak:** \[laughs\] I've just found this recently, because I was in his office... I've just found that recently.
+
+**Salvatore Sanfilippo:** He was doing the operations for GitHub. So basically, Ezra said to GitHub, "You have this delayed jobs thing, but you could use Redis", and he started to talk. But for six months, one year, I continued to work to Redis just because I liked it. And also, there was the contract with Telecom Italia, so I received the money in order to pay the bill... So who cares about actual work? I want to do Redis, and I did Redis.
+
+\[16:18\] And actually, as a kind of person, I am so erratic from the point of view of doing only what I want, that it's a miracle that I was able to work so many years to Redis. And what the trick was to just focus every time in a subsystem that in a given moment interested me. So it's like if you have many projects, basically.
+
+So thanks to this start of conversation, and I continued, and then I posted again the news... And also, one thing that I did was that every open source user that I had, I handled it as a customer. Like, it didn't matter if he or she is not paying me, I offered basically premium customer service \[unintelligible 00:17:11.01\] So people started to trust me, and this trust that they had in me started to migrate to Redis, because I believe -- you never trust just systems. You want to trust also the people that are behind a given system... And everybody started to use Redis more and more.
+
+But after like two years, it looked like it was inflating. Like, there was a very strong movement about very consistent systems, serializable systems. And I was very against it, in the sense that I thought "That's great that you have systems that have very strong form of consistencies, but not all the kinds of problems are made the same, so I believe it's a good idea we also have other kinds of systems." Because my motto was "The system usefulness is the sum of what happens when the system works without failures, and when the system works with failures." It's true that during severe failures, if you make complex partitions, Redis may lose \[unintelligible 00:18:23.04\] and so forth... However, on the other side, because of this sacrifice, you have things that otherwise are impossible to achieve from the point of view of latency, the kind of data structure you can support, and so forth.
+
+But many, many followed this path. For example -- I don't want to tell names, but they failed, because they were so compelled... At same time instead, for example, MongoDB prospered, inferior initially. Now I believe they are very strong; initially, technically, infrastructure. Because sometimes people want things that work, and not that work in the worst situation.
+
+So basically the thing inflated, to the point that one of the people that was contributing with me back then brought me one night and said "Save yourself and run away from Redis, because it's a ship that is sinking."
+
+**Jerod Santo:** \[laughs\] You didn't listen?
+
+**Salvatore Sanfilippo:** No, no. I said, "Okay, I understand. It can inflate, but then I will find a new project. I will not die if Redis is no longer used." But instead, after this stage, especially in the San Francisco area, to use Redis was worse than saying "I vote for Trump", or something like that. It was impossible to say "I use Redis" without people watching you very -- or it was even worse than not having a kitchen in your house, because it was very big to have a kitchen, you know?
+
+\[20:13\] And so people that I ran away from this Redis thing... But then, for example, there were very important companies that in the wave of this feeling was switched away from Redis, and then six months again they were back to Redis, because they created a lot of issues, trying to serve with other services. And so I continued. And at this point they said "Okay, this thing didn't die in this moment, so it's worth to --" Also, I was with VMware at that point, and they paid me a good amount of money, so I was no longer working for free... And then I switched to Pivotal, and then to Redis Labs. And it was a good setup. Basically, this was the start of the story. And then the story continued with incremental improvements to different subsystems of Redis, always trying to avoid making it too much complicated.
+
+**Jerod Santo:** So we started off talking about how times have changed, and you mentioned that when you first started it, or those first couple of years, you treated every new user like a customer, and that's part of its success story. Would you advise that today, for people who are starting open source projects, or do you think that was foolish, but it happened to work for you? What are your thoughts on that strategy today?
+
+**Salvatore Sanfilippo:** I believe it is very important to don't fall in the trap of saying "This is an open source project. I put it on GitHub, and then you don't ask me anything, because I'm already providing this work for free." But actually, the involvement of the users that try to use your system is not inferior to your involvement. They do things for work. When they use your system, they are exposing themselves to you faking up in some way or the other... So it's a big gift that they are doing to you. And I believe that minimum that you can do is to be as helpful as possible, as long as you maintain the project. Then if you don't want to maintain it anyway, you say "Okay, I'm done with this. I'm doing some other stuff." So absolutely, I believe.
+
+Also, the first users are not the mass users that you have in the next years. They often are extremely bright persons, that will provide you very valuable ideas. And there is so much value in that community, in that initial community. So I absolutely will do this again.
+
+**Jerod Santo:** So you had a couple of jobs along the way. You mentioned VMware... There's one in between that and Redis Labs. What was the one in between?
+
+**Salvatore Sanfilippo:** It was VMware, Pivotal and Redis Labs.
+
+**Jerod Santo:** Okay, so Pivotal. And so these organizations employed you full-time to work on Redis at your discretion, or was there strings attached? And then how did you get all that going? Because back then that was pretty rare. I think it happens now and again now, but that was pretty rare. So how did that work out in practice?
+
+**Salvatore Sanfilippo:** Okay, this is one of the best sides of capitalism. I believe that inside VMware there were two persons - basically Derek Collison, was one, and Mark Lukoski was another one, that realized that I did some good work, and they just did this for me without return. I believe it was like a gift. I will be always, you know -- thank you to them, basically.
+
+**Jerod Santo:** \[24:15\] Yeah, grateful.
+
+**Salvatore Sanfilippo:** I don't know if I ever said clearly thank you to these two persons... Because they didn't reason back then in terms of just "What's better for our company", but "Our company is big, and this money don't change anything. And what's a good thing for the software landscape, for the underground software landscape that we can do?" Sure, there was also a cloud strategy about VMware, and stuff like that, but I was free to do everything.
+
+**Jerod Santo:** That's amazing. And Pivotal was similar?
+
+**Salvatore Sanfilippo:** Initially yes, but they were probably a bit less happy about this setup. And so because Pivotal and Redis Labs had the same VC, they talked and said "If somebody has to pay this guy, let's pay you, that is Redis Labs, and not us, that do something else." And so I switched.
+
+**Jerod Santo:** I see. So the Redis Labs one is the confusing one, because Redis is your baby, and this is a third party that named themselves Redis Labs and hired you? Or did you start it? Can you explain that, how that worked?
+
+**Salvatore Sanfilippo:** Basically, they started the company and approached me immediately after starting the company in Germany. We were in Germany, there was some kind of conference, a NoSQL conference of some kind... And they asked if I wanted to join the effort earlier. But I was in a too good situation, basically taking money for free, and working on my system, so I said "It's okay for me if you do business with this thing. The license allows it, so..." But eventually, it was the most obvious thing that trying to make money with Redis would also pay me.
+
+**Jerod Santo:** So all three pretty good setups, some better than others. Obviously, the VMware one was the most, I don't know, philanthropic, or less capitalist. They just decided to do that and let you do your thing, which is awesome. Eventually, you grew tired of Redis? I mean, you left - was it 2020, as lead maintainer?
+
+**Salvatore Sanfilippo:** Not of the system. I was very happy of the work that I was doing. However, till the very last day, even if I had -- inside Redis, the company, there are many gifted people in the Tel Aviv side. There are very strong programmers that are in the company. Since I started Redis, they never left, and they have a very detailed understanding of the core of Redis, so they helped me with the most complicated bugs, and stuff like that. However, many of the user-facing features were still developed 80% from myself. And after 11 years, of all these users -- I, for example, remember that even when I took vacations of any kind, I always had my bag with my computer, and oftentimes I had to open it every day, because there was some issue that could be critical, some bug report from crash... And since a lot of systems run on Redis, for me the stability of Redis was the most important thing. Absolutely the most important thing. So every bug was investigated in details, with even weeks of efforts in order to replicate it, track the bug, to the point that after some time we realized that many of the bug reports were broken memory.
+
+\[28:12\] So when Redis \[unintelligible 00:28:10.08\] performs for 10 years or even more a memory check of the system it is running on in order to report broken memory. And now it's of no use, because most memory is error protected. But back then, a lot of cheap memory created problems.
+
+So basically, it was a life completely dedicated to this system. And I said, "I want to stop." Also, when I saw GPT-2, and then GPT-3, I said "This is going to change everything", so I wanted to write a novel about this thing. And I started to write while working, but I realized that the writing thing was a full-time stuff, so I said "I need to retire sometime in order to write this."
+
+What's happening - for example, in the novel, there is, I believe, the first description to date of prompt engineering... Because it was written in the novel before ChatGPT and -- there was just GPT-3 without the instruction tuning models... But it was obvious that we were headed in this direction, so I had the urge of doing this thing, and so I stepped back.
+
+**Jerod Santo:** So along the way Redis went through some license changes with various feedback from community members, and non-community members, from open source, to SSPL, and AG... There's lots of stuff we could talk about here. How interesting is licensing to you, Salvatore? Do you like this topic? Do you dislike it?
+
+**Salvatore Sanfilippo:** I believe that we that formed ourself in the '90s were extremely aware of the licensing stuff.
+
+**Jerod Santo:** Yeah.
+
+**Salvatore Sanfilippo:** Now I see younger developers not really caring about this, but we basically started to write code, and starting to be mini lawyers, understanding all the subtle things about GPL, BSD, MIT... So yes, I care about license, because the way that you express your willing about what others can do or cannot do with your code.
+
+Also, I think that without the copyleft idea, the computers technology could not accelerate to the point that it accelerated. Because when open source was created, basically in order to create a startup, you had to buy complicated workstations, Unix licenses, database licenses... So it was impossible to have the landscape that then it was created by the open source movement. So for me, licensing is very important.
+
+Initially, what I did with my software was to use GPL. Then I started to realize that GPL had two problems. It created problems to myself, because sometimes I thought -- and if this becomes big, I want to have a business model. And I don't want to get some paper signaled by all the contributors that I had so far. So I started to switch to BSD, saying "This is a protection for me. Also, it's a protection for other people based on the environment they use my software." And sometimes it can be a problem, even if they don't want to violate the license, and stuff like that.
+
+\[32:00\] When I started Redis, I was very into this BSD stage of my life, so I released the Redis as BSD also because there is also behind that an idea about accelerating society, improving society to be more important than basically what is going to happen to me, in some way. However, then it must be said that the cloud situation changed the landscape, because even if it was very complicated to create a product business model, even before - basically, Red Hat was the only one that really succeeded in this kind of game in the open source, and a few more... But still, if you wanted to sell services, you were the to-go person as a creator of the software.
+
+And then after AWS everything changed, because there was no longer need of somebody supporting you, because it was handled for you. And also, you couldn't even compete with the others, because in order to compete in cloud services, you have to pay for the instances, and they have them for free. Also, even the billing is complicated. There are many companies that just because of billing will just get what AWS has, and stuff like that.
+
+So I understood at some point that the BSD license that I picked, with the changing world of the software, created serious issues to create business. Now we can go a step back and say "But why it should create business? It's an open source software." And I believe that more or less every complicated open source software has in one way or the other an economic system behind it, because it's a lot of hard work for many years. Either people are paid very well, or they will not afford to do all this kind of work. So I believe that both things are needed to redistribute to the community, and also.
+
+And inside Redis, they didn't want to change the license, and I didn't want for a long time... But there was this discussion, but it was some kind of taboo. So it never happened as long as I was there that somebody asked me "But what do you think? We want to change the license." It's a conversation that didn't happen. I just created the module system, and Redis the company started to have the modules that were enhancing Redis capabilities, and that was it.
+
+Then when they changed the license, I understood that it was basically some kind of a forced move in some way, because with BSD it was too complicated to compete in this market. However, now we are realizing - me and also inside Redis - that SSPL was not accepted by the community in some part. And we care about this thing, because you know, I don't believe that SSPL is a terrible license, because it's very similar to other GNU licenses. It's just a couple of sentences... But the reality is that culturally it's not accepted. And so we are starting to discuss inside the company about this problem.
+
+\[35:52\] Also, one important thing is that because of that, we are going to add in Redis a lot of the features that are now only for the paying users. For example, now I'm working a lot to vector sets, which is the first fundamental data type that Redis gets after many years... And it will be released in the community edition, like normal Redis. And like everything I did in the past, it's no dependencies, so it compiles because the data structure -- so the HNSW data structure for vector similarity I wrote from scratch, the quantization I wrote from scratch, the hybrid search... So it's some code, it's like 6,000 lines of code in total, the other Redis data structure. So you can open the code and understand how it works. And it was impossible before to do that, but still, maybe there are setups that make everybody happy, enough protection and the community will be more happy. So there is an ongoing discussion inside the community now. I'm not sure what will happen, but we are focused on the problem.
+
+**Jerod Santo:** So you probably heard that Elastic went there and back again. They are fully open source, with Elasticsearch specifically, once again. And we had Shay Bannon on the show last year, talking about that decision, and he says that he thinks nowadays, in today's culture, the AGPL license, which is accepted from the OSI as an open source license, gives entities enough of a foothold or legal precedence or holding that you can safely, in his opinion, use that and not be too worried about the re-hosting scenario. Now, he detailed some of that in that show. I don't know all of his reasoning why he thinks that's the case, but it's possible that -- because SSPL and AGPL are very similar, are they not?
+
+**Salvatore Sanfilippo:** Exactly. I believe so as well. There was, I believe, some -- I was not part of the company when this happened, but I believe that this was a consideration. But then, you know, sometimes also the legal advice that you get... I will say something a bit maybe unpopular, but I think that the legal offices in companies sometimes create the worst possible setups for companies. Because the legal offices don't really try to usually balance the risk, the legal risk with the company potential, but they are mostly going always to err on the safe part, in order to be right later. If you are super-conservative, you never do anything wrong from your point of view. However, you are paralyzing the company.
+
+Many times when I had to do contracts, I skipped the lawyers that were in between and we could not understand each other, and said "Okay, I will write the contract. This is the contract. If you want, you can fix the form", and immediately, the thing started to immediately to work, the conversation. And I believe that sometimes it can happen that choices are -- but if you read the AGPL license very, very well, you actually discover that also the form is not extremely clear. It's not very well written. So I believe that it's very hard that somebody will authorize to use it in a software as a service setup without any kind of a signed paper by the folks that have the copyright. And this could be a good -- I think that they did a good thing switching back. Basically, in this kind of changing setup, you try, and you believe "This is very similar to AGPL. People will accept it." Instead, you find the cultural resistance, and then you step back, "It makes sense. I can understand that."
+
+**Break**: \[40:09\]
+
+**Adam Stacoviak:** So you said that you guys are considering licensing updates, let's just say, to paraphrase it... Is AGPL back on the possible table? Since the SSPL and the AGPL is so similar, do you think - to be culturally fit again, let's just say - that's a wise consideration?
+
+**Salvatore Sanfilippo:** There is, I believe, a part of the company that wants to do some change in this direction. However, there is to basically find an agreement about that. But personally, for myself, I am positive about doing this switch, and I will keep the conversation going about this.
+
+**Adam Stacoviak:** Yeah, I think it'd be a win. I mean, we started off with, you know, basically 15 years ago, and your journey to create it, and how it prospered in the community. Obviously, BSD license, open source - thrived, beloved, to this day. And many mourned its license change because it limited their ability to use it as it was prior to be open source. And I get it. You've got the Goliaths and the behemoths, and you've got the challenges, and license changes are sometimes necessary to protect all of your work, and the company you built, and the customers you've gained that pay you. And then we kind of forget, in a way, this open source land, this thriving, free ecosystem that's not just free in dollars, but also free in freedoms... I can feel that tension there, obviously, but I think it'd be super-cool if you guys went back to open source. Could you imagine, Jerod?
+
+**Jerod Santo:** It'd be awesome.
+
+**Adam Stacoviak:** Redis, open source again.
+
+**Jerod Santo:** There you go.
+
+**Salvatore Sanfilippo:** Yeah. To be honest, I'm investing my time, both in creating the cultural conditions inside the company, but I don't want to stop there. Also, with the vector sets, with \[unintelligible 00:45:24.22\] I want to in some way show the company also back the Redis way from the point of view of simplicity, of also -- one important thing about the vector set data, the API compared to the other systems that do vector similarity is the trust to the user using them. Instead of exposing a system that is an index, we expose the data structure, because Redis users have shown us in the past that they are very smart, they are very capable, and they know how to use building blocks in order to shape them exactly for the kind of problem they want to fix. So this kind of thing is also part.
+
+So pressing about the license, pressing about the different vibe in the community - in general, this changes. I want to be in some way the flag of this new revolution inside the Redis company. And I see that I have many, many people that are aligned with this vision. For example, today I received the second pull request from the CEO of Redis Labs, that hacked the C code in the vector sets in order to fix a bug. So in this moment Rowan, which is the new CEO, is hacking to the code that I'm showing to my colleagues, and sending pull requests. And since I see he's really a community person, that has this kind of background, I believe that the company can be aligned towards this vision.
+
+**Jerod Santo:** Mm-hm. Was this one of the reasons for your return, or did you find this once you got back, that you wanted to make this cultural change inside? Or did you see it as a reason to come back to Redis and Redis Labs?
+
+**Salvatore Sanfilippo:** Honestly, I started to think that my comeback could be useful from the point of view of recovering what was created in so many years. And I didn't want -- you know, after I left Redis, I didn't watch again the commits, whatever. Zero. Because I said "Okay--" During my term I tried to take the direction straight, but then now it's too simple to look at the commits and saying "This is wrong." If other people are working to the system, you have to let them work.
+
+\[48:06\] But then, when I saw the conflict inside the community, I thought that since I know a lot of people inside the company, and I was sure that many of the technical staff, for example, it's not what they wanted happening, I said "Maybe I can return. And I can return also without feeling too much stress, if from the point of the code, I focus on only a subsystem each time. And from the point of the other work I do, I focus on the community." That was the idea.
+
+**Adam Stacoviak:** Let's paint a picture, because you keep saying the company, and I imagine - you've got Salvatore here, you answer as the person we've known since the beginning... And then you say "the company." Can you quantify what that means, the company? How many people? Who's involved? Is there a lot of open source lovers or purists there that really appreciated or loved the days when Redis was fully open source? Help us understand what you mean when you say "the company."
+
+**Salvatore Sanfilippo:** Okay. That's a very cool question, because I believe that Redis is one of the worst communicated companies in the world...
+
+**Adam Stacoviak:** Okay...
+
+**Salvatore Sanfilippo:** Because it's much better in the inside than it looks in the outside. Basically, we are 1,000.
+
+**Adam Stacoviak:** 1,000 people.
+
+**Salvatore Sanfilippo:** Yeah.
+
+**Jerod Santo:** That's a lot.
+
+**Adam Stacoviak:** Holy moly, that's a lot of people. Okay, the company's big.
+
+**Salvatore Sanfilippo:** A lot of people.
+
+**Adam Stacoviak:** It's not Google-big, but it's big. That's a lot of people.
+
+**Salvatore Sanfilippo:** It's a lot of people, and the development part is almost completely in Israel. A lot of remote work is inside Israel, in many places of Israel, in the North, in the small cities, small villages... And there are truly talented people there. Incredible persons. Also very loyal persons to the company, that understand the project and understand it is very important. Also, they are people that kind of fight for the AGPL thing, because they thought it was a good thing. So people interested in the project. And I believe it is a very strong pool of talented people in Tel Aviv.
+
+Also, another thing that happened that I believe is extremely cool is that we are acquiring new people that are like that. What happened is that one of my colleagues, Oran, which is truly a genius of programming, together with Yossi - they are two really superstars that nobody knows. For example, Yossi was the one that wrote the abstraction layer in order to make Redis SSL as a module. So you load a module and it is SSL, otherwise you don't have the complexity, and stuff like that. For example, this Chinese person starting to do pull requests only on comments. This comment, this -- and another guy, which is a Portuguese, for example, also doing initially simple pull requests. But he recognized that they were smart, and started to help them to create more complex pull requests. And since Oran works a lot, since he's responsible of all the developments of Redis on Flash, which is, you know, the way that we have to offload part of the data from RAM to Flash; it's a complicated fork of Redis. You know, he had to do a true investment, and now these two persons are very strong core contributors of Redis, making complicated stuff.
+
+\[52:00\] Also, another guy, which is the one that implemented the expires in the hash type elements, single elements - he's very talented, and he optimizes this thing to a so low level. And for example, one thing that I still love about Redis, and I think that we have an edge here compared to, for example, Valkey, is the design idea. In Valkey there are a bunch of people and they do a lot of agreement design, which is a way of working that I don't believe will be optimal. And why I hope that Valkey is going to be great, because in some way this was -- you know, the thing about the BSD is that whatever happens to the company, to myself, the code in some way can go forward, because it takes other streets, like Amazon, or Google forking, \[unintelligible 00:53:00.11\] However, from the point of view of the software idea, I don't think that Valkey can do a better job than us. And one thing that I would love is we go back in to AGPL to be also OSI-approved it again, and then compete on the quality of the ideas, and the quality of the developments... You know, because Redis is going to diverge a lot.
+
+For example, one of the leaders of the project said, after seeing in my blog post the vector sets thing, "We will never do something like that", because they don't trust in this line of, you know, AI is this terrible hype. For example, this is one way to approach the problem. I am, for example, an AI enthusiast, and this already creates a different take on the project.
+
+**Jerod Santo:** Sure.
+
+**Salvatore Sanfilippo:** So I believe they will diverge a lot.
+
+**Jerod Santo:** That's really interesting. So -- and you're happy about that, because now you have basically... Like, Redis gets pushed forward, but also Valkey gets pushed forward in different directions, and different things work for different people, in different circumstances, so it's kind of spurring some innovation. Did you see Redka? Is that one that crossed your radar? Because there's a bunch of forks that came out back when Redis was first announced that the license changed. Redka was not a fork, it was like a re-implementation. Are you familiar with this one, Salvatore?
+
+**Salvatore Sanfilippo:** No, no. I heard about it, but I didn't go to check what they are doing.
+
+**Jerod Santo:** You didn't check it out. This was just a re-implementation with a SQLite backend, trying to just get the Redis API up and running with a different backend. It was kind of a greenfield thing that was inspired by the re-license, but I thought it was pretty cool, the explosion of new projects and ideas that happened... Which was kind of like open source working, in a sense. Like, this is the spirit of hackers hacking, and starting new things. And it's kind of cool that all of this has inspired you to come back to Redis and like push it forward in new directions. I think that's pretty cool.
+
+**Salvatore Sanfilippo:** Yeah. It's like, all the teams together optimize to search the space of potential possibilities, basically.
+
+**Jerod Santo:** Well said, yeah. That's cool. So where do you take it from here? AI, it sounds like, huh? Redis AI. \[laughter\]
+
+**Salvatore Sanfilippo:** I think that vectors -- I mean, learned embeddings. Because vectors -- okay, the way I am implementing vector sets is one way that allows the user to also use vectors outside AI. For example, I implemented binary quantization, and if you want, you can put binary vectors where just each bit is a product that a user got in an e-commerce system, and you still can do cosine similarity, and it takes similar products to this user. But of course, the main thing is learned vectors created by models that compress something, some object, be it text, image, whatever, in an embedding that captures the semantic value of the initial object.
+
+\[56:22\] I believe this is going to be useful in the future, even if I am not a fan of rug, because I believe that RAG created a lot of issues. For example, one of the reasons I believe Claude Sonnet is often in the real world so strong for programmers is that when you attach a file there, the model sees all the file. They don't use RAG. Instead, when you attach a file in the Open AI systems, it uses RAG. So you don't see it, but the model is just seeing fragments of the code. So in some way, Claude looks magical because of that.
+
+**Jerod Santo:** So do you know what technique Claude Sonnet is using to get that information that's not RAG?
+
+**Salvatore Sanfilippo:** They just put everything in the prompt.
+
+**Jerod Santo:** Oh, they put all -- I mean, do they have a huge prompt, I guess? Big context.
+
+**Salvatore Sanfilippo:** Exactly. That's the reason why Cloud Sonnet, after you fiddle with it a little bit, says "Return in five hours."
+
+**Adam Stacoviak:** Yeah, I was gonna say...
+
+**Jerod Santo:** \[laughs\] "Return in five hours..."
+
+**Adam Stacoviak:** My experience with Claude so far has been just that. They almost -- they start to yell at you, essentially, when you have too many back and forths. And I'm like "That's the whole point, is to have the back and forths", essentially. The prompt, the response, the prompt response, and you iterate and massage and form whatever it is you're forming. And it's like "Well, that's the whole point of this chat system." And we can debate the usefulness of chats if you'd like to, but they discourage long chats, essentially.
+
+**Salvatore Sanfilippo:** I agree, but I believe that this is a feature, and the way why so many programmers speak at the Claude. Why I'm saying this - basically, like Redis, they are putting you face to face with the limitation of the system. Instead, what Open AI does is to implement two things. One is memory, which is after the context is too long, the model will summarize it in some way, or use RAG itself in the history. And so you believe that the model is still aware of all your chats, but the model is no longer seeing the full details, and they start to act erratic.
+
+**Adam Stacoviak:** They do, that's true. I experienced that as well.
+
+**Salvatore Sanfilippo:** So I am more happy \[unintelligible 00:58:39.18\] but they can understand very well what's happening. And also, DeepSeek V3 and R1 does the same, that when you attach a file, the file is completely inside the prompt. And since attention is quadratic, it burns a lot of GPU compared to using RAG.
+
+**Adam Stacoviak:** How much are you studying this? Because this is something that I'm still sort of catching up to how the GPU interacts with a model. Let's just say you've got an off the shelf one, 24 gigs of VRAM. How does that translate to the model being in the VRAM to consuming the GPU's ability to compute? How are they different? Like, memory and compute on a GPU is not one and the same. Because you may say "Well, I've got one or two GPU's and I can consume 24 to 48 gigs of VRAM", but you still have compute on that GPU. Can you explain, do you know much about that, or are you becoming more familiar with that?
+
+**Salvatore Sanfilippo:** \[59:44\] Basically, what is happening right now is that all the frontier models are MOE, mixture of experts. They are no longer like, for example, LLaMA 3.3 or the initial models, dense models. Starting with GPT-4, Open AI started with the idea of using a mixture of experts. Basically, the model is partitioned inside, like a cluster, and then there is a layer, which is the routing layer, which selects at each layer, for each token, which part of the model to activate. In this way, a subset of the model becomes expert in the English grammar, another subset in Python, and so forth. However, this division is not so clear. The model learns how to split the knowledge, and it's very mixed \[unintelligible 01:00:41.26\] inside.
+
+So basically, with MOEs you still need the VRAM to contain all the model in your RAM. So for example, if you have -- let's talk about DeepSeek version three, which is more or less as big as Claude, and we have the actual data public for DeepSeek. It's a 600 billion parameters model.
+
+**Adam Stacoviak:** That's big.
+
+**Salvatore Sanfilippo:** Yeah, it's very big. And it seems it is 8-bit for a parameter. You need 600 gigabytes just for the model. And then you have the cache for the \[unintelligible 01:01:26.25\] And for a very long context, you need more RAM. So basically, with mixture of experts you need less GPU, and you need more VRAM. So this is changing. The GPU power is becoming, in some way, for inference, at least less important than VRAM... To the point that I have a MacBook M3 Max with 128 gigabytes of memory, and I can run DeepSeek in my computer. Not the distillated versions. The actual thing, extremely quantized to 1.5 bits per -- but still, because I have enough RAM, \[unintelligible 01:02:13.13\] inference, it's able to run this beast. So basically, what will be the dream for all us will be computers with like one terabyte of RAM. And even if the GPU is not so powerful, we could run frontier models inside our machine.
+
+**Adam Stacoviak:** Let's clarify runs. Tell me tokens per second.
+
+**Salvatore Sanfilippo:** Four tokens per second in my computer right now.
+
+**Adam Stacoviak:** Which is so slow.
+
+**Salvatore Sanfilippo:** It's very slow.
+
+**Jerod Santo:** \[laughs\] But it runs, Adam. It runs.
+
+**Adam Stacoviak:** Yes, I know. I want to be clear though, because --
+
+**Jerod Santo:** 600 gigabytes, he said.
+
+**Adam Stacoviak:** Yes. No, I get it. But, I mean, you've got people who are dabbling more and more into this, and runs and is usable is potentially two different things, because... It can run, and I do agree with that. Your hypothesis that more RAM and less GPU makes sense because the Mac, the M1 systems or the M systems allow you to essentially allocate this RAM as VRAM, as what would normally be VRAM for a GPU. So you can utilize that 128 gigabyte that you have available, and maybe a less powerful GPU, and still run an inference. However, you won't have super-fast tokens. And then if you want to program a Python program with that, you're going to spend a couple of days, a serious amount of patience, and potentially a ton of errors, who knows.
+
+**Salvatore Sanfilippo:** \[01:03:52.01\] Yeah... Also, a lot of people don't realize that if you are talking in English with the model, often a single word is one token. However, because of how the tokenization work of large language models, if you talk with them in Italian, for example, since the Italian was very underrepresented in the dataset, or even if it's writing Python, and often for each parenthesis, a new line, it's a token. This means basically that you have to generate a lot more tokens for each sentence, and it becomes incredibly slow, because every two characters are one token; not one single word.
+
+**Adam Stacoviak:** It really makes you weigh your words that you put into the prompt too, right? Because the weight and the structure -- it almost kind of brings back prompt engineering in a way, because you want to, not so much engineer, but you want to be particular with how you prompt the LLM in this scenario, when you have resource-constrained systems. Whereas if you've got Claude, you just realize that "Oh, you can't have long chat histories." Or if you're on open APIs, ChatGPT 4.0 or whatever model, you can go to infinity, basically, but like you had said before, it's not real. It's faking the scenario. It's ragging it later on, which is just like retrieving stuff in fragments.
+
+**Jerod Santo:** So what you're saying, Adam, is don't waste precious tokens by thanking your models when you're done.
+
+**Adam Stacoviak:** Precisely.
+
+**Jerod Santo:** \[laughs\]
+
+**Adam Stacoviak:** I longer -- that's a deep cut, Salvatore, because I am the person that says "Thank you" or is generous with my words, let's just say... "Can we please", you know... Or "Let's do this", versus "Generate", which is --
+
+**Salvatore Sanfilippo:** Me too, but this is nothing. The trick is when you start to have a mid-longer chat, to start a new one... Because since attention is quadratic, once they go -- okay, there is the KV cache, but still, you are accumulating a too much longer chat... And instead, you start anew, fresh. Also, sometimes the model can focus much better on the latest version of the code when you start fresh. It becomes more selective, and it gets less confusing. So my trick is this one.
+
+**Adam Stacoviak:** Yeah, I like that. So you pull out the latest -- you may have like a 10-minute session, let's just say, a couple of back and forth, but then you take the final artifact you're liking from that session, sort of whole brand new session, and bring that latest version of whatever you're liking to that new session and allow it to begin again... To have that new, finalized context, versus all the different bifurcations you could have taken along the path.
+
+**Salvatore Sanfilippo:** Exactly.
+
+**Adam Stacoviak:** Okay.
+
+**Jerod Santo:** That's a pro tip right there.
+
+**Adam Stacoviak:** There you go. Prompt engineering is back.
+
+**Jerod Santo:** It's back... Until we have sufficient VRAM that we don't need these things. So Salvatore, when do you think this new Redis vector set stuff you're working on will be baked? When will people begin using it? It's out there, it's somewhat open source? What's the status?
+
+**Salvatore Sanfilippo:** I think it will be released in one or two months, something like that.
+
+**Jerod Santo:** Cool. And what are some use cases or some people that are currently using Redis that they would be like -- what would this unlock for certain folks?
+
+**Salvatore Sanfilippo:** Okay, an example... Sometimes when I take weight, I want to lose this weight. Normally, I do calorie jacking with my \[unintelligible 01:07:27.17\] or other applications, and stuff like that. However, it's so boring. And I do that for 10 years at this point, but still it's boring. So I'm writing to a Telegram bot where I just say "I am eating 10 grams of honey, and one slice of bread of 40 grams", and it will do it for me. However, you cannot trust the model to have accurate informations about the calories of each kind of food. So basically, what I did was to create a vector set inside Redis, where it has each line of a very big database where there is each food with the macronutrients, the calories, and whatever.
+
+So what the bot does is that based on my query, it asks Redis for similar items. Then I take the similar items of foods that are similar, and it takes that in the context of the LLM when asking the question. "The user reported eating this thing. Based on this table, don't invent the macronutrients. Please tell me what are the total calories and proteins and carbohydrates and fats." This is one use case that works very well.
+
+**Adam Stacoviak:** Key phrase there, or key word there, Jerod, was please.
+
+**Jerod Santo:** He did say please.
+
+**Adam Stacoviak:** He did say please. \[laughter\] The debate is still out.
+
+**Jerod Santo:** It's only one token in English, you know? I don't know about in Italian, but... It's only one token in English. Okay, that's cool. That's very cool. I'm sure there'll be countless other people with ideas on how they can leverage this to do cool stuff, whether it's commercially, or healthily in their own time, to track their calories. I think it's pretty sweet.
+
+**Break**: \[01:09:24.24\]
+
+**Adam Stacoviak:** I don't know if this maps to this, but there's a particular problem I'm trying to solve for me. So as you may know, Salvatore, we are a sponsored podcast. And so we have brands, like Redis - which have never sponsored us - that we would happily partner and share with the world, the developer world, what you're doing, let's just say. And our ability to win these relationships is really predicated on my ability to be fast with accurately quoting people what we can do for them. For the most part, we have packages, but sometimes people are asking for this or for that, and so it's not always just "Here's package one and here's package two." And so what I've done recently is I've gone into Open AI's ChatGPT and I've created my own GPT. And it sounds like what you're talking about with this vector embeddings is like "Give me truths, particulars." Like "This ad spot costs this much. And when you multiply it by 12, it costs this much." That will never change, because it's like a row in a database. The value is X, and so quantify it by Y, for example. However, in my experience with this, even though I've made this GPT and I've told it how I want to format the response, and I just say "Generate it", it's driving me crazy, because it's doing the math wrong... I so bad just want to spit on my own Ruby app just to like generate -- like, this basic text is all I really need. But I want to prompt it to say "Give me 12 weeks of the Changelog, give me 12 weeks of this, and then generate this thing I would normally handwrite", which I already did, I formatted it, "but just do all the word math for me, do the word calculator stuff for me. Because I'm lazy, and I can be slow, and I want you to be fast, and I want it to be accurate every single time."
+
+So I feel like what you're talking about with this vector embeddings is like "Okay, in that world, building my own GPT or my own thing like that is possible because you've got these embeddings in there", and whenever I prompt the LLM, it's going back to those embeddings and saying, "Okay, I know this costs Y, and we're multiplying it." And the word calculator is my prompt, essentially, to say, "Give me this", and it gives me all my words back, but all the math is done. Does that map onto what you're talking about with -- is that an application of vector embeddings?
+
+**Salvatore Sanfilippo:** Yeah, yeah. In general, if you want to retrieve anything -- for example, another example that... For your use cases particularly, one of those systems that have a very large context window, like Claude Sonnet, and that don't use RAG and see all the files in the context could be the best spot for you. If you want to try, you can take some document, put all the information needed, and then you do few-shot learning. So you make examples. Okay? And then if you upload this file to Claude and you ask a question, it should be able to reply in a perfect way.
+
+**Adam Stacoviak:** I want that.
+
+**Jerod Santo:** \[laughs\] "I want that so bad."
+
+**Adam Stacoviak:** The math being off is just like driving me crazy. I just did it today, I was like, it's accurate 95% of the time, and then one time it's like, it does terrible math. I'm like "I could never charge our customer that or our partner that, because that's not the real number." And so I've got to check it every single time. I'm like, "I just want the word calculator part of it."
+
+**Salvatore Sanfilippo:** Another thing that you can do is to specify to the model that you want all the calculations performed using a program. This also is a trick that works.
+
+**Adam Stacoviak:** Okay. Interesting.
+
+**Salvatore Sanfilippo:** \[01:15:43.16\] And Claude unfortunately cannot run Python, but it writes JavaScript and executes it inside your window of your browser, and then takes the calculation back. But to make another example of vector sets, imagine that you want to do something like face ID. So basically, the iPhone face ID works with an embedding. The dot matrix in your face is provided to a given model that outputs a fixed vector. And imagine an evil government that wants to track you in every second, and it has a database of all the faces in a vector set, and then when you scan yourself, it locates that it's you, for example, or you are similar. Or if you have doors, automatic doors that look at your face and open the door. So it's not just text embeddings. You can use it for every application when you can turn an object into a set of features, and you want similar items. So it's possible to use a lot -- and as I said also, it works very well, for example, even outside the AI.
+
+A few years ago there was one guy that did this in Hacker News. It downloaded all the data, all the comments of all the users, and created vectors with the most used 10,000 words, setting the number of occurrences for each word. So in a given vector, there were many zeros, but other things valorized. And using this and cosine similarity, he was able to extract all the fake accounts that were actually cloned. You know, when you write with a throwaway account, because the cosine similarity is very similar. So also, you can use this in order to spot similar texts, and other stuff like that. There are many, many use cases. Basically, I have a very active YouTube channel right now, talking about AI, mostly. And when vector sets will be released, I will write a lot of toy applications showing use cases.
+
+**Jerod Santo:** Cool. What's your YouTube channel called?
+
+**Salvatore Sanfilippo:** Salvatore Sanfilippo.
+
+**Jerod Santo:** \[laughter\] Keep it simple.
+
+**Salvatore Sanfilippo:** The coding videos are in English, a lot of talking videos are in Italian, but what I do is to take the transcript, pass it to Claude Sonnet, and it cleans up the Italian, and then the auto-translate to English works very well. There are other people that don't talk Italian following me in the channel.
+
+**Adam Stacoviak:** So it sounds like you're a Claude Sonnet fan over Open AI's models. Like, you're hanging out in there.
+
+**Salvatore Sanfilippo:** Definitely.
+
+**Adam Stacoviak:** Okay. And how do you get around this limitation of what we talked about before when you volley back and forth? I don't know, you have a long chat history. How do you get around that? You just restart it? That's your way?
+
+**Salvatore Sanfilippo:** Yeah, in general, I use AI a lot for coding, but really big time, but in a strange way. Most of the time I am in my terminal, writing my code myself. When I've got a problem, I go there and often my prompt is like "Don't write any code. Let's discuss about this", because I use it like an improved \[unintelligible 01:19:23.13\] in order to create more ideas, or explore mathematical concepts that are at the edge of my mathematical skills, or stuff like that. For example -- or if I have to do some complicated optimization and there are tricks that I don't know, or stuff like that. So basically, I'm very conservative. I avoid, basically, the brute-force way of coding, of continuing trying to refine... Because I discovered that this often leads to losing too much time to get not great results. So either I spend a lot of time writing the prompt in a very subtle way in order to influence the design ideas of the model, giving a lot of hints, but positive hints, and also "Don't use that, because I already understand it's not going to work very well." So my chats are usually not super-long. And otherwise, I start from scratch.
+
+**Adam Stacoviak:** \[01:20:26.17\] Gotcha.
+
+**Jerod Santo:** So are you using this outside of your editor, or inside of your editor?
+
+**Salvatore Sanfilippo:** Not integrated, because for me this is a way to avoid becoming lazy, basically. I want to be sure that 100% of the code that I use, even if some code is written by the LLM, I understand what's happening. But I use it a lot more to track bugs. Like, I have a bug, oftentimes Claude Sonnet, if I post the code, "What's happening", it can debug it immediately.
+
+**Adam Stacoviak:** So good. Yeah.
+
+**Salvatore Sanfilippo:** Incredible. Like, five hours of debugging, five minutes. Because it will say -- after refactoring a few days ago, I forgot a statement; I removed the statement, but the statement removal didn't cause any error in the compilation, but it just created a memory error. And what they do in this case, which is -- I see that many people don't do that. You have to send the file and the commit diff, so that it can see the difference. And then it becomes extremely smart in understanding what you messed up.
+
+**Jerod Santo:** Well, I'm going to have to try Claude Sonnet now. I've never tried it. I've been happily doing the other things. I don't know if "happily" is the right word, but I've been doing other things, and that's one that I just happily ignored so far. But you've convinced me. I need to give this one a shot. Maybe a few shots.
+
+**Adam Stacoviak:** I can toot it a little bit, the horn of Claude a little bit. So I saw somewhere - I think it was on YouTube, or something like that, someone was saying it's really great for building documents. So I needed to create an agreement... And I knew what I wanted to say, but the way Claude works is it will have a file created, and it's open on the side; let's say you have your chat thread, and then on the side the document is open, and it will just update parts of the document. So from a user experience standpoint, you see not a regeneration of the whole document, that you now have to be like "Well, what changed? What did not change?", and you've got to scan it, and you waste your time... But you literally see Claude pull the document out that you're iterating towards, you see the little by little change, in a clause, or a heading, or the very specifics. And as a user, I'm so much more comfortable with the process, because I'm like, I'm working it through, and it's generating what I want it to, and it's not like generating this legalese kind of stuff, but it's only updating the parts that I want. Like, let's focus on this clause. I want it to cover this, this, this, and this, and here's what I want it to say, and it goes and just updates that one part. So from a user standpoint, it's great.
+
+**Salvatore Sanfilippo:** You see it removing the lines, like pressing the cancel key, because it uses the artifact as a tool.
+
+**Adam Stacoviak:** Yeah.
+
+**Salvatore Sanfilippo:** So it can remove, edit, change parts... And also, between Claude 3.5 and 3.7, the generation output length was increased like 10 times. So now it can generate like 1,000 lines of code in one shot or even more, which is pretty impressive.
+
+**Adam Stacoviak:** Wow. That is impressive.
+
+**Salvatore Sanfilippo:** It's a very great system. Also, when Dario Amodei started the company, they were also extremely focused on the ethical part, the safety part, which is also a plus... Even if I believe that what they released a couple of days ago, Claude Code, which is agentic coding, I believe this will start to ignite... It's the right word?
+
+**Adam Stacoviak:** \[01:24:12.24\] Ignite, yeah.
+
+**Salvatore Sanfilippo:** A lot of firing or mis-hiring of programmers. Because it's --
+
+**Jerod Santo:** It's that good?
+
+**Salvatore Sanfilippo:** Yeah, it's -- Claude Code is the first thing I see which is not conceived to empower the programmer that wants to do more, but to replace.
+
+**Jerod Santo:** Ahh...!
+
+**Adam Stacoviak:** So much for ethics... I'm just kidding. \[laughter\]
+
+**Salvatore Sanfilippo:** I understand that it's inevitable that sooner or later it's going to happen, but --
+
+**Adam Stacoviak:** Someone's gonna do it, yeah.
+
+**Salvatore Sanfilippo:** ...what is my interest is that if we have to go forward like that, that governments in all the world understand that they have to find ways for the people to pay their bills. And then it's okay if machines want to code; as soon as people are safe, I am okay with it.
+
+**Jerod Santo:** Well, on that demure note, anything that we haven't talked about that doesn't include the career extinction of us software developers...? \[laughter\] Anything else you want to mention, Salvatore, before we let you go?
+
+**Salvatore Sanfilippo:** Just that -- I see that, in some way, even if we say that our field in some way in these 15 years changed. However, imagine something as big as AI in the hands of other industries, for example pharmaceutical industry or finance. They will never -- it will never be like that that we have most of these tools for free. We have frontier models under the open source, a lot of other models and all the informations available. So I believe that still for IT there is hope, because we are in the environment that created the most incredible revolution is creating after like the industrial revolution, and still there is this idea of sharing, of making the tools accessible.
+
+Also, if you think at the price of Open AI and Anthropic right now, $20 per month is very low. It's very democratizing this tool. So in some way, I continue to be optimistic. I am more about the \[unintelligible 01:26:40.27\] side, which I believe that in the short run the problem that we will face with AI - as a society, we will be able to overcome them, but I will focus on the existential threat... Because the agentic nature of these models, as they become stronger and stronger, even if super-intelligence can be achieved without consciousness, still an agent that breaks all the computers around... This is the worrying part for me, that there is the -- we will survive social change, but we may not survive some extreme event due to AI.
+
+**Jerod Santo:** Well, we're excited that you're back at Redis... \[laughter\] We're excited for the subtle changes, the incremental changes that you're making to help Redis be more useful to more people... Thank you once again for creating such an awesome system and for open sourcing it, so that the world can take advantage and benefit. I know I've benefited greatly from it in my career, even though I only used it for a relatively short time. I think the API and the way it works and the way you designed it has influenced my thinking as a programmer... You can't go back from that. That's like fundamental change... So I appreciate you for that reason.
+
+**Salvatore Sanfilippo:** Thanks.
+
+**Adam Stacoviak:** Thank you, Salvatore.
+
+**Salvatore Sanfilippo:** Thank you. It was very great to talk with you, and thanks for the hospitality.
+
+**Adam Stacoviak:** Oh, yes. Glad to have you back.
+
+**Jerod Santo:** Yeah
+
+**Outro**: \[01:28:17.29\]
+
+**Jerod Santo:** So you recently wrote a post called, "We are destroying software", which - you know that, but I'm telling you that as a setup... Which was almost a poetic piece of maybe a little bit of a rant, with like 19 things that we're doing, roughly... I didn't count one by one. But ways that we are destroying software, for instance with complex build systems, with absurd chains of dependencies, with rewrites of things that already work. I'm curious what were you doing the half hour before you decided to write that thing? Like, what happened in your life that's like "You know what...?" Because you probably thought this way for a while, but finally you're like "I'm writing it." What inspired this?
+
+**Salvatore Sanfilippo:** In general, everybody I look, I see that the complexity is increasing without a proportional utility, usefulness of the systems. For example, one of the most important things that we have right now, which are LLMs, are like 3,000 lines of code for the inference, and even if you look at DeepSeek today, they released all the training code, the distributed training of mixture of experts models, and it's a simple code. So why the Facebook Android application has 1 million lines of code? They are rewriting it from Java to Kotlin, and it's 1 million lines. And everything is like that. We have the same forms, the same buttons, the same likes of 10 years ago... However, if I check every tab -- now the recent Chromium tabs show you the memory usage, and the Claude tab, the chat is 419 megabytes, just a single tab. And everything is like that. And then also, software developers can --

Autonomous drone delivery in a Zip (Interview)_transcript.txt

@@ -0,0 +1,479 @@
+**Jerod Santo:** Today, we're joined by Keenan Wyrobek, founder and CTO of Zipline, a company I'm very, very interested in. Welcome to the show, Keenan.
+
+**Keenan Wyrobek:** It's a pleasure to be here.
+
+**Jerod Santo:** Pleasure to have you. Autonomous flying machines, all kinds of deliveries happening... I saw Zipline and I thought, "You know what, this one makes sense." There's a lot of people trying to do delivery... Of course, you have autonomous cars upcoming, you have people that are doing tunneling, they're trying to put tunnels in the ground and ship it... Kind of Futurama style, but underground Futurama... And then we have drones, which - I don't think Zipline's the only one playing the drone game, but you guys are, I think, way ahead of the game. Been doing it for a long time. Tell us about the start of this company and where you're coming from.
+
+**Keenan Wyrobek:** Yeah. So we started almost 12 years ago now, and we started in a niche of a niche, of delivering blood to hospitals in Rwanda. That's where we started.
+
+**Jerod Santo:** Why there?
+
+**Keenan Wyrobek:** Yeah... So really, Rwanda earned this for -- we knew we wanted to start in health. At the very beginning of this we started exploring this space based on really just stories from family and friends. My wife's an epidemiologist, and she would tell me these stories about health intervention campaigns that would get stuck on logistics. These would be lik vaccine campaigns where they had the vaccines, they had the doctors, but then the logistics got in the way, and they couldn't be successful. So yeah, that's how we started. Went deep on that... I'm personally not an early adopter, I'm very tech skeptic... So I went deep in Central America and Africa at the time expecting to come away with a thousand reasons why we couldn't possibly make a difference. And the opposite happened. The deeper we went, the more conviction we got that we could make a big difference. And one of those places we were exploring with was Rwanda, and they turned out to be a phenomenal first customer for us, and really a partner figuring this out together.
+
+And blood, of course... Why blood? It's just this rare commodity everywhere. There's never enough of it... It's actually very expensive. We think of it as free, because we donate it, but the cost of the collection, the testing, the transportation is very expensive. Hundreds of dollars a unit; even in places like Rwanda, and way more in places like here. It has a very short shelf life, as little as seven days from the time it's donated, to the time it expires if you don't give it to somebody who needs it. So by centralizing blood in one place and then delivering it when you know what blood type is needed for what patient, at what time, it turns out you can virtually eliminate blood waste, saving a ton of money and a ton of lives for the health system. That's why we started there.
+
+**Adam Stacoviak:** What was the tactical nature of that when you went to Rwanda? Was it discovery first, and then tech later? How did you come to a technical solution? How did you think about it before you built something?
+
+**Keenan Wyrobek:** Yeah. I mean, we had built nothing. I literally called a friend of mine who was good at drawing, and he drew pencil sketches, and made a slide deck of the concept... And we were using that to talk to people. In the early days, it was just talking to the operators in these health systems, that actually run the logistics, run this testing, run these labs... But very quickly, we ended up having meetings with the offices of presidents in these countries, and that was a big sign that we were looking at a problem that was a really big deal for these health systems. And yeah, no technology whatsoever. A year or two before we were doing this, Amazon was like "Hey, drone delivery coming next year." So that drone delivery was an idea that was out there. It wasn't a real thing yet anywhere... And so we kind of thought of ourselves as like "Hey, if we're going to do something, let's find a niche where the value is really high to do this." Yeah, no technology. And then of course, once the interest was there, we got moving fast.
+
+**Jerod Santo:** Was there a point that you and your co-founders looked at each other -- you had saved countless lives with delivering blood around Africa, and thought "Burritos. Burritos are next." Or was that from the start?
+
+**Keenan Wyrobek:** It was both there and not there. In a lot of the countries we operate around the world, we start with health systems. That's how we started in the United States. We started with health deliveries during COVID, and then expanded from there.
+
+In the islands of Southeastern Japan we deliver healthcare supplies and bento boxes. In a lot of places in Africa we operate we started in human health and then expanded into various agricultural use cases, things like genetic diversity for milk cows; huge impact, but super-obscure. We've just been layering on other use cases, and slowly worked our way into things like e-commerce and burritos.
+
+\[07:55\] What's compelling -- when you think about Zipline, we talk a lot about the impact on health, which is really exciting. We started with the impact of blood. It's a super-visceral, high-impact, but very rare problem. Most of us don't need blood transfusions in our life, but when you need it, you really need it. Then we slowly worked our way in health to preventative care, which is actually where the real cost savings for health systems is. Then we've worked our way into other areas, like animal health.
+
+One of the things that's really magical about this, as we step into new things like auto parts delivery - you know, what limits the mechanic's ability to service cars is usually how fast they can get the parts. The way I think about it is basically for life, in the home or professionally, we're in the mission of getting you what you need, when you need it, and we do it in a way that's not just fast - we skip over the traffic; we're very reliable in that way - but just wildly environmentally friendly. We don't talk a lot about that, and we can get into that if that's interesting. We're talking about a 20x improvement in the overall environmental footprint of the supply chain, and almost every aspect, including things like bird safety. That's something we don't talk about a lot. There's just so many layers to this.
+
+**Adam Stacoviak:** Well, even -- I live here in Texas, and we have a lot of wind farms, because here in Texas we have multiple ways... Probably like Nebraska, Jerod - you've got multiple ways you're getting electricity. I'm not so much a fan of them; they're big and crazy, but I hear people get upset with the bird issue, essentially of that...
+
+**Jerod Santo:** Oh, they kill lots of birds.
+
+**Adam Stacoviak:** And maybe to the environmental aspect of it, that there's still so much of a cost to produce even wind energy that it's almost not worth it when you compare to how you talk about your efficiencies.
+
+**Keenan Wyrobek:** Yeah. I care a lot about birds, so I pay a lot of attention to it... We estimate that when we displace delivery by car, we reduce bird injuries by something like -- somewhere between 10x and 100x lower rate. We study that a lot. There's a lot of aspects to why we do what we do, and why we're excited about scaling what we do into all these different verticals.
+
+**Jerod Santo:** Let's talk about the design of the drone itself, because we mentioned the tactical approach coming out of the need of delivering these things in Rwanda... When I saw your guys' design, I was very surprised by it. There's kind of a more typical quadcopter, larger quadcopter-looking thing; almost like an airplane with blades on it. And then it doesn't actually do the delivery. It delivers the delivery mechanism, which is a smaller kind of a baby-copter that comes down on a zip... That's, I presume, where you guys get the name. It lowers it down to the ground, and then raises it back up again. Where did that design come from?
+
+**Keenan Wyrobek:** Yeah. That's sort of a double drone; a drone inside of a drone design. It came from just an obsession of figuring out what would actually work for our customers. Our first platform, that does long range delivery, has a lot of great attributes, but it requires about two parking spots worth of space to deliver in, because it literally just flies over, drops the package with a little paper parachute on it, and the package floats to the ground. But so many of our customers in the health space and otherwise were like "We want home delivery. We want in metro delivery", where you don't have that kind of space. And at the time, we're kind of like "Well, we don't do that." Eventually, they asked us enough times si we started thinking about "How would you do this?" And we knew from that platform that one of the things that makes people love our service is you don't hear our drones. And that's way harder than it sounds. And we're like "Well, we don't know how to do a super-precision delivery without hearing the drone." And we've seen other ways of doing this. It's usually loud, and noisy, and kicks up a dust storm, and can't deliver very precisely. So this two-part architecture enables a bunch of things. One thing it enables is quiet. It lets the drone itself stay up at 100 meters, at 300 plus feet up in the air, which is one of the many things we do to make that just silent in even suburban places. Then that little mini drone that comes to the ground - we call it the delivery zip. It comes to the ground, it actually delivers the package for you. It has its own little propellers on it, so in windy days and stuff it can still be super-precise and get into a tight space.
+
+\[12:14\] So really, those two things - helping us be really quiet, and be super-precise; we can get into a really tight space with that little drone that comes to the ground, while the big drone stays up high, that does the heavy lifting. It's literally there to carry the weight, the distance. And yeah, that's why we have the little drone inside of a drone solution.
+
+**Adam Stacoviak:** It reminds me of how you see things happen in space, when you see a little jettison engine pushing it, nudging it... I'm a sci-fi guy, so I see these films all the time, and I'm like --
+
+**Keenan Wyrobek:** Exactly.
+
+**Adam Stacoviak:** It's kind of like little slight adjustments, but you just have -- is it one single large propeller, or is it multiple on the, I guess... What do you call the mini drone? What is that name for that guy?
+
+**Keenan Wyrobek:** Yeah, internally we call it the droid delivery zip. We don't have a great name for it. I like delivery zip, but... Anyway, this little thing that comes to the ground - yeah, it's actually got three little thrusters, very similar in concept to what you're saying. It's got one big thruster in the back, and that thruster's job is to fight big, heavy winds; and so point it at the wind and it just makes sure it doesn't get blown off course. And then it's got a little bow thruster and a stern thruster, if you like boats, to think of it that way, that help it basically stay oriented, and do the side-to-side adjustments, as it's coming down and staying precise. A lot of our team comes from space, and the control theory behind how you stay precise on that little delivery zip as it comes to the ground is very similar to how you think about reaction control thrusters on a spacecraft.
+
+**Jerod Santo:** Because that thing can't spill your coffee. I mean, sure, it can flip a burrito, no problem, but if you're lowering a coffee down - it has to stay oriented very well.
+
+**Keenan Wyrobek:** Oh, absolutely.
+
+**Keenan Wyrobek:** I assume - blood, they probably want that stuff spinning, or something, but there are certain fragile items that can't go... So on the software side then --
+
+**Adam Stacoviak:** "Certain fragile items..." \[laughs\]
+
+**Jerod Santo:** Like coffee.
+
+**Adam Stacoviak:** Coffee's fragile.
+
+**Jerod Santo:** It is. On the software side, are those two divergent operating systems then? Because it seems like they have perhaps different concerns, but they probably have a lot of overlap as well. How did you tackle the software side of these two different drones?
+
+**Keenan Wyrobek:** Oh yeah, they're very different control problems. Yeah, the drone that's up high - you're right, it has a wing, so to fly a fixed wing on the wing really efficiently it can hover in place for takeoff/landing, and while it's doing deliveries. Yeah, it's a long story, the controls approach, and really it comes down to just massive amounts of testing. Our high volume test sites have done hundreds of thousands of deliveries. We chase weather with a mobile test rig, so we go chasing like hail, and icing conditions, and these kinds of things. A ton of simulation... Literally, half of all the engineering we do at Zipline is test. Just full on half of it. That's how we can do this, is basically building the test systems, the various software platforms for simulation and things like that, the actual test scenarios, all the ground testing we do... We have an entire building down the street from where I am, that all it is is just hundreds of ground test systems to simulate what we do in the air, on the ground. Yeah, the answer to your question is a long one, but it's literally half of what we do is all the testing it takes to actually develop the control systems for these things to know that they're going to be both capable, but also safe.
+
+**Adam Stacoviak:** I also think about updating the software... I mean, you mentioned the test facilities, so I imagine you do a lot of iteration there; you're constantly deploying potentially over the air updates... I have no idea how you accomplish that mission, but... Divergent operating systems, and also a truly distributed update platform. How do you get new to the individualized? Is it an operating system delivery? Is it an individualized delivery? Can you talk about that at all?
+
+**Keenan Wyrobek:** \[15:55\] Yeah, absolutely. When we think about software and software deployment, there's two worlds for us. One we call flight software, and then cloud software. Flight software is all the pieces, a few parts of the cloud, but mostly things on the actual aircraft that we consider flight-critical. That software release process takes us about six weeks. So every six weeks we release software to that system. And just in those six weeks we do tens of thousands of flight tests, we do a bunch of hardware in the loop testing... Think of it as like the matrix, where you take the electronics of the aircraft and basically plug it into a simulation, literally, and fool it. It thinks it's flying over in Dallas somewhere, but it's actually over in our basement, plugged into the simulation.
+
+**Adam Stacoviak:** That's cool.
+
+**Keenan Wyrobek:** Plus, of course, software in the loop simulation, where you're doing millions of flight tests. And yeah, that whole journey takes us about six weeks to get to the point where we're like "Yup, this new software release is good to go." Then we release it, and then we have an over the air update system that updates all the processes, the software on the aircraft, and deploys. People often ask, "Do we do that in the air or not?" It's like ""No, no, no. When our aircrafts are docked, they update themselves." They check themselves, they're like "Okay, good. Software's good to go", and then they'll fly again.
+
+**Adam Stacoviak:** How many of these problems are novel to you? I know that the defense systems that we have here in the US probably have these problems and maybe have solved a lot of this, but maybe you don't have access to the DOD's tech platform, or maybe lack thereof... How much of this is invented here and how much of this is maybe a partner or vendor that you bring in to support over the air updates, for example?
+
+**Keenan Wyrobek:** Way more of this is done in-house than I would have ever thought we would have to do. Actually, when we started Zipline, I was like "Great, I'm going to go find somebody who has a drone, buy it from them, modify it a little bit for deliveries and start serving our first customer." And the best quote I got back then from the companies making drones for defense was a $200,000 drone with a 200-flight warranty if I did not fly it in the rain. And to be clear, in most of these countries it rains pretty much every day. So that wasn't going to work.
+
+Yeah, we do a lot in-house... And a lot of that's because we fly in conditions that no one else flies in. Stormy weather, near the ground, or over mountains, stuff like that - no one flies in that. So we've had to learn a lot of this stuff ourselves, and build our own datasets and experience and simulation approaches to actually develop these things. But then the other side of what makes us very unique is just the scale. We operate at a very large scale. Over the air updates is a good example, where we need that to be very, very easy and robust, in a way where if you're operating a smaller scale, a human can do a lot of double checking of that, and things like that, and it doesn't matter too much. Even our over-the-air update system we've built ourselves.
+
+**Jerod Santo:** Do you know how many drones are in your fleet roughly, worldwide, to give us an idea of the scale you're currently at?
+
+**Keenan Wyrobek:** Yes, it's a couple of hundred.
+
+**Jerod Santo:** Okay. And you've just started in the United States recently. 2021, I think, was the first non-medical flights... This is what I've been told by Ian behind the scenes. And now you are in Dallas, and rolling out -- I just saw a video from your CEO and co-founder, that there's a whole bunch of drones getting built right now to just go crazy. But 200 worldwide. And how many of those are over there in Rwanda, doing their thing?
+
+**Keenan Wyrobek:** I think it's close to 400 or so. It's about 300 outside of our first generation, and well over 100, but I don't think 200 yet, of our precision delivery, the drone-in-the-drone system.
+
+**Jerod Santo:** Gotcha. So how much of your software is orchestration? I assume that's the cloud side that you referenced, the control of the cloud.
+
+**Keenan Wyrobek:** Yeah.
+
+**Jerod Santo:** I guess maybe tell us more about the cloud side and what all that entails.
+
+**Keenan Wyrobek:** \[19:48\] Oh, there's a lot to it. There's obviously the pieces for people to place their orders... Every partner we have, we have basically systems to integrate with our partner's data systems. Then we have our basically dispatch... I literally think of this as like there's cloud autonomy, or fleet level autonomy, and then aircraft level autonomy. At the fleet level, you've got to decide which aircraft is going to go on which mission, at which time, and some of that is based on which aircraft is capable of that mission. Some of that is based on making sure if we told someone "Hey, we'll deliver to you in this 60-second window", that we actually dispatch the right aircraft at the right time, get the order ready at the right time so that we'll hit that window.
+
+And some of it is stuff that -- other cloud stuff for us is weather forecasting. We do our own weather forecasting. We can talk about that, it's a really fun problem. Also, basically, we have to design our own highways. If you think about autonomous vehicles, you take for granted that if you want an autonomous vehicle to drive from here to over the mountains, if you use the road network that's already been designed, you'll get there. There's no such thing for drones, and so we have to do a lot of computation in the cloud to basically give enough prior to our aircraft, so that when they're flying, they have the equivalent of a highway system. It's not quite that simple... So that when they're doing an avoidance maneuver, for example, and planning that live, that they're doing that with enough prior that they're not going to get off track and not be able to make their delivery, if that makes sense.
+
+**Jerod Santo:** And they generally fly at 300 meters, is that what you said?
+
+**Keenan Wyrobek:** They fly at about 100 meters, so 300 feet.
+
+**Jerod Santo:** 100 meters, 300 feet. Thank you. What kind of stuff do you have to avoid? Obviously, birds, but at that -- I mean, cell towers, airstrips, perhaps? What's the kind of stuff that has to be avoided at that height?
+
+**Keenan Wyrobek:** Yeah, for sure. So yeah, you can have very tall cranes, cell towers. You can have inaccurate maps, so you may not be at the height you think you are, so there's a hill or a mountain there... Yeah, definitely other aircraft. That's a big part of what we avoid. There's not a ton of passenger aircraft at that altitude, but there are some, and it's really important that we avoid them as well. Those are the kinds of things we avoid.
+
+**Break**: \[21:57\]
+
+**Jerod Santo:** How much does regulation play into this rollout? I assume every area probably has different rules about how you can go about flying your drones at 300 feet...
+
+**Keenan Wyrobek:** Yeah, I wish everybody had rules.
+
+**Jerod Santo:** Really? \[laughs\] Usually, people say they want less rules when they're trying to scale up.
+
+**Keenan Wyrobek:** Yeah... I mean, when you're a startup and you go to a regulator and you're like "Cool, I would like to do this thing", and they're like "Well, that's not allowed", and then basically from there it's like all by exception. And so you have to kind of work it out, the rules by which you operate at the detail...
+
+**Jerod Santo:** I see.
+
+**Keenan Wyrobek:** And as a startup, that's a slow process, and startups need to go fast to survive... And yeah, that's why I said I wish there was some precedent. In some ways, obviously, there's some luxury in getting to help create the precedent... But yeah, it's a lot of working with the regulators. Some things about airspaces between countries are very similar, harmonized, but other things are very different. And so yeah, things we have to do here in the US have been very different than Rwanda.
+
+Yeah, it's a big part of what we do, but it's also... We've gone from that being a big challenge maybe 10 years ago, to now it's something that we've gotten quite good at, and quite good at working with these partners and these regulators as partners, and working with them proactively enough that we can generally work out what needs to be worked out before we need it to serve a customer. But it takes a lot of effort, I will say that.
+
+**Jerod Santo:** Yeah, I was thinking of the luxury of not having rules that are already defined, because those can slow you down. But actually, not having rules and not having a clear path towards engagement - that means you have to trailblaze. And to trailblaze, you've got to get out a machete and hack away a bunch of stuff, and that can be even slower and more laborious than if somebody else had already cleared the path for you. So that makes sense.
+
+**Keenan Wyrobek:** Yeah. There's a great story -- here in the United States, starting about six years ago, we started working on the sensor system on the aircraft that can sense and avoid other aircraft. This is a system that uses machine learning, and machine learning is a class of software that the regulators have no experience with, literally. And it took us about five years of working with them to go from the beginning of this to them approving us to operate. And that was the first time they've ever approved AI/ML-based software for safety in the airspace. And so it was near daily conversations with their teams, figuring that out over that five-year journey... And there's some people at the FAA and at Zipline who I respect a ton for working through the minutiae they had to work through to figure out how what would the requirements be, and how would you validate the system against those requirements. And then actually doing all that work and all that analysis to bring it back to the FAA and be like "Okay, we did what we agreed to. Let's go through this together." And yeah, that was a five-year journey there, and it was a hard won battle. When I say battle, I don't mean like Zipline versus the FAA. It was very -- we were very much in it together, because they definitely wanted to figure it out, too. Yeah, it was something else.
+
+**Adam Stacoviak:** \[27:50\] This is an example where the word "regulation" - which you haven't quite said yet - gets really weird. And I don't want to go into politics, but it seems to be the word that gets thrown around. Regulation, less regulation, more regulation... But as a developer, I really lean back on what you said there was not a presence of, which was specifications, which is really what regulation is. It's an adherence to a specification that everyone agrees on, for obvious reasons. But you're in this kind of unique high stakes scenario where maybe regulation could be less pressured, so to speak, and in your case, you've got to be a high stakes scenario where if there's a crash, or if there's a break a leg situation, which - I'm not sure you like the word crash. I wouldn't like the word crash if I was in your business, personally, but I just think about the need for that. Can you speak to how that's played as an ally versus a foe?
+
+**Keenan Wyrobek:** Yeah... I think there's a couple of layers for how we think about this. The first one is because everywhere we've been, everywhere we've gone, there hasn't been some, like you said, specification to follow, we've taken the mantra that our first job is to convince ourselves. And not in like a handwavy way. It's like "No, no. This is over our house, over our kids. Are we convinced that we've taken a rigorous approach before we suggest that approach to a regulator?" And that's gone a long way. That mindset is attracting some of the best talent, as figuring some of this stuff out is really complicated. It's trying to track the talent we need to actually figure this out, and then it's also forced us to do the work -- to convince ourselves, the analysis, all the testing, all that data review to convince ourself is also what the regulators want to see. And so if we've done that to ourselves and we've really been skeptical internally first, it helps a ton when you're then having these in-depth conversations with these folks on the regulatory side. They have the completely thankless job of "Okay, we've got to figure out the specification for this new thing that no one's done before." So if they underdo it and someone gets hurt, they look really bad. And of course, if they overdo the regulation, then everybody in the industry complains of "Oh, they wrote down these regulations that are completely impractical." So they're stuck between a rock and a hard place.
+
+And so by doing that legwork upfront and internally proving to ourselves something that we really believe in first, and then bringing it to them - that's been core from our perspective. And then the only other piece of it that I would say is it's all relationships. These are all people on the regulatory side, they've got a job to do, we've got a job to do, and so we spend a lot of time basically working together to make sure we understand what they're trying to do, they spend a lot of time working to understand what we're trying to do... Be really open-minded, make sure we're hearing each other out, like in any relationship, and try to make progress every day, even if it's not as fast as you might like.
+
+**Adam Stacoviak:** How would you frame -- this isn't an exact one-to-one when I say this, but how would you frame test coverage in terms of a coverage level, confidence level? When you ship something and it's in a regulation standpoint, or you have a specification you've got to adhere to - is there like 95%, 98%? Like, what do you strive towards when it comes to adherence to a spec, or defining the spec? What is confidence to you and your team?
+
+**Keenan Wyrobek:** Oh, great question. So there's a model for thinking about this that was called the Swiss Cheese Model. I did not come up with that name, although I'm known for naming things with really dumb names... It was actually a NASA name, the Swiss Cheese Model. Conceptually, basically what it is - every approach to testing is like one slice of your Swiss cheese. It's going to have good coverage of some aspects, but if you look at it, it's going to have holes, for every type of testing you do. You obviously have many different layers of testing that you do, and the intent is if you stack up those layers of Swiss cheese and look at them through the stack, then you're not going to see through holes, if you're doing it well.
+
+Examples of our layers of Swiss cheese - you're going to be familiar with things like unit testing, and static testing, and things like that, that are commonly done in software... Software in the loop testing - so this is where you have full on physics level simulation running of your full system, even your full fleet, and you're able to then write test scenarios on top of that.
+
+\[32:10\] We talked a little bit about the hardware in the loop testing, where you take the hardware itself, the brain of the aircraft, plug it into a simulator and fool it into thinking it's in the real world... That's great for testing the actual hardware level components. If you short out a certain bus, does the actual full software stack detect that short and do the right thing? And then a lot of things in flight tests. We do all kinds of different flight tests. We do on the order of 10,000 plus flights a week at this point across our flight testing, and almost every single one of those flights is pushing the system to an extreme in some way. We actually have this fully automated software system we call chaos monkey, where basically what it's doing as we're dispatching all of these flights is it is -- chaos monkey is literally throwing chaos into the system as those missions are happening, that either will cause the airplane to do something extraordinarily dynamic against the physics limits of the aircraft, or turning off subsystems of the software, killing various rotors, things like that... And making sure that the system, despite any given flight having dozens of weird things happen - nominal events, as we call them - in that flight, does the whole system handle it gracefully and safely?
+
+And this is just on the software side. When we think about hardware, very similar things - many layers of testing of the hardware, to make sure we understand deeply its performance and its reliability. And there are entire teams of folks who look at all that testing, and basically are constantly questioning "Okay, in this release we have a new capability we're releasing. Did we add enough testing to all the different layers of the Swiss cheese, such that that's going to be responsible to ship?" And there are times we think we're ready and we do that review of everything we did, and we're like "Nope, this isn't ready. We're pulling this from this release. It's got more work to do", just on the testing side, to get that confidence.
+
+**Adam Stacoviak:** These releases, are they -- share as much as you like. You seem to be very sharing, and I like this. I think about releases and I think about "Well, if I released my test level from idea to new capability", as you said, into production, there's probably a several layers that are in that middle ground, which is like a true release, but not in a real world scenario, so to speak... How do you plan releases? How do you graduate releases from tests, delivery, observability, test again, potentially... You said you do 10,000 flights a week. Seven days in a week - that's 1400 flights a day, roughly... Gosh, that's a lot of flights. But how do you release and do that in a way that's like "Okay, it's in production, but it's not real production." I'm assuming these things, so... Walk me through that kind of idea.
+
+**Keenan Wyrobek:** Okay. So depending on the scale of a given feature... To talk about one of the biggest features we've ever released - we just released it recently. So we talked about the drone that flies up high, it has a wing, and can hover. Until about a month ago, in our releases and our operations, we were only hovering. We were not flying on that wing. So we were flying it more like a traditional quadcopter. It's a pentacopter, but we're flying it like a traditional quadcopter, just hover, which means the range is very limited. But while we were working on actually being able to fly on the wing - and the software and systems to transition from hover onto fixed wing, as it's referred to in industry, and then back into hover, including when anything could be going wrong, like any of those rotors could not be working and that kind of thing... There's a lot to that, and we've been working on those features, at some level of development and testing, for two years to enable this. So two or three months before, that release cuts. So when I talk about releasing every six weeks, we cut basically at the beginning of that six weeks.
+
+\[35:55\] We will look at all the new features that folks think are ready for a release, and we'll look at like "How tested are they? How proven are they? How good is the test coverage?" and all these layers of Swiss cheese. And then we'll decide "Yup, this feature is ready. This feature is ready. That feature is not ready. It's not going to go in this release", and we will cut a release. And then once we cut a release, there's a bit of a stabilization that takes us about a week, to make sure that release is really stable, and then it goes into this very structured release campaign that takes about five weeks on average, which - it depends on the release, but something like 50,000 flights, a whole bunch of hardware in the loop tests and software loop tests and all these other tests we will do again as well, as part of that process. And then the people -- all the teams are basically paying a ton of attention to all the data coming out of all that testing. And over the years, we've developed all kinds of tooling and dashboards that help all the teams look at that data, and look for anything weird, anything that might we consider a near miss. Obviously, the real miss is if we have a parachute landing during that testing, we're going to know exactly what goes on there. But we're looking for anything that's at all weird. We call it a near miss. And studying it, and seeing if it needs to be addressed, or if it's fine.
+
+You can kind of think of that flight test as like the last leg. If we find something in flight test, that's a red flag to us. Like "Hey, the other layers of Swiss cheese aren't good enough." So if we find something in the flight test during that campaign, we're immediately investing in those other levels of Swiss cheese to make them better, and improve their coverage. And then only once the data says we've got a system that meets or exceeds the overall target level of safety of our last release, do we then say "This is ready to actually ship."
+
+**Adam Stacoviak:** Do you have the idea of a black box in your thing? I mean, I know that airplanes do, so I imagine you've got some version of a black box, where if it does - I'm sorry - crash, that you've got some sort of thing in there... And then I also think about observability, because you've got all this testing happening... And I think about like traditional software in the cloud - it's kind of easy to orchestrate observability. But over the years I'm sure you've got some version of WiFi or access to the thing digitally, via a network of sorts, telemetry going... Is there a black box, and how does observability happen in those scenarios?
+
+**Keenan Wyrobek:** Yeah, absolutely. So we log a ton of data on the vehicles, at all times. And that's of course done in all this testing, but also in production. And the philosophy is that anything that could possibly go wrong, or be a near miss, you've got to study. And you mentioned crashes - I should mention we have a parachute system for the aircraft, that's kind of the backup to the redundancy. So there's redundancy in the aircraft, so it can nominally fly when things don't work, but the backup to the backup is a parachute. So it'll float to the ground like a -- it's actually made for us by a skydiving parachute company. So it's like a skydiver coming to the ground.
+
+So yeah, if we ever have a parachute landing in tests, or operations, or anything that's even close to a weird data, we go study that, to basically deeply, deeply understand it, because quite often it's a new insight. We're so far in the long tail of the problem that many of the problems we see now, they happen one in hundreds of thousands of flights. And so you basically -- like, if we didn't log it, think of it as like "Oh no, now we have to wait for this to happen again, in hundreds of thousands of flights", which is a total shame.
+
+**Adam Stacoviak:** Total shame. It should happen more often... \[laughs\]
+
+**Jerod Santo:** That's right.
+
+**Keenan Wyrobek:** Exactly. No, that's not the shame. The shame is not logging it. So we log these things. Now, in addition to the logging, to enable us to like go through those logs and understand everything that happened -- we use those logs not just for like root causing a specific problem, we use those logs to understand things in aggregate. So we have tools where you can mine those logs longitudinally, across thousands and thousands of flights, to deeply understand the statistics of a certain thing. That's really important as well. And then of course, during live operations - yeah, the aircraft are reporting over radios, live, what's going on. We have a remote operation center that's monitoring all of that... And so during live operations we have certain things that basically -- our pilots, I think of them more like an air traffic controller, where they can be like "Okay, you go back and dock", and that kind of thing, and understand what's going on at a very sort of fleet level.
+
+**Jerod Santo:** \[40:16\] Have you ever logged a tornado?
+
+**Keenan Wyrobek:** Hah, we've definitely -- yeah. So the closest thing we've logged -- we've logged some very small... I don't think you'd call them a tornado. They're basically -- the things are too small to be considered tornadoes. At some of our testing, we're actually chasing some tornadoes right now, as we look for hail and more extreme weather events with our mobile test rigs.
+
+The craziest thing we have a lot of log data on is what's called a microburst. So it's the beginning of a thunderhead formation, about first 90 seconds of a thunderhead. You have the vertical winds at like 50, 60, 70 miles an hour vertically, in the middle of the microburst; on the sides of the microburst, the same thing goes down. Those are some of the most extreme weather events, that are hard for the drones to handle.
+
+**Jerod Santo:** What do they do in those circumstances? Do they turn around and leave, or do they land immediately, or what's their protocol?
+
+**Keenan Wyrobek:** So it really depends how big it is. this is actually one of the first things we developed AI to forecast. It's because if this is really big, if it's kilometers across, and really strong, if we're in it we're going to end up parachute landing. And we know this because in a lot of our operations around the world when we're doing emergency deliveries for a patient on the table, delivering blood situations, we will take -- we will basically ignore our weather limits and fly in very extreme weather. So we go through these things. But then for non-emergency deliveries, we rely on our forecasting to keep us out of this stuff, in those really extreme events. If they're small enough - yeah, the control system will fight its way through it. But of course, the aircraft has its physics limits. It can only climb so fast, it can only do so much. And so in certain situations it can't overcome them, and yeah, that's when the parachute kicks in.
+
+**Jerod Santo:** I assume there's some sort of wind limit that there's just no chance for anything to fly through successfully. It's gotta be around a hundred miles per hour or so, maybe a little bit less... I don't know, but there's certain storms where it's just like, you'd better just parachute out, because there's no there-there; there's no success as a possibility, regardless of the size of this thing. 30 seconds at 150 miles per hour is knocking down massive trees here... And so I'm assuming it's taken a drone off its course quite a bit.
+
+**Keenan Wyrobek:** Oh, yeah. There's a certain mountain pass in Rwanda I'm very familiar with, because we have to fly through this mountain pass to serve about a third of the country there... And in this mountain pass the winds get whipping. The winds can be higher than we fly. We fly that aircraft at 60-70 miles an hour, and the winds get up to a hundred miles an hour... And you'll see our drones flying backwards as it tries to fight through it, developing the control system and logic to... Because most of those winds are gusty, so if you can stay in the air long enough and then punch through it at the right moment, you can get to mountain pass. And yeah, it was a big breakthrough when we figured out how to make that logic robust. And you see the drone is flying backwards for a while, flying backwards for a while, flying backwards for a while, and then flying through at the right second; it's estimating the winds dying down online, and then it punches through the mountain pass and gets through.
+
+**Jerod Santo:** Do you have any video feeds? Because you're going to come up with some great drama in those moments if you could capture a video, just when it's really gnarly, you know?
+
+**Keenan Wyrobek:** I've got a great video of the logs of that. It's not \[unintelligible 00:43:28.29\] You can see the position on the maps, going backwards and things...
+
+**Jerod Santo:** Yeah, I don't think Hollywood's going to call you for the log file.
+
+**Keenan Wyrobek:** No... \[laughs\]
+
+**Jerod Santo:** Let's talk more about that wing. So did you guys just put the wing on there because you knew eventually you were going to use it, and so you just had it in manufacturing for a long time, but couldn't use it?
+
+**Keenan Wyrobek:** \[43:47\] Yeah, absolutely. We knew we needed the range. You get about 10x more range flying on a wing, for the same energy, than hovering. And our customers need that range. So many of our use cases need that range. And so yeah, we always knew we needed the range of flying on a wing, and we decided to launch without having developed that system yet... Really just because that's how we're wired. We want to get out into the real world fast, because we just know - when you're in the real world, serving customers, you're going to learn a lot of things you just can't learn any other way... And so we want to get out there earlier. Yeah, and that hardware was there, and the teams that spent all this time developing that hardware were like "We want to see a fly on the wing..." And of course, we didn't have the test sites for a long time as we've been developing that capability. It is one of those things where you kind of develop it, and then wait, and then once you can get -- yeah...
+
+One of the things... You're talking about regulations and safety - that is also part of our ethos. And a difficult thing in a startup is we hire people who like inherent urgency, because if you set a deadline on safety, you can end up in a really bad situation. If you're like "Cool, we must ship this on this date." Some of the stories you hear about, of 80 companies that have hurt people in the past - they take that mindset, and then they'll fool themselves into thinking they're ready, as opposed to saying "No, no, no. We set a safety bar, and we decide how we're going to measure ourselves against that safety bar, and that's when we ship." It means your dates aren't quite as predictable, but you still need to move fast, and so you need people with a very inherent sense of urgency to work through those problems and get there, as opposed to a more traditional "We have a date and we're going for it" management approach.
+
+**Jerod Santo:** So what was the range prior to the wing, of a typical drone? ...assume good conditions - how far do they go, and how long a time, and then... You said 10x, so I can do the math, but... That's a huge win.
+
+**Keenan Wyrobek:** Yeah. Great. Yeah, so we're doing about a mile and a half hovering, and we'll be able to do -- right now we're at about five miles on the wing, and we'll be able to take that to about 10 miles over time... Again, as we continue to validate and sort of work our way up to it, once we feel confident in it.
+
+**Jerod Santo:** And I assume each drone launches from a location, and then has to return to the same location, at least until you get multiple hubs in a city, or something.
+
+**Keenan Wyrobek:** Yeah. That's actually a software feature that's coming in and released next year, is that ability to go from one doc, go make a delivery, and then dock somewhere else. But yeah, you're exactly right. We don't have the density of docking locations yet to need that feature.
+
+**Jerod Santo:** Right. So that cuts your range in half, because you have to get out there, and you've got to get back.
+
+**Keenan Wyrobek:** Exactly.
+
+**Jerod Santo:** Okay. So when you say five miles - if I live five miles from your facility, do I get deliveries, or I've gotta be two and a half miles?
+
+**Keenan Wyrobek:** No, you can live five miles.
+
+**Jerod Santo:** Gotcha.
+
+**Keenan Wyrobek:** Yeah. So once we have that... If we can dock at a different location, nearby the delivery site - yeah, we can do more than 10 miles. That's actually an interesting thing... We've learned over the years that it's way better when you're working with these partners to talk about a conservative range, than like where you're going to go in the future... Because it's so complicated to talk about that when you get into the details of your terrain, and your physics, and your local weather, and all that stuff. It's like "No, no, no. We're just talking about this little circle." And then they're thrilled that we can do more than that.
+
+**Jerod Santo:** Right. I was just flying my drone the other day - I have a little DJI Flip, and I was flying it... Too far away, slightly too far away, to where it was telling me to come back... And I'm like "Yeah, I'm coming back." So I started flying it back, and then the wind picks up, going against it on the way back... And I'm sitting there, thinking "Oh man, it's just not going to make it back." Because the battery percentage continues to go down, and it's still 1200 meters away, and... It was very dramatic. It did make it back to the house, thankfully, but just barely. So I've got to imagine that's a stressful -- you know, we're talking about range anxiety with all of our new EVs. I'm sure you probably have extreme forms of range anxiety with your -- because these are expensive machines, right? What's an all-in cost on one, if you can share that? ...to build and manufacture. What are we talking about?
+
+**Keenan Wyrobek:** I can't share that specifically, but it's lower than you would think. And this kind of -- this is part of the reasons we develop our own drones. Both for this range problem you're talking about, and for the cost. But for the range - take our fixed wing platforms. This is what we do with these very long range deliveries. There's no hover capability, it just flies like a plane... We advertise, basically, a 50-mile service radius, but if it wasn't windy and there wasn't a mountain in the way, it could easily do a hundred mile service radius. So we literally have twice as much energy as you would need nominally, so that we can make it back on those windy days reliably. And that's just really important, because once you start serving a hospital, delivering blood, you can't be like "Oh, it's too windy today. Sorry, guys, we don't serve you. We only reach the people closer."
+
+**Jerod Santo:** Totally.
+
+**Keenan Wyrobek:** \[48:22\] It's like, no, we turn you on, and we want to serve you all the time.
+
+**Jerod Santo:** Yeah.
+
+**Adam Stacoviak:** Does anybody get underwhelmed by the range, when you say five miles? I'm not throwing stones here, but I'm just saying... I know you're innovating, but does anyone ever go like "That kind of seems small"?
+
+**Keenan Wyrobek:** It all just depends on the use case. The 10 miles was developed going really deep with our health partners, and exactly on all the use cases that they want to use the drones in metro areas. And of course, when you're going out of metro areas, that's where the 50-mile range drone comes in handy. And all of this has been developed just basically technology second, customer first. \[unintelligible 00:49:02.29\] with the customers. For this precision delivery platform, we had contracts with partners while we were still in what we would call the conceptual phase of that project. Nothing was locked in yet. And having those contracts with the partners gets them to give you -- that gets them invested enough in us to give us the feedback and the data to make sure that we make something that's going to be useful for them. And so yeah, some people get overwhelmed, but we just have that confidence because we've had so much time in these use cases, analyzing them, going deep with our partners to just have confidence we're making a platform that'll serve their use cases well.
+
+**Break**: \[49:42\]
+
+**Jerod Santo:** Could you walk us through the user experience that either exists now, or will be existing as you guys roll out domestically in the United States? ...from the end user perspective, but then also as we go, maybe explain the software systems that have to support that... Because I'm fascinated where software meets the real world and there are all sorts of weird situations that you have to account for. I know there's an app. I assume it's similar to a DoorDash kind of a thing, where I'm going to order - let's just say I'm going to get a Chipotle burrito, because I already cracked that earlier... And I'm a fan. Not a sponsor, but I'm a fan of the old Chipotle burrito. And I know it's under eight pounds, so it's going to fit into my zip... Unless I get the double. I don't know, can you get a 10-pound burrito from Chipotle? Maybe not yet. Someday. Someday we can dream.
+
+I open up the Zipline app, I assume... Or do I open up the Chipotle app? Maybe that's where we start. Do I start in the Chipotle app? Do I start in the Zipline app? Where do I start?
+
+**Keenan Wyrobek:** Yeah. So soon you'll be able to start in the Chipotle app. Today you start in the Zipline app. And depending on where you live -- if you live in the best place in Dallas-Fort Worth area, you open that app and you're going to see a bunch of different store fronts. You'll see Walmart, very soon Walmart Pharmacy, you'll see Blue Bell Ice Cream, you'll see Chipotle...
+
+**Adam Stacoviak:** There you go, Adam.
+
+**Keenan Wyrobek:** ...you'll see just a whole bunch of great restaurants. Buffalo Wild Wings, and others. And yeah, you pick what you want... That part of the experience feels a lot like DoorDash, or someone driving the thing to you... Then they'll start prepping it, our partners will prep the food or prep the order... And then we put it in this thing that we call a zipping point, which is this fun little kiosk you put the package into. And similar to how we deliver, the drone stays up high and the delivery zip comes down 100 meters, grabs the package out of the kiosk, and brings it to you. And this is where things are very different than delivering by car. We actually have to warn people "Hey, your food may be too hot to eat."
+
+**Jerod Santo:** Really?
+
+**Keenan Wyrobek:** So we deliver so fast that we've had people be like "Oh, yeah, I just assumed my hot wings would be soggy", and they just stuck them in their mouth and burnt their mouth... Because they're so used to how long it takes for deliveries to happen the old fashioned way, that when it takes only a few minutes, and it's well insulated and controlled the way we do it, the food's actually fresh, and sometimes very hot still.
+
+**Jerod Santo:** Yeah. That's a great problem to have. So show me the Chipotle side. From your integrators, your restaurants... And I'm just picking on them because I know they're a partner of yours. Obviously, there's going to be a bunch of them. But imagine that I do order a Chipotle burrito out of zip line. I assume an API call kicks off over to the Chipotle, and makes the order, and then they have to somehow tell the zip when it's ready... How do they do that?
+
+**Keenan Wyrobek:** Yeah, so with Chipotle they use their software system the way they use it for any order. If you're ordering over the counter, or you place an order to the Chipotle app, and it shows up on their -- they call it their make line. That's the people that actually put the order together. So they're going to make this like any other food they're making; our system gets hit over the API, so we know when they started that, and we know when they're done... And then when they're done, we'll put it in that kiosk, dispatch the zip, the drone over to come get it, it'll come pick it up... That's usually about a minute or two after it's ready. And then we deliver it anywhere from one to five minutes later, depending how far away you are.
+
+**Jerod Santo:** Right.
+
+**Keenan Wyrobek:** We deliver it in your in your yard. And yeah, we have a fun little app that shows you the drone coming, and it gives you that really super-precise ETA.
+
+**Jerod Santo:** I think one of the things I didn't appreciate about this kind of delivery that's really cool is just how human-centered it is... As opposed to whenever that driver is going to show up, and sometimes it's like "Wait, why did they take that crazy route? My kids are hungry." We tell you "No, no. Within a minute \[unintelligible 00:57:38.12\] actually show up and you can walk outside and pick it up. It's just there."
+
+**Jerod Santo:** Now, can you hold your hand out and it'll put it in your hand?
+
+**Keenan Wyrobek:** \[laughs\] I don't recommend it.
+
+**Jerod Santo:** That can be a cool feature there, Keenan, come on. Put your hand out, and we'll put -- you have a third little... So you have the big zip, and the little zip, and then like a hand that comes down; it just drops a little bit lower, and...
+
+**Adam Stacoviak:** Keenan's like "Nah..."
+
+**Jerod Santo:** \[laughs\]
+
+**Keenan Wyrobek:** \[58:06\] I literally tested -- the way we talk about this is we don't want people to do stuff like that, but we test in case people do that kind of thing. So it's delivered into my arms, and I love it...
+
+**Jerod Santo:** It felt good, didn't it? It felt special.
+
+**Adam Stacoviak:** Do you pause the delivery then, when you detect anomalies you don't want to be present in? I assume that's probably a phrase or terminology...
+
+**Jerod Santo:** Like a child standing there, underneath your target? Yeah.
+
+**Keenan Wyrobek:** Yeah, it's actually kind of funny... So we do a little bit of avoidance. So if we're coming down and we see the person standing right there, we'll scoot over and deliver next to them.
+
+**Adam Stacoviak:** They keep scooting closer, you keep scooting further...
+
+**Keenan Wyrobek:** Exactly. If you do that -- you can make a really fun game for kids, and we don't want to be a fun game. You know what I mean? We don't want to make this to be a game of "Oh, it turns out if you chase this thing, it'll become a game", and then everybody wants to chase these things, and we can't deliver... And so basically, what we've found to work really well is just slow down. One of the things we like about that little delivery zip is it's very benign. If someone touches it or something, it's not a safety concern. And so yeah, we just slow down and keep going, and make that delivery, and get out of there. And basically, the way I think to talk about it is like "We be boring", so that we're not creating a fun game for every kid. As a parent of young kids, I've brought them to the test site too, and I know exactly what -- if you did that avoidance with my three-year-old, she would have so much fun, and we would never make deliveries.
+
+**Jerod Santo:** Right. Well, it might be a consideration for down the road, of getting even more human-centric; it's like, allow the orderer to set a mode. "Do I want to have fun with it? Do I just want my food?" I know you guys want to save energy and battery and all that, but... You're already pretty happy to have your food delivered from the sky. I can imagine that's a moment already. But to have a little bit of a personality with the thing - it might take it over the top, to where it's the only way we're ever going to get food, is to just give us... That's why I think the handoff would be actually amazing, but... Yeah, too dangerous perhaps to do that today.
+
+**Adam Stacoviak:** For sure.
+
+**Jerod Santo:** It's a tomorrow thing.
+
+**Keenan Wyrobek:** I love that idea. Your head is exactly where my head is. To me it's not a question of if, it's a question of when. So I look forward to it.
+
+**Adam Stacoviak:** Well, someone will always mess around too in those scenarios, but... Did you mention at all when, time of day, you can't deliver? Is there like danger zones where you can't deliver in the evening, or after dusk? Or is there light concerns? Do you have lights on these things?
+
+**Keenan Wyrobek:** Yeah, so we can deliver anytime.
+
+**Adam Stacoviak:** Anytime. So 24/7.
+
+**Keenan Wyrobek:** Yup. Yup.
+
+**Adam Stacoviak:** Wow. Okay.
+
+**Keenan Wyrobek:** Right now that's just limited by our partner hours. And most of the places we operate in the world, we deliver 24/7 with the health systems. And it's always interesting to see - you'll work with a restaurant partner, and all of a sudden they'll realize just how much volume we're adding to their normal business, and they're kind of like "Oh, wait a minute... If we operate with you all, and you can go 24/7... You're already operating 24/7 for other people, so we could be open at night..." There's a huge part of our economy that works at night. Night shift workers, and things like that... We know that because all the health people we work with, they work 24 hours a day... But obviously, lots of manufacturing operates 24 hours a day, and a lot of our partners are seeing it like "Wow", there's a whole part of commerce they can address, and a big part of the demand they can address by operating more hours. And obviously, partnering with us, we can do that in a way that's people really like. Yeah, it's really cool. You'll get whatever you need, 24 hours a day. As a parent -- just simple things, like kid was spiking with a fever at night, and you've ran out of medication to bring that fever down... The only thing worse than having a kid with a fever is taking your kid to a 24-hour pharmacy at night to go grab that stuff you don't have.
+
+**Adam Stacoviak:** \[01:01:56.01\] Yeah. Croup is a big deal for parents. I know that we've battled several late night ER visits for croup, and generally - I don't want to go into the medical scenario here, but basically, you have to be seen by some sort of medical professional to get a special medicine that is for croup, that sort of helps your vocal cords and things reduce, or whatever. And it's kind of a pain in the butt, because you kind of know, as a parent -- you're like "You know what, I'm Dr. Mom or I'm Dr. Dad. I've been down this road before with my child. I just need that particular prescription." So if you had a relationship where you could do telehealth, which is super-popular, we would not have to leave the house at all. If that were a scenario, we can make a phone call in our telehealth, and maybe 10 or 15 minutes later we've got a delivery from zip that says "Hey, here's your croup medicine." You didn't have to leave the house, you can do it efficiently... You didn't have to be on the road...
+
+You talk about the efficiencies and environmental effects... Like, we didn't have to drive, we didn't have to lose sleep and maybe ruin our day the next day more; so we got the medicine -- and there's so many unintended consequences in that scenario that just make it so interesting to think about how you can reorient the world around this being a thing.
+
+**Keenan Wyrobek:** Oh, totally. I mean, Adam, I'm so excited about this. And we have partnerships now with about a dozen regional US health systems, and all of them have this use case in that partnership. They have some other ones as well, but this is -- they call it the physical side of telehealth. Telehealth is obviously becoming a really predominant thing here... And like you said, for so many things - there's a physical side, too. And so if you can get that inhaler delivered right away, or that croup medication delivered right away, it's huge for the family. It's huge for -- yeah, and you're touching on something too that I didn't appreciate when starting Zipline, that I now I appreciate a lot... It's just how much health affects basically the economy. You mentioned just - yeah, the parents not being exhausted the next day; their work is going to be better, and they're going to be more productive in the economy, and just keeping people healthy in that sort of preventative way is something that I've become extremely passionate about, because I've just seen this cause and effect in my own life, and in all the patients we serve around the world... Just how powerful that is when a parent doesn't have to stay home, or a parent doesn't have to basically be affected in their professional world by what could be a very preventable sort of health -- some of these things kind of feel like inconveniences even, right? Blood, obviously - very acute, big, life threatening... But a lot of this stuff, it's just -- a little bit of convenience just goes a long way for people staying healthy and productive.
+
+**Adam Stacoviak:** Have you -- I imagine you probably have, but have you imagined a world where Zipline's at scale, and you don't have a couple hundred in production, you have, I don't know, half a million; I don't know, a lot. Just an astronomical number that you don't even consider today. How the world changes, maybe even how we see the world... Because I can't imagine \[unintelligible 01:04:58.18\] Zipline delivering something to me. I'm sure it'd be cool once I have it in my own life, but... Once this kind of technology, whether it's you or just generally, like flights, isn't only owned by United or Southwest, for example, when this kind of delivery or scenario becomes ubiquitous, have you imagined that world and how that changes the world?
+
+**Keenan Wyrobek:** Yeah. I mean, I'm super-excited about it. This is part of the reason we spend so much time on things like making the system silent, is -- one of the things we know is I don't want to notice these things. Just like a bird flying over your house. How many birds flew over your house today? You have no idea. Why? Because you don't hear them. And we spent a lot of time on things like that, because we just know that as these things scale, there's certain things that we've learned from our experience that just matter a lot, like it being quiet.
+
+\[01:05:49.24\] So there's certain things like that that we spent a lot of time on, to make sure that when this scales, it's a world we actually want to live in. And then, of course, the net effects of this kind of instant delivery... I live in a small town here in California, and - yeah there's, there's two bakeries... They're so excited for when they get their little kiosk and they can just send out fresh bread in the morning to people.
+
+**Jerod Santo:** Yes, I was just thinking about that.
+
+**Adam Stacoviak:** Timely deliveries.
+
+**Keenan Wyrobek:** Exactly. And there's so many things on the personal life side like that, like getting fresh bread delivered, there's so many things on the work side... There's so many jobs that just require stuff. Obviously, doctors, medicines - it's so obvious. And auto mechanics - it's so obvious. Just yesterday, my HVAC system in my house went out; a little blower had failed, and the guy came out and he's like "I'll be back in two days with the part." And it's like "No, no." You would deliver the part right while he's there. And I looked it up, the part he needed - it weighs less than eight pounds. We can deliver it.
+
+**Adam Stacoviak:** Did you deliver it?
+
+**Keenan Wyrobek:** No, we don't serve that area yet with HVAC parts.
+
+**Adam Stacoviak:** Oh, that would have been so cool. If you're like "Hey, I've got a solution for this. Hang on. There's no delay. I got the part."
+
+**Keenan Wyrobek:** Yeah. Even just talking to him about his business. This is an independent guy, who does a lot of these kinds of HVAC repairs, and fixing heaters and stuff... And his customers, they hate it when he charges for these visits just to figure out what he needs. If he could just fix it there on the spot, not only would he be able to serve so many more customers, but those customers would be so much happier. And he has somebody who hates that part of the customer conversation, of like "I've got to charge you the 200 bucks for just coming by and seeing this thing." He'd love that to go away, because he hates that part of his job. And anyway, there's so many pieces like that, I just can't wait to see what people imagine, and how they will use it. Because every country we operate in, people are using it in ways I never expected.
+
+I'll tell you one fun story from Dallas. This summer we were looking at our usage data, and most people would either order at one address, or order to maybe two addresses. But we saw this one person who was ordering to a different address almost every single day. Like, what is this? What's going on here? And so we send this user an email saying "Would you share how you're using the service?" And so this is a postal delivery carrier, and his New Year's resolution for this year was to eat healthier... And he was getting a salad delivered to him along his route every day, for lunch. And I just -- anyway, I just loved that. I would have never have guessed. I would have never guessed. We call him the Johnny Appleseed of Zipline, you can see...
+
+**Jerod Santo:** What about gifting? It's a great way to gift. It's just like "Hey." Call somebody up, "Go outside." "Why?" "There's a delivery coming for you." And this thing just drops a gift out of the sky.
+
+**Keenan Wyrobek:** Oh, we did this. So we did a really fun thing for Easter this year, where we basically loaded up the little delivery zip with a bunch of Easter eggs; no packaging, just Easter eggs. And by the way, one cool thing about what we do is it doesn't require packaging. That's another part of the future I'm excited about. None of this cardboard jungle we all have in our garage from Amazon boxes.
+
+**Adam Stacoviak:** Oh, my gosh... Save me.
+
+**Keenan Wyrobek:** And so we put a bunch of Easter eggs in there, and it was really interesting to see how many -- we had grandparents in other States, outside of Dallas (we were just delivering in Dallas) place orders for their grandkids to deliver this pile of Easter eggs into the yard. And the kids loved it, and the grandparents loved it... We had just sort of this out of state gifting, and it was so cool.
+
+**Jerod Santo:** That's awesome.
+
+**Keenan Wyrobek:** And there's so many things like this, that I can't wait to try.
+
+**Jerod Santo:** There's so many places you can go with eight pounds, but at a certain point that becomes your limit; I don't think you guys are anywhere near that point. And we all know that. But at some point, have you thought "Okay, what's the next -- where's the zip line jumbo? What's the weight that we need to go to expand to a new set of things?" And then "What do we need to manufacture in terms of drones to actually get that done?" I know there's a long range one. Is there like a heavy weight one, or is that just too far down the line to care about?
+
+**Keenan Wyrobek:** \[01:09:49.14\] It's a great question. Eight pounds gets us really far, as you're alluding to. Something like 80% plus of Amazon packages are less than eight pounds. So many of the use cases that we are really focused on... A meal for a family of four, from almost any restaurant, is less than eight pounds, as an example. And then, yeah, if you need more, we just send two deliveries, back to back. It's two little bags of food. It's no boxes, none of this other stuff.
+
+It'll be interesting to see how it evolves. If I was a betting person, I think we will increase our payload over time beyond eight pounds. There's a good chance it won't go a ton past that. Maybe 10, maybe 12 pounds... And really, what it comes down to is what doesn't fit. Once we get to that scale, we can deliver a very full bag of groceries at that scale. And the advantage of having one platform, as opposed to like different variants, is we can drive the cost down further and further, just from scale. And so as the cost per delivery goes down, there's a chance that we'll never be able to compete with -- if we made a drone that could deliver four bags of grocery at once, that will always be more expensive than a smaller drone delivering four bags just 30 seconds apart onto your picnic table in your backyard. And then there's stuff left out, like the flat screen TVs... We obviously can't deliver those.
+
+**Adam Stacoviak:** I was thinking about the TV, honestly. I was like "Can you bring the TV to me?" I'm just kidding, I was not thinking that. That's cool.
+
+**Keenan Wyrobek:** Yeah. We've been looking at the car -- the partner will be launching within Dallas, and we can deliver something like 90% of their car parts... But obviously, 10% we can't do. And yeah, we think about that. Zipline has always just kind of been like customer-first. So once we get the data, once enough customers ask us for the same thing over and over again, and we go deep with them, if we have conviction that there's a thing that will work for them, that will make sense for the business... You know, that's how this precision delivery drone inside of a drone system came from... So yeah, we'll do more things in the future.
+
+**Jerod Santo:** Is anybody else doing this? Like, who's the Southwest to your United? I know that Amazon announced the thing, like you said, years ago, about drone deliveries, and then that got canceled, or sidelined, or whatever; maybe it's happening, but... I hope there's other people that are tackling the same problem, because we want diverse and unique solutions, and learning from each other... So who else is working in this space that maybe is either your competition, or even that you admire what they're up to?
+
+**Keenan Wyrobek:** Yeah, so Amazon has started doing some deliveries, quite limited as far as I understand. Google has a project called Wing. I admire a lot of how they're thinking about the problem, and we collaborate with them a lot on some of the projects with the FAA. I really like how they think about some of some aspects of safety, for example, and things like that. I admire that a lot.
+
+There's a fair bit out there... I'll be honest, I don't think a lot about the competition. For us, it's just about the customer, and just "What can we do? How can we do better every day?" and really measure ourselves on our own velocity towards delighting those customers.
+
+**Jerod Santo:** Yeah. What's 2026 going to be about? Scale? Other things? What's it going to look like?
+
+**Keenan Wyrobek:** It's a lot about scale. Our long range platform is scaling around the world, so that will continue in 2026... And yeah, this precision delivery platform is scaling... Yeah, it is scaling really fast. So a lot of what 2026 is about is building up the supply chain capability to manufacture faster, building all the internal muscles to launch the charging sites, and things like that, to enable scale. Yeah. And just getting better and better at that user journey you were asking about, of like "How do people order, and how does that all work for all the different use cases?" We've found -- you can't be super-generic there. Different use cases need a tailored user experience for it to really be magical... And so we'll spend a lot of time there as well.
+
+**Keenan Wyrobek:** How do you pick your next city, and why is it going to be Omaha? \[laughter\]
+
+**Keenan Wyrobek:** It's going to be Omaha, because the --
+
+**Jerod Santo:** Yes...!
+
+**Keenan Wyrobek:** Yeah.
+
+**Adam Stacoviak:** "Yeah..." I love that. \[laughter\]
+
+**Keenan Wyrobek:** \[01:13:50.14\] It's really about the testing and validation story. So I can't announce the next city, but we'll be launching our next cities in the first half of 2026. Part of how we choose those cities is they're very similar to the Dallas-Fort Worth area. Similar in terms of weather, similar in terms of terrain, that kind of thing. And that's because that's what we validate the system for. In parallel, we're validating the system for expanded weather, more complex terrain, smaller yards, all those kinds of things. And so basically, once we validate the system to go into metros that aren't like -- we refer to like the Sunbelt metros; then we'll move into these other metros as well.
+
+Yeah, it's really fun. We have a test site up in the Cascades, where we're doing a lot of testing to validate the system for metros that aren't at all like Dallas and Fort Worth, but... Yeah, it'll be Sunbelt metros first, and then we'll keep going from there.
+
+**Jerod Santo:** Yeah. You've solved a ton of technical problems to get here, and now you're trying to take it in lots of places. What are the big technical problems that are ahead of you? If you had a checklist as CTO, like "Here's the things we need to solve." What's unsolved for Zipline?
+
+**Keenan Wyrobek:** Almost every front, there's more work to do. I've been thinking a lot recently about weather forecasting... So we spend a lot of time to validate the system for certain weather. But then when you operate, you've got to know how -- like, how do you not go in that weather you can't go in? And that's harder than it sounds to do at scale. There's some easy things, like wind limits. Forecasting that is fairly straightforward. But there's much more complicated things, like icing conditions. Right now we're testing to understand exactly what level of icing conditions we can handle, and how to handle that well, but there will be a limit there that we're gonna need to forecast. Forecasting icing conditions is notoriously very hard to do. This is like when you have fog, and clouds, and cold weather; oftentimes you fly through that and you just cover your aircraft in ice.
+
+Another good example is lightning conditions. We are currently chasing lightning with our mobile test rigs right now, to really understand how does lightning interact with the system. We think it won't based on analysis, but lightning is one of those things where you don't trust the analysis. You've got to go in the real world and get real amounts of data to get confidence in that analysis.
+
+So there's a lot of work, both on the capability side, to expand the kind of weather we can fly in, and on the weather forecasting side. There's a lot of work there. There's a lot of work in the airspace side, of just "How do you manage this fleet at scale?" You were talking about going from one charging point to a delivery, then landing at a different charging point... There are some things there that are just way harder than they sound, where you can end up with your fleet kind of like migrating organically in a direction away from where the supply is... That can happen.
+
+**Jerod Santo:** Like a beehive. They're swarming. They're going somewhere else. Oh, no...!
+
+**Keenan Wyrobek:** Exactly. So you've got to manage that carefully, to make sure you can keep delivering for people. So there's a lot of -- almost every front of what we do at Zipline, there's a reimagining happening for the scale, for new capabilities. We've talked about some of healthcare use cases... There's a lot of new capabilities we need to build, both into how we operate the drones, and into the tools that the humans use on the healthcare side, systems side, and on the patient side to use the system... So there's all kinds of exciting things ahead of us. I think of us as like -- we're at the very beginning of this journey, and there's a really exciting roadmap ahead.
+
+**Jerod Santo:** I love it, man. I'm excited. I love that one of your teams is like a group of storm chasers, and they're just like "Our job is to find lightning in the world, and then go fly these drones up into it and see what happens."
+
+**Adam Stacoviak:** "What do you do for work?" "That's what I do."
+
+**Jerod Santo:** Storm chaser for Zipline.
+
+**Adam Stacoviak:** "That's my job." That'd be a rad job.
+
+**Jerod Santo:** Yeah.
+
+**Keenan Wyrobek:** It's a badass team. And they're hardcore too, because chasing storms when you have basically a big old trailer, we have these docks mounted to a trailer that they use... And obviously, with the storms, it's just hard to get around in these places... It's really cool, how they work and their dedication.
+
+**Jerod Santo:** Is there a manual override on those things that you could just take over if you needed to, or no?
+
+**Keenan Wyrobek:** \[01:17:50.26\] Yeah, but it's very high level. You basically have three buttons. A button to tell it to go back and dock, a button to tell it to loiter right where it is, like stay put, and a button to deploy that whole aircraft parachute and land where it is. That third button we've never pushed, but every regulator wants it... And then mostly it's the other two buttons. You can't jump in there and stick fly the thing. It'd be extraordinarily hard to. These aircraft fly themselves in a sort of superhuman way. How fast they make adjustments, how fast they can self-diagnose, and handle faults, and still fly nicely... Yeah, if you try to put a human in the loop of that, it would be extraordinarily hard to do.
+
+**Adam Stacoviak:** I was actually just watching a video from Mark Rober, and it was a goalie. It was an AI goalie, essentially. And I forget the soccer player's name, because I'm not a soccer fan to that level where I know the person's name, but I imagine it's probably like Messi, somebody who's super-famous. And they were going against the goalie, and they could not beat the goalie, because the reaction time was just so uncanny, like how fast they could react. And I take that kind of example, which is sort of a fun experiment, which Mark Rober is a scientist, and he's super-serious... But at the same time, it's meant to be fun, and not really an experiment of seriousness on a large scale.
+
+**Keenan Wyrobek:** Yeah.
+
+**Adam Stacoviak:** And to what level your advancements may be well beyond that, to truly say what you said. Because I'm sure there's a lot of people who say "You know, Keenan, I hear you saying that, but I still don't believe it..." Because they can't believe it until they see the truth of the superhuman ability to fly these aircrafts. I just think back to maybe the Mark Rober thing; that soccer goalie that was AI totally demolished the human, could not beat the AI goalie.
+
+**Keenan Wyrobek:** Yeah. It is surreal to see these things handle the -- when I saw the first... I mean, all this fault tolerance we test extensively. \[unintelligible 01:19:51.18\] one rotor fails, and the system handles it, and goes back and docks. When that is working well, when you're watching it happen, if you don't know what happened, you won't even notice it. You have people with binoculars watching that thing, because the only way you're going to see it is if you actually see the propeller blade, that one probably stopped spinning. But the aircraft handles it so seamlessly, it barely moves. And if it's windy at all, it doesn't move more than it does normally in the wind.
+
+This is one of the exciting things about autonomous systems, is they can get super-robust, which means we can operate and serve customers and deliver to patients and hospitals in all kinds of really crazy conditions. Again, the conditions that -- you know, flying low altitude in these storms, nobody does it. And when I say nobody, I didn't even realize this when we started Zipline. I thought like at least Medevac helicopters and this kind of things, the Coast Guard must do it... And we ended up calling some of them, because we were struggling to figure this out... And we're like "Cool, do you have any data?" And they were just like "No, we stopped flying those missions in the '70s. Way too many people died." And yeah, you can only do it if you use an autonomous system. And yeah, it's quite wild to see these systems perform, because they just -- they're so fast, and so superhuman. And then they can get these deliveries through these conditions that if you're on the road... Some of these conditions are so crazy if you're on a good road, you're kind of feeling like "I should not be on this road", but the drone is doing alright.
+
+**Jerod Santo:** Yeah, I used to listen to the Bill Burr podcast quite a bit; I don't anymore, but he is a hobbyist, but a trained helicopter pilot. And so he would chronicle some of his -- it's just him talking for an hour, so he needs a funny person to listen \[unintelligible 01:21:32.00\] so it's enjoyable. But he would chronicle some of his efforts at just flying a helicopter... And I mentally understood that it was difficult, but hearing somebody try to learn it, and the maneuvers they're going through just to get down the basics of flying a helicopter...
+
+\[01:21:52.03\] It's so fraught and dangerous to fly helicopters in adverse conditions... I mean, it's a death knell. And we know that we've had famous people like Kobe Bryant dying in helicopter crashes, just because of fog, or because of bad systems. And humans struggle, and have to have extreme training to be able to fly those things... And computers can fly them so much better now, that it's like "Why even put a human behind the sticks?", so to speak.
+
+**Keenan Wyrobek:** Yeah, absolutely. And what our system can do literally in the blink of a human eye is the tech side handle almost everything that can go wrong, but faster than a human blinks. And that's powerful for making autonomy scalable and safe.
+
+**Jerod Santo:** One last question for you, Keenan - what's your tech stack? So if you were to describe -- I know you have lots of different tech... If you were to pull out all the tools in your CTO tool belt - programming languages, frameworks... What's Zipline built on?
+
+**Keenan Wyrobek:** Right. Yeah, I mean, for high-performance stuff in the cloud we go for -- almost everything on the aircraft is moving to Rust. There's various parts of the tech stack... Almost every part of the tech stack kind of has its own little world. A lot of -- we've built our own in-house simulation system we call Phoenix. That was how we do basically a lot of the simulation and testing. Boy, the tech stack is broad. I worked a lot on ROS, the open source robotics platform back in the day, and everybody always asked me "Cool, you must run on ROS." And I'm like "Oh, man... No, we don't." If ROS 2 had come along early enough, we would have, but unfortunately we kind of went down that path before ROS 2 was ready for that.
+
+Yeah, we use open source where we can; of course, we contribute to open source where we can. I'm a big proponent of open source. One of the things we actually did on the tech stack, what we spent a lot of time on, is basically the future of autonomous flight. Even today in Dallas we share the airspace with some of these other drone operators we mentioned... And I think of the -- no one really knows how autonomous flight's going to scale in the airspace... And I can tell you what we do today in the airspace is not going to scale. It barely works for the scale we have today, which is very low compared to where the future is going. And there's a lot of need for folks in industry and Academia to like bring the best ideas to the table, make sure the best ideas win, so that we are both scalable in the airspace, and safe in the airspace as we scale.
+
+So Zipline's working on a bunch of projects like that, that we have or intend to make open to the community, so we can kind of prove it internally as like "Hey, this is how our drones don't collide at scale with each other", and make those technologies available so everybody can share that. That's a big part of the tech stack that we think a lot about, because -- well, again, I just love the... To me, the open source mindset comes from this thing of like "Hey, we're all in this together." And as I think about the future of the airspace, we're all in this together; it's going to be great for Zipline if all companies that need to use the airspace and all users of the airspace can use the airspace safely. But the future of autonomous flight is still to be figured out, and we're spending a lot of time on that.
+
+**Jerod Santo:** Do you see a world where there's some sort of open orchestration communication channel where you're not just detecting objects and avoiding them, but maybe you're actually publishing your future path, and so other systems can know, like, "There's zips going this way and that, and so I can avoid those paths"?
+
+**Keenan Wyrobek:** Yeah, absolutely. So today there's a version of that that we use in Dallas. And it's not live, it's done before you take off. Basically, there's a federated cloud-based system where the different operators' systems talk to each other. And basically, it's very rudimentary today. You can kind of think of it a little bit like a game of Battleship, of like "Okay, I'm trying to share my intended path in the air", sort of a tube in time, and I share it with the others and they're kind of like "Nope, that collides with something I've already reserved." And then you try to find another tube, and then it finds a tube, and then you're expected to stay in that tube about 95% of the time. Not very scalable. It's not a very efficient way to use the airspace, which is plentiful.
+
+\[01:26:13.12\] Internally, our aircrafts share information live with each other over a point-to-point radio link, where they share their intent of "Where I'm going in the next 10 seconds" or so. And they do that also over the IP network. I can't announce the partner yet, but there's a commercial drone operator in Dallas and we're going to be announcing a partnership later this year. They'll be using that same protocol, so we can share the airspace without those big reservations. And that's to kind of explore "Hey, is this more scalable? Alright, can we build up the data and then share it with the rest of the regulators and industry to kind of bring everybody along and make the case for what we think is a better idea than what we have today?" And then we'll move to that idea.
+
+So yeah, I think what you're describing is going to be the beginning of where we go with this. I think there's going to end up being a fair bit of structure, a lot of real-time sharing to make sure we use the airspace really efficiently, and scalably.
+
+**Adam Stacoviak:** It's like a DNS in the cloud, basically; in the sky, literally the clouds... Pluralized, not the internet cloud as we know it. But it's kind of like that; it's like a phone system, in a way, or a route system, or a flight path you now have to have predefined or known things. It's like "Well, we have these coordinates, and we know where we're at in the world", so maybe they're logistically like that, or something like that.
+
+**Jerod Santo:** Yeah, but constantly changing. Yeah.
+
+**Keenan Wyrobek:** Exactly. Yeah, we spent a lot of time talking to the original folks who created ICANN, which created the DNS system, to like "How did that actually come to be?" Because in some ways it's kind of a miracle, how open the internet is, and how it evolves...
+
+**Adam Stacoviak:** Undoubtedly, yeah.
+
+**Keenan Wyrobek:** And it's like "Alright, cool. We want to learn all the lessons from that", because we really want the future of the airspace to be similarly democratized and open and innovative and evolving.
+
+**Adam Stacoviak:** Right. So you're studying DNS to learn how to do things in the sky?
+
+**Keenan Wyrobek:** Absolutely. Yup.
+
+**Adam Stacoviak:** Wow.
+
+**Keenan Wyrobek:** And not just the technical layer of DNS, but also like -- if you've heard of ICANN, the international... It's a corporation of names and numbers, I think. Anyway, it's basically the people behind sort of the internet, that agree internationally of how it's going to work. They figured out how to be a group of people that doesn't make bad decisions, you know? And we want to learn from that as well, because I can see -- if you don't get the right incentives and the right people and the right mindset, and the right kind of structure, I think it's really clear how this kind of gets stuck in gridlock, and doesn't evolve and innovate.
+
+**Jerod Santo:** Right. So when you operate on a 2D plane, everybody gets one place. So like my house address leads to one place. Now, there's many paths to that place, so you can alternate routes, and that's all well and good. But on a 3D plane, if you have conflicting paths, how feasible is it to just change the Z axis and say "Well, we're going to fly five feet lower than we were going to, and now we just completely have free airspace"? Is that like foolish thought, or is that legit thought...?
+
+**Keenan Wyrobek:** \[01:29:03.02\] No, that's exactly how our drones don't hit each other.
+
+**Jerod Santo:** Okay. So they're just like "Well, I'll go up or down a little bit."
+
+**Keenan Wyrobek:** Exactly. There's a distributed algorithm, they share their 10 seconds out intent with each other, and there's a distributed algorithm based on what they hear from the people around them, that they'll adjust their course in a way where you get an emergent behavior you're talking about, of like "Okay, great--" You end up with that nice passage, and very efficiently and dynamically. Yeah. I think you're describing exactly how this needs to happen. But the challenge is it has to happen in a very robust way. It can't be like "Oh, if you can reach a server on the cloud, and your cell connection is working, and the server's working, and everybody's talking at the same time, and there's some central orchestration that doesn't have a problem, then it works..." We have a lot of experience to learn the hard way that you just can't scale that way. It's gotta be simpler, and the aircrafts themselves have to have many layers that they can do without any radio connections to things that will still give you layers of safety. That's really important.
+
+**Jerod Santo:** When you bring up layers, it makes me think "Is there an advantage in formalizing around some kind of layering, vertical layering of lanes, so to speak?" ...so you're not just having the emergent behavior of height, but you're actually pre-planning like "Well, I fly at this altitude, and this other company always flies at that, and so we're pretty much cool."
+
+**Keenan Wyrobek:** I think that'll happen. And I think it'll be more than vertical. I think it'll be more like sort of highways or lanes. And the reason -- there's a lot of things we do to optimize for safety, but one thing you'll do is you'll avoid certain areas. The classic example is a stadium. If there's a game happening, you're just not going to fly over that stadium, because why would you? It's not that you're going to have to parachute landing there, but you might, statistically speaking... And flying around it, statistically speaking, is going to be safer.
+
+**Jerod Santo:** It'd be great advertising \[unintelligible 01:30:51.08\]
+
+**Adam Stacoviak:** "Not a crash, just a parachute."
+
+**Jerod Santo:** \[laughs\] Not a crash.
+
+**Keenan Wyrobek:** Yeah. Exactly. So everybody in the air is avoiding similar things, and so you'll end up with this congestion, kind of going around these keepouts, or these soft keepouts, as we call them... And sometimes hard keep outs. So there's certain government places where the government is like "You cannot fly over this area. We consider that to be something you can't fly over." So these hard keepouts and soft keepouts end up funneling their traffic into these places, where I think you'll have -- in the early days, vertical separation might be enough, but in the long term, it may be like "Yeah, if you're going this way, stay to the right, and up high. And if you're going the other way, stay down low and to the left." And you end up with literal kind of corridors that create more and more efficiency, especially when you have these kinds of forced congestions. So there's a lot of cool things that are going to be needed to really scale.
+
+**Jerod Santo:** That's so cool. So cool. Well, we've taken enough of your time... We appreciate you answering all of our questions.
+
+**Adam Stacoviak:** So much fun.
+
+**Jerod Santo:** I think this is one of the coolest things there is. I'm both happy with you, and jealous of you that you get to work on Zipline all the time, because I just feel like there's so many hard problems, there's so much benefit on the other side of it... And so congrats, and good luck, I guess, on scaling out/up.
+
+**Keenan Wyrobek:** Thank you. It's a fun place to work. Just yesterday I came to work a bit earlier, I was on a call in my car, and I was watching everybody walk into the office and everybody kind of has like a skip in their step... And I was just like -- anyway, just reminding myself how lucky I am to work in a place that's like that, where people are fun, the culture is fun, and the technical challenges are real, the customer challenges are real... Yeah, and everybody's skipping to work. It's like a good sign that life is good. So... I'm feeling lucky.
+
+**Jerod Santo:** For sure.
+
+**Keenan Wyrobek:** Yeah. Great to talk with both of you, too. I really enjoyed this conversation.
+
+**Adam Stacoviak:** It was awesome.

Biocomputing on human neurons (Interview)_transcript.txt

@@ -0,0 +1,410 @@
+**Jerod Santo:** Today we're joined by Ewelina Kurtys, or Kurtys... Or Kurtys...
+
+**Ewelina Kurtys:** Ewelina Kurtys. Nice to meet you.
+
+**Jerod Santo:** Yes. Third time's a charm. A scientist-turned-entrepreneur with a PhD in neuroscience, with 20-plus peer-reviewed papers. So you're the real deal, Ewelina. Real deal.
+
+**Adam Stacoviak:** Wow.
+
+**Ewelina Kurtys:** Thank you.
+
+**Jerod Santo:** You're welcome. Thank you for showing up, coming on our show and talking about... Neurons. Neurons. This will be an interesting conversation. I'm a little bit out of my league here, if I'm not going to lie... Because I saw on your website, finalspark.com, on the neural platform page it says "Instant access to human neurons." And I was like, "What does that even mean?" I have no idea what it means. So please, demystify a little bit that, and we can dig into the science as well.
+
+**Ewelina Kurtys:** So that means that our lab is available remotely. So everyone from all over the world can access our laboratory. Through the website browser they can log in and they can write Python code to do experiments, because everything is connected to real neurons. Because we are trying to build computers using living neurons. We want to use neurons as a processor, because they are very energy-efficient... So that's the reason. And at the moment it's still R&D, of course. We don't have these computers yet. But for the moment, it's possible to do experiments only, to try to program neurons. But it's not possible yet to process information like images, or sounds, or videos... But we hope to do this in the future.
+
+**Jerod Santo:** So we recently had Greg Osuri on the show, talking about the AI energy crisis and all these ways that we can potentially power this new compute demand which is burgeoning... And it sounds like those ways were hard, and maybe if we figure this way out, it's way better. Exactly how much more energy-efficient is it to compute on neurons versus silicon?
+
+**Ewelina Kurtys:** So neurons are one million times more energy-efficient. Of course, this is all an estimation, because we can have some idea about this by looking at the human brain, which is built out of neurons. And we can have some idea what will be the efficiency of the processor.
+
+**Jerod Santo:** Okay. So when I think of a platform where you're provided instant access to human neurons - well, I just added the word 'human' in there. Maybe they're not human.
+
+**Ewelina Kurtys:** No, they are human. Absolutely.
+
+**Jerod Santo:** They are human. So I'm thinking about a bunch of brains floating in water, or some sort of formaldehyde...
+
+**Ewelina Kurtys:** No, actually, there are no brains...
+
+**Adam Stacoviak:** No brains. \[laughter\]
+
+**Jerod Santo:** I told you I'm out of my league here. I'm just -- everything science fiction is coming out of me here.
+
+**Ewelina Kurtys:** No, sometimes you can see on social media such pictures of brains enslaved in the lab... But that's not what is happening.
+
+**Jerod Santo:** Okay. Phew.
+
+**Ewelina Kurtys:** We are just using the same building blocks, which are in the brain... But it's like bricks - you can build a house, or you can build something else. So we just use these building blocks, but we don't want to build brains in the lab. We would like to build computers, which would be totally different; probably much, much bigger, because we imagine these neurons can have huge structures in the lab, as we don't have to make it so small as a human brain.
+
+So these are only building blocks. So we are not trying to reproduce a brain. It would be actually very difficult and impossible at this stage, actually, of science, because the brain is very, very complicated. There are a lot of little structures, so we don't try to make this. We just use living neurons. And they are human neurons, indeed, and they are derived from the human skin. So you can reprogram the cells of the skin so that they become stem cells. And from this, you can have, theoretically, any cells you want.
+
+**Jerod Santo:** \[08:01\] Okay. So no brains, but human skin cells, and the neurons that are in them.
+
+**Ewelina Kurtys:** No. Skin cells, which later become neurons.
+
+**Jerod Santo:** Okay, they become neurons. Adam, do you know any of this stuff? I'm over here like a... Imbecile.
+
+**Adam Stacoviak:** Kind of. I mean, what we know about human anatomy and why there is so much curiosity, and why she's studying neuroscience, and how this science fiction era is just simply that our brains can compute so well, with such little power requirements... That's why there's the lore.
+
+**Jerod Santo:** Twenty watts is what I read. Twenty watts to power the human brain...
+
+**Adam Stacoviak:** Right.
+
+**Jerod Santo:** That's nothing.
+
+**Adam Stacoviak:** Very little, comparative to ChatGPT, or something like it.
+
+**Ewelina Kurtys:** Yes.
+
+**Adam Stacoviak:** Dramatically different.
+
+**Ewelina Kurtys:** And to simulate a human brain, you would need a little nuclear plant. So...
+
+**Adam Stacoviak:** So you don't need the full cognitive brain... So -- here's what I understand ,about the brain at least. Tell me if this even maps to the science that you're doing to discover this stuff... It's that you've got this humanity, which is your frontal lobe; that's what helps you have rationale, reasoning etc. If I don't have my frontal lobe, I'm angry Adam. I'm not nice Adam. I don't make good choices, I make very poor choices. How do you get to this level of compute without the full brain? How are these cells able to do so much without what I would typically call like the human brain, I guess?
+
+**Ewelina Kurtys:** Well, so the human brain is actually for many things, not only thinking. It also runs all our body, it controls everything... So that's not always necessary for the computer. What is the most interesting for us is indeed this cortex part, which is responsible for thinking, for processing some abstract information. So we are most interested in this.
+
+So we would like to, in the future, process information through the neurons. Information only. So we don't try, for example, to control a human body, or stuff like this. So there are a lot of things in the brain which are not really related to the biocomputing project.
+
+**Jerod Santo:** So you're effectively using the neurons just as logic gates; you're just doing ones and zeros at the end of the day. They're not doing --
+
+**Ewelina Kurtys:** Well, yes. We would like to try to play, to reproduce the logic gates. However, neurons work totally different. And that's why it's so difficult actually to build the computers, because indeed, in the computer you have zero and ones. And this is one of the reasons why actually they use so much energy. But the brain is encoding information totally differently in space and time... So when we have neurons in our head, it matters when and where (exact location) they are active
+
+And this is information. So this is a totally different type of encoding. So no zero/ones, actually, but there are a lot of ways how we can look at the activity of the brain. For example, how often you have spikes, or what are the time in between the spikes... So this electrical activity of the neurons. So we know for sure it's totally different. And that's why this project is so difficult, because we have to figure out a totally new way of programming, a totally new approach. It's the same actually as in quantum computing. It's the same situation, that you have totally different hardware, which is working differently... So it's necessary to figure out a new way of writing algorithms. And that's why it's so difficult, because actually, someone has to come up with some idea which would be totally, totally different.
+
+But indeed, at the moment, when we do research on living neurons or sometimes on some simulations of neurons in silico, people usually try to follow the rules of digital computers, like try to reproduce logic gates... Which is actually not really correct, but that's the best what we can do at the moment.
+
+**Adam Stacoviak:** \[12:09\] Let me see if I understand this... I'm grokking some of the stuff from what you're sharing, and then also from your very awesome website, finalspark.com. It says they transform stem cells into mini brains, that learn and adapt, growing neurons in an orbital shaker - which I have no idea what that is. It sounds so cool...
+
+**Jerod Santo:** It sounds cool, yeah.
+
+**Adam Stacoviak:** ...over a three-month period. And these mini brains, organoids - I'm not sure if that's a term you all came up with or not, but that sounds cool, too...
+
+**Jerod Santo:** It sounds \[unintelligible 00:12:37.09\]
+
+**Adam Stacoviak:** ... of 0.5 millimeters in size, with about 10,000 neurons that function as real brain tissue. So you've found a way to take stem cells, grow them over a three-month period... They used to have a half-life that was even shorter, like a few hours. Now you can actually -- they'll live 100 days. And they get connected to this neural platform with 24-7 access. So essentially the same way we treat a CPU in AWS, you're doing with stem cells, turned neurons, turned mini brains, turned organoids... That can be compute platform.
+
+**Ewelina Kurtys:** Yes, absolutely. Although now it's for experiments, so we cannot really process information the same way as in digital. But yes, it's available remotely, and we imagine that actually in the future our lab or our biocomputer will be available remotely as a cloud service today.
+
+**Jerod Santo:** Right. Not quite AWS yet, but working your way there.
+
+**Ewelina Kurtys:** Yes, absolutely.
+
+**Jerod Santo:** How did you get involved in this? Where are you coming from?
+
+**Ewelina Kurtys:** So I come from Poland. I was always on the medical side, let's say. I studied pharmacy and biotechnology, and I always wanted to be a scientist. I enjoy a lot working in the lab. I was very fascinated by cracking my brain, teasing my brain with some ideas and challenges... So I always wanted to be a scientist, and I realized at some point that the brain is the most interesting part to study... So I did a PhD in neuroscience, I was working on brain imaging... And later, when I moved to industry - because I always wanted also to see what is outside Academia, outside this academic world... I started to work with startups on actually -- initially on imaging. And then there is a lot of AI in the medical imaging, in industry, so this is how I learned about AI. And then I became fascinated by that, and I started to discover which opportunities it brings, beyond imaging. So I started to work on commercial applications of artificial intelligence, and after, I started to work on the next frontier of AI, which is actually closely related to neuroscience, so on biocomputers.
+
+**Adam Stacoviak:** Can you talk about the imaging? I think you mean -- when you say imaging, you're probably referring to like MRIs, like brain scans... Is that right?
+
+**Ewelina Kurtys:** Yes. So actually, I did my research on positron emission tomography. This is something you do using radioactivity. You put some radioactive substance in the body, and this substance goes to some specific places in the body, and you can detect this non-invasively. So you can get the picture, for example, of the brain, which parts are active, for example... Or you can visualize some receptors without opening the brain.
+
+**Adam Stacoviak:** Right.
+
+**Ewelina Kurtys:** But it's actually similar to MRI. MRI is just a little bit different... So you don't use the radioactivity, and you can see a little bit different things. But the idea is always the same, to look inside without opening the body.
+
+**Adam Stacoviak:** \[15:58\] Right. Yeah, the MRIs are a little different... Which one's more accurate? Is the imaging or the MRI more accurate? Does it matter?
+
+**Ewelina Kurtys:** Always imaging. I'm not sure it's a good comparison, because I think it really depends on the protocol... Because there are different types of MRI, and also there are different types of PET, so it really depends. It depends on the parameters. And also, they measure different things. Because PET is always functional. So when you use the radioactivity, there is always some chemical substance, like even glucose. Everything is actually a chemical substance. Everything that is flowing in the body. So you always observe some process, a biological process. And in MRI, it's not always like this. Sometimes you just observe the tissue, when you have different tissues, and it's static. It's not always functional. It's not always \[unintelligible 00:16:54.09\]
+
+**Adam Stacoviak:** Right. This imaging -- that's why I asked this question, because this imaging is really kind of like the rage, I would say...
+
+And my version of the rage may be way different than your scientific version of the rage... But what I mean by that is that there's a lot of study around mental health, ADHD, ADD, trauma - you name it - that folks are trying to image brains in these scenarios. Is that kind of what got you into this curiosity of like how the brain operates from different trauma levels, or different prescriptions, or descriptions of health concerns, mental health concerns, whatever it might be? Is that what got you interested in this imaging process to understand more clearly how the brain reacts to, I suppose, life?
+
+**Ewelina Kurtys:** Well, actually, I was working on something a bit different to what you're talking about... Because you say about different activities of the brain, during different maybe diseases, or maybe different tasks, cognitive tasks... But I was actually working more on inflammation. So I try to visualize microglia. Microglia are a type of cells which are around neurons in the brain. So they actually take care of the neurons. And sometimes they become very activated, which means inflammation... And it's believed that this process is actually involved in many neurodegenerative diseases, and also depression. So when you have an inflamed brain, you can develop some disease.
+
+**Adam Stacoviak:** Yeah. Alzheimer's... That's interesting. Yeah. Inflammation is like the number one issue for most people.
+
+**Ewelina Kurtys:** Yes.
+
+**Adam Stacoviak:** They get inflamed everywhere.
+
+**Ewelina Kurtys:** Yes, my research was actually about the effect of nutrition on inflammation... And it's funny, because I started to put attention on everything, what I eat, after I started to do this research. It's really interesting, because I started to look totally different on my groceries... Because actually, diet can be pro-inflammatory or anti-inflammatory. It's very important what you eat. And it can affect also your brain health.
+
+And I think now, after -- it's maybe almost 10 years since I did these studies... Now I see that there is more and more -- more and more people are talking about this. About an anti-inflammatory diet, about how much is important, what you eat, for also your brain health.
+
+**Adam Stacoviak:** How did that lead you into discovering -- I mean, it kind of seems obvious, but how did that lead into AI and your discovery there? Were you leveraging trained models? Express how you got curious about AI.
+
+**Ewelina Kurtys:** Actually, I started my first job in industry. I started in the company which was doing medical imaging. And actually, it was a very good start, because there was at least one thing which I understood at the time... Because I had absolutely no idea about how companies work, and anything about this industry world. So there was at least one topic which I understood well, which was medical imaging.
+
+\[19:56\] And this company was doing a service of analyzing images from different medical studies, and they talked a lot about AI. Because when you have imaging data, when you have a high number of imaging data, you can analyze them automatically. In some way, you can use AI for that. And actually, that's the way how I learned about artificial intelligence. Because actually, when you go for different industry events, you see people talking constantly about AI... And I was very lucky also because at that time I was in London; this is a very good place for learning new stuff and for networking. So I could get a lot of exposure and to see what people are doing with AI. And I could discover that there is much more beyond imaging. So that's why I started to be interested with anything, what you can do with AI.
+
+**Adam Stacoviak:** Well, being able to scan a lot more - no pun intended really, but just grasp a lot more of these imagings that you're doing, to see the anomalies and see the connection points that you can't really see individually... I mean, that totally maps to me, because the more you can see across different scans is good.
+
+**Ewelina Kurtys:** Yeah, this is a general thing about AI, because it can see much more than us, and can scan a lot in a very short time.
+
+**Jerod Santo:** So how long have you been working on this problem?
+
+**Ewelina Kurtys:** On FinalSpark I met the founders in 2019, at the conference in London, so I started to work with them, initially on some other projects... So actually, on FinalSpark I could say I'm working like three years, around.
+
+**Jerod Santo:** Okay. And there's a platform right now for experiments... You're hoping to get to compute down the road, and a service for that... Is there a straightforward path towards that, or are there like breakthroughs that still need to happen to get from where you are right now to where you guys want to go?
+
+**Ewelina Kurtys:** No, it's a very difficult, very challenging project. That's why we expected to build these real computers in around 10 years. So it's a bit of a challenge when we talk with potential investors... Because it's quite a long-term project, and it's very, very difficult. Because nobody knows how really neurons encode information. So this is the biggest challenge. So we know that neurons are active electrically, we know they are spiking... Spike means that there is electrical activity. And we know quite a lot about this, how it happens. However, we cannot really translate this into some specific information. So for example, you have text or image... Maybe about imaging there is some understanding already in neuroscience, but for example when you have words, text, it's hard to say how some word can translate to specific activity of a neuron.
+
+So at the moment, as I said, many people do a lot of random experiments. Also us, we do a lot of trial and error. So this is why actually we built an automated laboratory. Initially, the idea was to just be able to do as many experiments as we can. And also, a lot of research on neurons are often inspired by what happens in the digital world, which is not really correct, because neurons are working totally differently... But it's still, at the moment, the best you can do. So this is the biggest challenge, that we don't really know what activity of neurons mean.
+
+And also, another thing very important is that brain, or neurons, any kind of form of - also our neurons in the lab - they are not stable systems. So a computer you can consider as a stable system. It's a dead matter, so it works today in some way. Tomorrow it will work the same way. But the living tissue is not like this. It can change, so the dynamic inside can change.
+
+\[23:54\] So for example, today we do some experiments, we send electrical signals to neurons and they can react in one way, and tomorrow they can react to the same signal totally differently. So that's also a big challenge. The fact that living matter is plastic, so it changes behavior, actually also like us and our brains. We also change. During time we can be completely different people.
+
+**Adam Stacoviak:** Can you talk about how you get them to compute?
+
+**Ewelina Kurtys:** This is actually a challenge. So what we do - we try to send them electrical signals. Because neurons are placed on the electrodes - you can see this on our website, finalspark.com. There is a section Live, you can see the readout. So we send them electrical signals and we measure how they respond. So how they change the activity. How they change electrical activity. Another thing how we try to compute neurons is also by sending them some chemical signals. At the moment, we can send them dopamine or serotonin. So programmatically, we can program in Python that neurons will get dopamine at some point. So at the moment, this programming is not really to do some specific task as with computers, but actually to change the behavior of neurons. So this is actually the first step. So we want to be able to consistently control how neurons behave. So how is the electrical activity of neurons.
+
+**Adam Stacoviak:** So you may have one lazy organoid and one very non-lazy organoid.
+
+**Jerod Santo:** Productive. Yeah.
+
+**Ewelina Kurtys:** Yes, absolutely. It is biological tissue. Sometimes it can vary. Sometimes they can just die. So we are still learning... And yes, so there is also variability. Yes, absolutely. A lot of things. It can be also a lazy organoid.
+
+**Jerod Santo:** Yeah. Or depending on the day. Like, yesterday it was really productive and then today it's lazy.
+
+**Ewelina Kurtys:** Yes. As I said, every day is dynamic. It's not a stable system. However, we have some success. It's not only so bad. We were able to store one bit of information. So just to give you an idea about the stage at which we are - we stored one bit of information in neurons. That was quite consistent. And we were able to reproduce this many times... So we are happy there is some kind of progress. But yes, it's very challenging to get something.
+
+**Jerod Santo:** So when you store a bit of information...
+
+**Ewelina Kurtys:** Yes.
+
+**Jerod Santo:** How do you read it back out again? Or how do you get it back? Or how do you know that it's stored?
+
+**Ewelina Kurtys:** Yes. So actually, that's quite technical, because I can tell you every blob of cells - because there are such a blobs, these neurospheres, organoid, they are 3D structures... And each is placed on the eight electrodes. And all these eight electrodes, they measure activity from neurons. And depending on how strong the activity is at each electrode, you can mathematically calculate something, what is called center of activity. So this is quite a hard science approach... And we were able to shift the center of activity. So yes/no. That was one bit of information. And this is quite complex, as you can see... But yes, we were able to have this consistently, this kind of results.
+
+**Jerod Santo:** Consistently across different neurons and across different times?
+
+**Ewelina Kurtys:** Yes. And different days. Because that's always -- in general, in bioscience, every time when you work with biological tissue, this is important, that you have to be able to repeat. Because many things work once or twice, but it's important to be able to repeat your results. Yes, on different days, on different neurons.
+
+**Jerod Santo:** Mm-hm. And is the process slow?
+
+**Ewelina Kurtys:** \[27:59\] Well, to be honest, I cannot tell you how much time it took... No, actually, I don't know. But I guess it's in seconds or milliseconds. But I don't know. But I can say that, in general, neurons are slow, and in general, neurons will be good for tasks which don't have to be fast. Because also, when we look at the human brain, and we look at computers, we can see that computers are very good in speed. In doing repetitive things - very, very fast. And we will never be able to compete with digital on that. However, the brain is better in complex tasks, because we can solve complex problems using very little energy. So that's where is our strength. But definitely, speed or, for example, memory is not something where neurons are better. Because also, when we look at our brain, they are very limited, actually. A computer can remember 20 books very easily, and for us, it would be difficult to remember every word in 20 books. But for the computer it's easy.
+
+**Jerod Santo:** What does this wetware look like? I'm programming my Python experiment, and I'm sending it over the internet, I suppose, or some sort of VPN connection to you guys...
+**Ewelina Kurtys:** Absolutely.
+
+**Jerod Santo:** And so, of course, it's going over copper wires, and Wi-Fi, and whatever backbones, and then back into your interface, which eventually translates it into... I'm imagining there's a needle at the end of a thing that like sprays some dopamine... I don't know what happens. What happens at the end, the last mile of this API?
+
+**Ewelina Kurtys:** Yes. So when you send an electrical signal, you have a digital to analog converter. So you have to translate the things from digital world to analog world, because neurons are analog.
+
+**Jerod Santo:** Right.
+
+**Ewelina Kurtys:** So you have this digital to analog converter, and it is translated into electrical signals, which goes through the electrodes. So basically, the stream of electrons are going, flowing through the electrodes. And when you want to send a signal with dopamine or serotonin, then it's connected to the lamp. So we have an UV lamp.
+
+**Jerod Santo:** Okay.
+
+**Ewelina Kurtys:** And when the UV light is open, the dopamine is released... Because dopamine is closed chemically. It's encaged in the chemical, so that it's not active. But when it sees the UV light, then it's released.
+
+**Jerod Santo:** Oh...
+
+**Ewelina Kurtys:** So it's a way of releasing very, very quickly dopamine to the neurons. So this is how it works. So then it's connected to some controllers, which are connected to the lamp, and then the lamp switch on.
+
+**Jerod Santo:** Okay. Same thing for serotonin?
+
+**Ewelina Kurtys:** Yes. But for serotonin, we have different wavelengths. I don't remember which one. It's not UV. But then we have different wavelengths, yes... So that they don't overlap. So you can have both in the medium. And of course, we plan to have more of this, and of course, much more neurotransmitters... But we started with dopamine, and now we added serotonin.
+
+**Adam Stacoviak:** Are those hormones? Are those chemicals? What is the proper terminology to call dopamine and serotonin?
+
+**Ewelina Kurtys:** No, they're neurotransmitters. So you have several in the brain, and they affect learning. And actually, the whole idea of using them is because we use them for feedback. Because, actually, the way how humans are learning is by feedback. So you have interaction with the environment, you get feedback. So things are going good or bad, and then you learn if you should do this or not. And the same way, actually, neurons are learning in vitro, on the very basic level. Because for example, when something good happens, then there is a dopamine release, and that reinforces the connection between neurons. So if they have done something good, then it kind of reinforces this behavior.
+
+\[32:02\] So the idea - at least our idea here, because it's a bit complicated... And actually, there are different opinions about how to give punishment or reward to neurons... But our idea is to give dopamine as a reward. And no dopamine as a punishment. So that's used for the feedback loop. So for example, you stimulate neurons with some electrical signals, you measure the behavior... For example, you want that they increase activity. So if they do this, you give them dopamine. If they don't do this, you do nothing. And then you send an electrical signal again, and there is such a loop, over and over, and you see if they're learning.
+
+**Adam Stacoviak:** And these neurons learn.
+
+**Ewelina Kurtys:** Well, that's the problem - sometimes they learn, sometimes they don't. This is still a challenge, you know. This is still a challenge, learning. So learning for neurons is changing the connections, changing the behavior of neurons, so behavior will be electrical activity... And this is still a challenge. It doesn't always work.
+
+**Break**: \[33:04\]
+
+**Adam Stacoviak:** Remind me... Dopamine is positive, so that's used for rewards... What is serotonin used for? What do the two levers do?
+
+**Ewelina Kurtys:** No, actually, that would also be for the --
+
+**Jerod Santo:** Two different versions of reward.
+
+**Ewelina Kurtys:** Yes. However, if we have to go to the details, it's a little bit tricky... So we are still --
+
+**Adam Stacoviak:** Okay... Let's go into the details. Let's get tricky.
+
+**Ewelina Kurtys:** \[36:11\] Yes. Because actually, dopamine is -- it also depends when it is given, and also there are different receptors. So receptors are on the surface of the cells. So if they have dopamine receptors, that means they can recognize dopamine, actually, because they need always a receptor to recognize a neurotransmitter... So it's a little bit tricky, because there are different types of receptors, and different timing... Sometimes it's milliseconds, or microseconds... So the timing also is important for cells. But we assume that dopamine is a reward.
+
+**Adam Stacoviak:** So it's not stable yet.
+
+**Ewelina Kurtys:** No.
+
+**Adam Stacoviak:** Some days they do, and some days they don't... And you're learning. So this is 10 years before it's usable in production. This is total lab learning... What is it that -- and maybe you're still early and you can't answer this question, but what is it that makes it such a variable? Is it just because it's bio, and we don't know what we don't know?
+
+**Ewelina Kurtys:** Yes, because it's bio. Because first -- there are two reasons, I would say, main. First is because it's bio, so it's unstable. Second is because nobody knows yet how neurons encode information. This is totally different than digital. So this is such a challenge, because you have to understand a new way of programming.
+
+**Jerod Santo:** But you do have some indicators that they at least generally work the same.
+
+**Ewelina Kurtys:** Well, you mean neurons?
+
+**Jerod Santo:** Yeah. Well, if each neuron was like a snowflake - you know, every snowflake is unique, and it melts, so it changes... Then there really would be no -- like, 10 years, 100 years, a billion years, there would be no getting there, because there's no determinism at all. Because everyone could just work completely different every time you prod it, you could never get information -- but you've actually gotten a bit back out again, so you have proven...
+
+**Ewelina Kurtys:** You know, I think it is deterministic. It's just that we don't know the rules yet.
+
+**Jerod Santo:** Right. That's why I'm saying, you do have an indicator that they do kind of work the same generally, though... At least for this one thing.
+
+**Ewelina Kurtys:** Yes. But the indicator is our brains, actually... Because we have no doubts that neurons can process information very well. This is why we can talk.
+
+**Jerod Santo:** True.
+
+**Ewelina Kurtys:** So we can say nature is a proof that neurons are working.
+
+**Jerod Santo:** That's fair.
+
+**Ewelina Kurtys:** There is no doubt about this.
+
+**Jerod Santo:** We think differently, but we all think, I guess...
+
+**Ewelina Kurtys:** We just have to learn how to program them.
+
+**Jerod Santo:** Have you tried telling them to ultrathink? I'm sorry, that was a joke...
+**Adam Stacoviak:** \[laughs\] Good throwback.
+
+**Ewelina Kurtys:** That was from a previous show.
+
+**Adam Stacoviak:** I have been using that, by the way... I've now said -- sorry for a slight aside... "Triple check and ultrathink." That's my new phrase.
+
+**Jerod Santo:** That's the key word, Ewelina, for certain AIs...
+
+**Adam Stacoviak:** Triple check that stuff and ultrathink.
+
+**Jerod Santo:** When they're not thinking best... But of course, with a neuron, maybe you just keep it analog and just whisper to it. "Ultrathink...!" Just walk up to it and whisper.
+
+**Ewelina Kurtys:** Yes, you can whisper, but they don't have ears, so they can only understand --
+
+**Jerod Santo:** Well, you need some ear cells. Get some ear cells going.
+
+**Ewelina Kurtys:** You have to translate, because usually our ears also translate to electrical signals.
+
+**Jerod Santo:** True.
+
+**Ewelina Kurtys:** That's why our neurons in the head can understand. So you have to learn. That's the whole point - how to encode information so that they can understand.
+
+**Jerod Santo:** Yeah, so you guys are just running -- I imagine you're just running experiments nonstop, right? Because you're trying to figure out how these things work.
+
+**Ewelina Kurtys:** Yes, absolutely. And we are actually constantly building, because we have done huge progress since we started. We built a whole laboratory, a very stable system for working on neurons. And also, now it's available remotely, so we are also busy with many users from all over the world... So we invited nine universities from different countries to work with us. They have access to our lab for free, to study also neurons. And we also have first industry clients who pay us to get access to our lab. So we are also busy with this.
+
+**Jerod Santo:** \[40:13\] Really?
+
+**Ewelina Kurtys:** Yeah. We didn't plan for this, but people started to write to us that they would like to try, that they would like to get access to the lab... And yes, now we have two types of subscription, and we have users who are coming to us and testing neurons.
+
+**Adam Stacoviak:** Is this the Betamax versus VHS all over again, in terms of quantum computing versus - would you call this bioprocessors? How would you frame this? Because it seems like you're both trying to solve a similar problem.
+
+**Ewelina Kurtys:** Bioprocessor, very good. Or biocomputing. We call it biocomputing, bioprocessor... So no, I wouldn't say it's in competition, because this is a totally different mechanism, different things. So we know quantum computing actually is very fast, and it can maybe be good for encryption of information... So it is a totally different type of task. So I'm not sure it will be in competition, but...
+
+**Adam Stacoviak:** Okay. I was thinking more like one may win, or one may actually prove to be fruitful in terms of viability... That's kind of what I was thinking like.
+
+**Ewelina Kurtys:** Yeah, actually, I think that the future will be that we will have very different type of hardware, because generally, you can see this kind of direction. It's not only quantum, not only biocomputing. People are also working on many specific chips, also digital, which are optimized for some specific tasks... So I believe that we will have variety. So today we have mostly CPU, GPU, and in the future we will have hundreds maybe of different chips, I believe so, which will be optimized on some specific task.
+
+**Adam Stacoviak:** Jerod, I think this was on a \#define where we talked about this, but do you recall talking about slime molds, and subway systems?
+
+**Jerod Santo:** Yeah, like a couple of years ago.
+
+**Adam Stacoviak:** Yeah, just this really -- it was recent. I want to say in the last year... We were talking about the concept of slime molds being very sophisticated.
+
+**Jerod Santo:** They use slime to design the subway systems, or something?
+
+**Adam Stacoviak:** Right. Routing, essentially. Like, efficient pathways to X. And they compare that to subway systems, and the way we route, which is more like cause and effect, really. We're very reactive... But it's very similar in terms of like bio. You've got this intelligence of sorts. Not intelligence like it's got a body and it can come fight you. Slime is not gonna do that. But it can do its own growth mechanisms... And I'm not a slime expert, so I'm not trying to pretend... But just being enamored by the fact that there's some level of intelligence in slime that can predict maps. Just this idea of bioinformatics, biointelligence that can supply this... In this case it's obviously a neuron that can provide feedback, and computing, and stuff like that, but very similar in nature in terms of trying to leverage intelligence built into nature, the world around us.
+
+**Ewelina Kurtys:** Yeah, I think people are also inspired by insects or different biological things...
+
+**Jerod Santo:** Yeah, why not.
+
+**Ewelina Kurtys:** ...in computation. Yes, there are such projects also.
+
+**Jerod Santo:** Ewelina, what industries are interested in this? You said you have these users all of a sudden... Which industries want this as a thing?
+
+**Ewelina Kurtys:** I would say we have three types of users. Individuals/fascinated engineers, or small startups. Some of them want to do something related to biocomputing... That's why they want to use our platform. And big companies, which have R&D teams; very large companies which have R&D teams which want to do some project on cutting edge technology. The same actually how people do with quantum. They know that it doesn't work yet, but they want to know what is going on, they want to know how it works, because they believe it will work in the future.
+
+**Jerod Santo:** \[44:16\] Okay, so that's cool... What kind of stuff are they trying to do? You don't have to give specific examples or anything like that, but...
+
+**Ewelina Kurtys:** No, actually this is confidential. What our clients are doing is confidential.
+
+**Jerod Santo:** Sure.
+
+**Ewelina Kurtys:** But we have universities which are using our lab for free, and they are going to publish. So actually, that's why we chose them. We chose those who have the highest chance to publish.
+
+**Jerod Santo:** Sure, that makes sense.
+
+**Ewelina Kurtys:** And actually, there will be some papers coming for what people are doing, so I hope everyone will be able to see.
+
+**Jerod Santo:** That's cool.
+
+**Ewelina Kurtys:** And we will be promoting this, for sure.
+
+**Jerod Santo:** What's at the other end of my Python API call? So we talked about what was at the neuron platform end... Like, a UV light turns on, or some sort of electrode electrolyzes... What do I get back? Like, I make a call... Is it like a one/zero? Is it like a success/fail? Is there more information coming back to me? Like, what do I get back at the other side so I can actually start mapping results or trying to make sense of it?
+
+**Ewelina Kurtys:** Yes. So what you get in response is the electrical activity of neurons. So this is what you can see also in our website, on the live section. So the way how you can measure activity of neurons is a few different ways. You can get a yes/no response. So this is spike trains. This kind of data you get just a dot, and you know this was spike. Every time there was a spike, you get a dot. And there is already quite a lot... You can analyze the patterns, you can see if they're more active or less active... This is actually the most common way how you collect the data. And it's quite efficient also, because you just have one dot, one point for each occurrence of the spike. And very often it can be enough. But if you want to be more specific, you can also measure the shape of the spike. Because spike means that the neuron will change the charge, and this will always have a shape. And you can also analyze the shape of the spikes. So this is much more heavy data, but you can also get this.
+
+And of course then you can have -- you know, people try to have different ways of interpreting the data. This is actually -- it's a big room for creativity. For the moment, we look for example how late, what was the delay before we saw the signal... Or we can see the distance between the signal, for example, how often a neuron is active. So we try to characterize all these patterns on how they are active. So that's what you get... And then you can -- yeah, there is a lot of signal processing, a lot of analysis of the data.
+
+**Jerod Santo:** Sure.
+
+**Ewelina Kurtys:** And a lot of data.
+**Jerod Santo:** Can you target a specific neuron or organoid, to like make sure that your call goes to the same place every time, or no?
+
+**Ewelina Kurtys:** No. Actually, you have eight electrodes. So every electrode is in a little bit different place of the organoid. And then you can target specific electrodes. You can, for example, use only a few of them, or you can use, for example, four of them for sending signals, and four of them for receiving signal... Or some other combination. So that gives you some room for playing.
+
+\[47:57\] And what is also interesting is not every electrode is always active, because sometimes you might have less signal, or no signal at some of the electrodes. So it's really complicated. It's very difficult to work with a living tissue. Sometimes they are just not active also.
+
+**Adam Stacoviak:** So they have -- how do you know when they're about to die? You mentioned inefficiency, you mentioned one day doesn't work the same as the next... We know earlier in your research they would die in hours, now they die in hundreds of days, I think... Help me understand terminology, is that right, and how do you know the inefficiencies aren't because they're about to die? Again, I don't know if that's the right terminology to use or not.
+
+**Ewelina Kurtys:** So there is at least one thing which is easy... So this is easy, to see if they die or not, because they are not active. So living neurons - they are spontaneously active electrically, so they will always produce some spikes, and you will see them on the electrodes, on the measurements.
+
+**Adam Stacoviak:** Okay.
+
+**Ewelina Kurtys:** So this is quite easy to say that they are dead. If there is no activity, you assume they are dead.
+
+**Adam Stacoviak:** Okay.
+
+**Ewelina Kurtys:** And actually, you are right, because also batches are different. And this is also what you can see on our website... Because a few of our neurospheres are monitored there. You can see that the activity is not always the same. Sometimes it's active, sometimes less active... So all this you can see very easily on the electrodes, when you measure the activity.
+
+**Adam Stacoviak:** Yeah. One thing I think is interesting too is the environment it has to live in, which - we talked a little bit about quantum computing, and then compared it to biocomputing... That there has to be a sterile environment, no viruses. Can you talk -- I know you are in a lab, or at least early days of research and stuff, but what is the environment? And how will that potentially scale to usable product at the long tail of usage? What is the environment these things live in?
+
+**Ewelina Kurtys:** Yes, so environment is very important. Neurons are very fragile. And the environment has to be physiological, so the same as in our bodies. So there has to be physiological temperature, there has to be, of course, always liquid around... Neurons are in the medium, so this is water with the different substances which keep them alive, which also feed them... And all this is very, very important. pH, temperature... Everything. Even small vibration. Everything is really important for the neurons to be stable. And it has to be very strict, otherwise the activity can change or they can die.
+
+And this is why also we believe in these bioservers, in the central servers idea, because we think that it will be easier to control these conditions of the neurons when they will be in a central server. So that's also the reason why we --
+
+**Adam Stacoviak:** So not likely to have a home version of this in the early stages of this. Like, you want to centralize it at some sort of data center, or a space where the environment can be better controlled.
+
+**Ewelina Kurtys:** Yes, absolutely. And we imagine we have the same what we have today, but much bigger.
+
+**Jerod Santo:** How big is what you have today?
+
+**Ewelina Kurtys:** Well, now we have two rooms for the laboratory, so we are growing... We started with one, a little lab. And our neurospheres are a few millimeters diameter. 10,000 neurons each. So they are very, very small, but for experiments it's enough. And in the future we imagine to have huge structures, even 100 meters long of neurons. So that's how we imagine the future. It'll be much, much bigger.
+
+**Jerod Santo:** Yeah. I have so many questions about the details of that, but you can't really ask them until you guys know how they work exactly... Because I think a lot of the decisions will be based on how they work. Like, how many neurons will I need to do a thing? And it's like "Well, we don't know, because we don't know how they work exactly yet."
+
+**Ewelina Kurtys:** Yes.
+
+**Adam Stacoviak:** \[52:17\] It seems like it's going to be exotic use cases. And I imagine, as somebody who's been at the doctorate level, the PhD level of this, from neuroscience to this laboratory stage, that you probably see at least some very exotic use cases. It has to have a unique environment, you plan to centralize it to offset that... But I'm sure there's unique scenarios where like this may be finely tuned or very specific to a certain type of task, versus general computing. It's not going to be in my iPhone. Maybe at that point it will be an iPhone, a literal iPhone... Anyways. Are there any unique, exotic scenarios or use cases that you already see, even though you're in the science stage, where this may apply?
+
+**Ewelina Kurtys:** Actually, we aim for general computing. But of course, not everything. By looking at the human brain and also thinking about what is done now in digital, we think that every -- maybe not every, but many tasks which are done by artificial neural networks will be much better to be done on biological neural networks. For example generative AI we believe could be better on the real neurons.
+
+**Adam Stacoviak:** Really? Okay. So maybe that's the first place where -- because we've started off talking about our energy crisis that is obvious; that seems to be the obvious reason why biocomputing is the platform... It's potentially a lot less required energy usage.
+
+**Ewelina Kurtys:** Absolutely.
+
+**Adam Stacoviak:** And so you're going general.
+
+**Ewelina Kurtys:** Yes, yes. We go for general computing, which will be much, much cheaper, and very competitive to digital.
+
+**Jerod Santo:** Are any of your customers well-capitalized evil geniuses who just want to electrocute some neurons because they're just enjoying...
+
+**Adam Stacoviak:** Gru. Maybe Gru is --
+
+**Jerod Santo:** Like a Doofenshmirtz, or maybe like a Moriarty... Any Moriartys? I'm just messing with you, Ewelina. I've just run out of actual questions... Adam, anything else for her? I mean, this is interesting stuff. I think we're definitely -- a lot of work to be done.
+
+**Adam Stacoviak:** Yeah, I think it's really about the -- I was just thinking like where could it be used, where do you see it being used... I'm surprised at general, because it seems like it's -- that's the long road. Like, the short-term road would be specialized use cases, where you can control the environment, have potentially really rich clients that have blank checks that can give you four rooms versus two, kind of thing... I'm thinking like that versus general computing, but I guess I was wrong.
+
+**Ewelina Kurtys:** No. Actually, we aim for general computing. We think general computing could be a real revolution.
+
+**Adam Stacoviak:** Post-silicon, Jerod. Post-silicon.
+
+**Jerod Santo:** Biocomputing human neurons.
+
+**Adam Stacoviak:** Wow. I'm looking forward to it. I can't wait to see what happens next. I'm so surprised by what's happening today...
+
+**Jerod Santo:** Yeah, I had no idea.
+
+**Adam Stacoviak:** ...let alone what the future may hold from this. So cool.
+
+**Jerod Santo:** Thanks, Ewelina. Thanks for coming on the show and telling us all about it.
+
+**Ewelina Kurtys:** Thank you so much for the nice questions and nice discussion.

Bringing Atuin to the desktop (Interview)_transcript.txt

@@ -0,0 +1,390 @@
+**Jerod Santo:** Well, I am joined once again by Ellie Huxtable with Atuin, which has been making my shell magical for years now... Ellie, welcome back to The Changelog.
+
+**Ellie Huxtable:** Hi, great to be here.
+
+**Jerod Santo:** Great to have you. You sound closer. You were so far away last time... I think we had a schedule around our different time zones... But now you're in the States. What happened? What changed?
+
+**Ellie Huxtable:** I moved to San Francisco. That's been very recent, just like a month ago, but it's been good so far.
+
+**Jerod Santo:** Good so far. And where were you before? Somewhere in the UK, right?
+
+**Ellie Huxtable:** I was in London.
+
+**Jerod Santo:** In London. Okay. And you were often on a motorcycle, I remember that part...
+
+**Ellie Huxtable:** All the time, yeah. Still am. I bought one like four days after moving here, so...
+
+**Jerod Santo:** Okay... Did you upgrade? Did you downgrade? Did you side grade?
+
+**Ellie Huxtable:** Side grade. I have a 1290 Super-Duke now. The roads here are more fun, and the weather's more amenable to motorcycling, so it's a good change.
+
+**Jerod Santo:** Okay. And there in SF, it's got to be -- you're in the San Francisco area, is that what you said?
+
+**Ellie Huxtable:** Yeah, yeah.
+
+**Jerod Santo:** Lots of hills.
+
+**Ellie Huxtable:** Yeah. They're fun, too.
+
+**Jerod Santo:** Fun a.k.a. dangerous, depending on your appetite for destruction, I guess...
+
+**Ellie Huxtable:** Pretty much.
+
+**Jerod Santo:** So Atuin - such a cool tool. Very few times do I cover a tool in Changelog News and then install it and then use it, and it integrates itself into my life... And it's been years now. So I just use it happily, Control+R for the win. I love it. I don't use probably any of the fancy stuff that you've added since then, but that was like your deal originally, was Atuin the CLI, which is like synchronized and prettified and just better, in my opinion, shell history. And the sync was really a big part of what you were building, which is the part that I don't care about. I'm a one-computer person.
+
+**Ellie Huxtable:** You can use it to back up as well. It's pretty good for that.
+
+**Jerod Santo:** Yeah, that's true. And you had some cool stuff with how you're handling sync, and how you do that without violating privacy, and there was all kinds of talk about that... And then you continue to work on it, and now you're working on something that's very related and very cool and different, which is Atuin Desktop. Can you take us on the journey from when we last spoke - which was about 18 months ago - on the show about Atuin, the CLI, and where you went from there, what you built?
+
+**Ellie Huxtable:** Yeah, sure. So the CLI records every shell command you run, is kind of the whole point. It makes it really easy to recall them. But what is trickier with that is sort of recording whole workflows. So if you're looking for an individual command, it's great. It's really easy. But if I was looking for "How do I set up a new staging environment? How do I onboard a new developer? How do I do something that's a long sequence of steps, and maybe needs annotations and various other things?", the CLI helped, but didn't solve the problem. We had a whole bunch of discussion over years, to be honest. I've had this idea for a really long time, of like "How can we solve this essentially documentation problem?" And Atuin Desktop is the answer there, at least in my opinion.
+
+\[07:50\] So what it is is it's essentially a runbook editor. What I mean by that is we started off by just building something that was good for recording notes and documentation. You can just write English words down, and share them with people... But where it gets really powerful is the executable blocks we added. So we have things like terminal blocks, which are - if you imagine Notion with its database blocks, instead of a database, we have an embedded terminal that you can actually execute a command. And this sort of takes it from a confluence page that goes stale, to something that you can actually run and read in the same place.
+
+A lot of people have asked how this is relevant to the CLI, and I can kind of see that it might not be immediately obvious. However, where it does become quite powerful with the CLI is all of the data you've been recording in your case over years in the CLI, is accessible in the Desktop app, too. So you can screw around on your machine, figure something out locally in your terminal, and then when you go to document it later on, you can write down what you did and you can pull in your shell history into the executable blocks.
+
+**Jerod Santo:** Yeah, that's really cool. So if I recall correctly, you had full-time jobs at one point, which was at -- I remember PostHog. What was it, before? Was it Coinbase?
+
+**Ellie Huxtable:** Coinbase before that.
+
+**Jerod Santo:** Okay, so a couple of serious jobs. And then you had quit your jobs in order to do this, open source, try to make it sustainable, living the dream, so to speak... And there was a question on "Well, where do you take it from here?" And desktop seems like a good place to take it, because you're not just syncing for people; you're providing collaboration, you're writing tooling around docs... And these are the people who are, generally speaking running infra at corporations, right? So is your end user, basically -- like, is DevOps still the term? I don't know, a platform engineer? I'm not sure what they're calling themselves today.
+
+**Ellie Huxtable:** Maybe a separate conversation, but I think DevOps kind of like failed.
+
+**Jerod Santo:** DevOps is over? Okay...
+
+**Ellie Huxtable:** \[laughs\] But no, it's sort of platform, SRE, DevOps... Whatever you want to call it. People that do SSH-ing and Kubernetes and stuff for their jobs.
+
+**Jerod Santo:** We used to be called sysadmins back in the day. I feel like we abandoned that because we wanted to get a raise, you know... But at the end of the day - I mean, you're administering systems, so I like the term sysadmin. I always have. I probably always will. How does it work? I mean, how does -- do you pronounce it Atuin?
+
+**Ellie Huxtable:** Atuin, yeah.
+
+**Jerod Santo:** Atuin. Kind of like Tatooine.
+
+**Ellie Huxtable:** Pretty close, actually. Maybe we should change the logo to have like a turtle and a desert planet. That would be pretty cool.
+
+**Jerod Santo:** That would be cool. The turtle is cute though, so that always gets me going... But Atuin Desktop - how does it work?
+
+**Ellie Huxtable:** Ooph... It has a lot going on. \[laughs\]
+
+**Jerod Santo:** It was an easy question for me to ask. It sounds like it would be a hard one for you to answer.
+
+**Ellie Huxtable:** Yeah. So, I mean, at a very high level, it's a desktop app. It's not an Electron app, it's a Tauri app - I'm not sure if I'm saying that correctly - which is a web view, but it's your system web view, instead of a copy of Chrome. And the backend's all written in Rust, which I'm very happy with, instead of Node.js. We'll talk about that more if you want later. But we essentially launch a web app locally in Tauri, and it contains an editor called Blocknote, which is a really cool open source project built on top of Tiptap and ProseMirror, that provides like a block style editor... And we extended that with all of our custom blocks.
+
+So to start off with, you're just editing -- you have a choice between a local file, which is currently essentially JSON. We're working on that. YAML, sorry. But again, we're working on that. Or you can use our hub, which you sign up for, you can create an organization or a single player, and it provides sort automatic syncing, offline-first editing, and the idea of that being like collaborative style features. You could have like multi-cursors with people all zooming around, which we can see being really useful for incidents, or helping someone figure something out... As much of it as possible is implemented in Rust.
+
+So right now we can't -- again, we released this as beta. There's still a lot of like "Right now it doesn't, but it will" going on. But a lot of the execution is written in Rust, and the idea there is we can bring it to be a runtime which is embedded in a desktop app, that runs in a CLI, all over the place, for executing runbooks.
+
+\[12:03\] My kind of long-term vision there, I guess, is that current automation is kind of an all or nothing approach. You start off with "Someone might maybe have documented something", and that's that alone is a stretch. Or you kind of have to write a million lines of YAML and have it 100% automated. And we find that there's very little in the middle; there's no happy medium. And a lot of the time, things just end up documented or not even written down because the automation is just such a high barrier of entry.
+
+The idea of the desktop app is to kind of make it easier to meet people where they might be comfortable, and that might just be writing down what you need to do, but it might be automating the whole thing with our blocks, and then running it with our runtime. So it's designed to be flexible enough to handle most sort of abilities.
+
+**Jerod Santo:** Mm-hm. So there's probably a group of our listeners who, when you say "Runbooks that run", which is one of the catchphrases for Atuin Desktop - they live in runbooks, they write runbooks, they copy and paste stuff all the time, they use their shell history all the time... And then there's other ones, more like myself, who -- I mean, okay, I used to be a system administrator, but most of the stuff that I had to deploy or manage at this point is like ./setup.sh, or deploy.sh, and I can just have a little script and it's good enough. And so that's at least my perspective. And there's probably a continuum of those folks.
+
+For those who don't understand runbooks all that well, besides like "It's a script of some kind", or "It's a list of things that I must do", can you expand more on what kind of stuff a runbook can do, and how they're usually used?
+
+**Ellie Huxtable:** Yeah. We've found a bunch of use cases. It could start off just being super-simple, almost like what you might write a readme for, in our system anyway. So go install these dependencies, run this migration command, run this command to generate some fixtures etc, etc. So it can be very sequential, "run this, do this, do that". But we also find that people use them a lot for like database operations... So maybe you need to run a whole bunch of SQL to set up a new schema, or migrate from one system to another... And we've made it flexible enough to make that easier, too. So our dropdowns you can power with the output of a shell command to select an environment you might want to run in, or a URI that you might want to use in a query. There's a lot of database use cases, too... But also, we use them a lot for our local development. So there's a collection of maybe values in a local database you need to tweak, or files that need generating... You can use them a lot.
+
+So you might have a home lab somewhere that has a bunch of setup required, and things can run over SSH, too... So it can show you -- yeah, you can use it for setting up a remote system, too. So maybe you need to set up your ZFS, and then you need to install K3S, or Docker Compose, or however you're running things, and it can help you with that, too.
+
+**Jerod Santo:** So one of the challenges I think with this, especially in the diverse world of computers that we live in nowadays, is different stuff on different people's computers. So this helps with sync -- I mean, if I'm using this as a team collaboratively, like through your hub... I'm not sure if there's a separate thing I can self-host or whatever; we can talk about that. But I'm syncing the runbooks across a team, and we're collaborating on those things, and that's all well and good... But oftentimes our environments are like drastically different. And so we've tried to throw a containerization at that, tried to help fix that... Is there any sort of concept or help with regards to that inside of Atuin?
+
+**Ellie Huxtable:** We don't currently, but it's something I'm really actively thinking about. We find a lot of people right now just have like a dependency step that does its best to set up the dependencies that you might need... But obviously, like you say, someone else might have a completely different local system than you.
+
+\[16:07\] The only approach I can think of so far is having a container integration. So we have the concept of contextual blocks, which you would use for specifying that something needs to run over SSH... I'm currently thinking of doing something similar for containers, where I can say "This should run in my local development container", and it will sort of transparently make that work. I'm not a huge fan of it as an answer. I find containers for local dev to be clunky, especially on Mac, when you're running on a VM.
+
+**Jerod Santo:** Amen.
+
+**Ellie Huxtable:** But I'm also not sure there's actually anything better, so...
+
+**Jerod Santo:** Right. It turns out maybe Linux is the answer in that case. I mean, I've been on a Mac for a very long time, and I've had this exact complaint for a very long time, and it doesn't seem to be getting resolved at any point... And so everyone else is like "Well, we just use Linux for dev." It's like "Well, that seems like an answer."
+
+**Ellie Huxtable:** Especially recently, I've been more and more tempted.
+
+**Jerod Santo:** Yeah. As has a lot of people. I think the luster and the excitement around Apple's products and ecosystem has just kind of like tarnished a little bit. I used to be very excited -- and maybe it's just like "Well, they haven't really changed much." Even the iPhone events, and stuff like that... I used to plan my Monday around the keynote, and "Let's watch it on the big screen" and "I can't wait to see it." And we'd even have shows... We used to comment, and we still do at times... But just like react to the Apple announcements. And honestly, especially the last one, I didn't even watch it. I was just like -- I don't know.
+
+**Ellie Huxtable:** Yeah, I didn't even know what changed for a few days.
+
+**Jerod Santo:** It's like "It's another new phone. Maybe I'll get it because it's been three years, but my current phone still works." Although they just released some MacBook pros, and so maybe I'll go look at those and change my mind, but... Yeah, it's definitely -- maybe that just shows age, and exposure. Maybe not just age, but exposure over time, for all of us, to these things. It's been like decades of kind of the same thing, just iteratively better... Anyways, a bit off topic. I definitely think that Linux is gaining mindshare once again. Maybe not once again, but in a new group of people that it previously didn't have mindshare in. And fast containers for dev is one reason why that might be nice.
+
+**Ellie Huxtable:** Yeah. I actually found the new Asahi - however you say it, the Linux on macOS - I have another Mac that's running that, and it's incredibly good.
+
+**Jerod Santo:** Is it?
+
+**Ellie Huxtable:** A lot of my experience was like running Arch on a MacBook in 2015, and that was dramatic. Whereas you install this and it just works, which was really cool. I recommend it.
+
+**Jerod Santo:** Yeah, that is super-cool. That is super-cool. We tried to do a show on the Asahi Linux, and those folks are hard to get a hold of. They don't like to come on podcasts. If you know anybody, and you can put in a word... Because it is a fascinating technology. But they're a shy bunch, that's for sure. What have been some of the challenges with this? I mean, you've been working on it for a couple of years... What have been the hard parts?
+
+**Ellie Huxtable:** With the CLI, the hard parts realistically have been the sheer amount of data people have. So if we ever want to change anything with database schemas, it's a migration which professionally, when we run a really big database migration, it's like "Let's really pay attention to what's going on here." Whereas when we release something, there's like hundreds of thousands of database migrations going on, and if one of them goes wrong, that's someone who's lost data, and we have to figure out how to recover that... So it's very -- like, mistakes I might've made a few years ago are sticking around.
+
+**Jerod Santo:** Right.
+
+**Ellie Huxtable:** With the desktop app, it's a different set of challenges, I think. It's very -- the CLI is quite simple. The command line environment is -- it is very varied, but it's still a terminal. You still have command control sequences and various other things... With the graphical app, it's a bit different. People running it on macOS, it's reasonably a consistent experience. But people on Linux, who I really want to support as best we can, the graphical environments differ hugely. So making sure things run consistently for different people has been kind of hard... We kind of -- we joke; there's two of us working on it... We joke that we're sort of building a terminal and a text editor at the same time, and neither of those things are known as particularly easy things to build... \[laughs\]
+
+**Jerod Santo:** \[20:25\] Well, they're kind of dramatically different, too. I mean, the desktop, besides like you said, kind of the spiritual ties and the fact that you can pull in some of that shell history into the desktop... I mean, they're kind of like wildly different products, aren't they?
+
+**Ellie Huxtable:** I think so. I think they solve very similar problems. They're all about knowing what to do next and not forgetting what you've already done. And I think when you're looking at a shell prompt that does not have any shell history and then does not have any good documentation, it's very much like what you can remember, what you can figure out and what you can google; or I guess these days ask an LLM. But we're trying to solve the problem of making the shell really easy to use, making repeatable workflows quick and easy... And when you look at it that way around, they are kind of the same thing. It's just trying to solve like an ecosystem of problems, rather than just one.
+
+**Jerod Santo:** Mm-hm. And so why did you go to the desktop? You could have just taken the shell, and maybe done a TUI, or expand it... I mean, it kind of is a TUI already, right? But expanded it and put this kind of functionality in, where it already existed. Why'd you decide to start fresh with a graphical interface?
+
+**Ellie Huxtable:** Yeah. So the end goal is to have both a TUI, a CLI, and a desktop app. But I think the desktop app is the most, like, revolutionary change. I'm kind of putting myself on a pedestal, I guess... \[laughter\] But there's a lot we can do in the graphical app that you either can't do with a TUI, or it feels clunky. The editing experience we can give people in a desktop app is a lot better than we could provide in a TUI.
+
+We have another example, we have Prometheus blocks. So imagine you have a runbook that is how to react to an incident, and you can have a chart in your documentation that's real time, with a terminal underneath it, that lets you change stuff, and you can see what you're doing and what reaction that has at the same time. And doing that in a TUI - I'm sure we could probably render a chart somehow, but it wouldn't have the same impact and it wouldn't be as explorable. And a lot of what we're trying to do is make it exploratory, and make it something where things aren't quite as strict, they don't necessarily have to run in order... And again, fitting that into a TUI is not as straightforward.
+
+An early prototype was actually a web app... So if you think like a Jupyter notebook style experience, where you run a server and then you open it locally... And I didn't really like how that felt. I think we couldn't integrate with the system quite as well, even with the web app angle... And the desktop feels a lot nicer. We're not necessarily shutting off any of these things as future sort of expansions, but the desktop's, I think, where it starts. I like to think of it more as you're putting a terminal into some docs, and not putting some docs into a terminal.
+
+**Break**: \[23:08\]
+
+**Jerod Santo:** How has it been working with Tauri as a platform? We've done a couple of shows on it. We have a few production apps out there... There are. They are out there. It's not trying to belittle it. But hasn't been a huge upswell. I mean, Electron still is the 800-pound gorilla in the room.
+
+**Ellie Huxtable:** Generally very good. There are some issues that the team are aware of and are working on... So firstly, the development experience has been really nice. Being able to write Rust has been really good for us, especially we're doing a lot of systemsy stuff. Being able to include all of the libraries we've already written has been awesome... It's a lot lighter as well. So when someone downloads like a 20-meg package, it's a lot nicer than asking them to go download like five, six, seven hundred, or whatever Electron apps that are now. The resource usage is lighter... I don't know if it's necessarily as light as people might say, because you are still running a web browser, it's just not another instance of Chrome.
+
+**Jerod Santo:** Right. As you proliferate Tauris, you're not going to have multiple Chromiums in the background.
+
+**Ellie Huxtable:** Exactly.
+
+**Jerod Santo:** But if you're just going to run one, you're still going to run a full Chromium, right?
+
+**Ellie Huxtable:** Exactly. Yeah.
+
+**Jerod Santo:** Gotcha.
+
+**Ellie Huxtable:** MacOS has been nice because every user is -- you know, they're using the same web view that's consistent; it's what we develop on, so it's fine. Linux has been -- honestly, I thought it would be worse. So it's a lot better than I expected. Tauri on Linux currently uses an old WebKit version, which has some performance problems... I think the performance is when there's a large number of DOM elements. There are several very long GitHub issue threads, if anyone's interested, but... It's luckily for us been mostly okay on most systems, but it's not quite as performant as it could be on macOS. The team are looking to use Chromium embedded framework on Linux, which would solve the performance problems, but I'm not sure if that's just doing Electron again. But what has been good for us is, again, the Rust on the backend means that anything that needs to be fast, we can make sure it's really fast, rather than it's all JavaScript.
+
+**Jerod Santo:** So another opportunity you had with this project - it was a fresh start codebase-wise - was Atuin CLI. I would say the reception to it that I see, and that I have, was like universally positive. It's just like "This just makes my life better." Open source... But the question is "Can we make money off of it? Who knows?" And now you come out with a desktop app, and it's like an opportunity to make some money, but also open source... You're following that same pattern. So your thoughts on that?
+
+**Ellie Huxtable:** Yeah, so the CLI was -- I mean, I could definitely charge people for hosting. I've had a number of people say "Oh, let me pay you single-digit dollars a month for shell history hosting." And that was nice. However, the way I looked at it was I don't want to work on shell history by myself forever. So it needs to have the potential to do more, and it also has to have the potential for it to be more than just me. And we have something like 26,000, 27,000 maybe people syncing their shell history... And I worked out that realistically, a hundred percent of them are not going to sign up to pay. It's going to be quite a low amount. And we can't charge very much for it, so it's quite a niche, realistically.
+
+\[27:56\] It also really doesn't cost very much in terms of infrastructure to run. I self-host the Postgres on a Hetzner machine. It's cheap, it's really performant, and other than my time, it's not a big sync. Desktop, I think has a lot more potential in terms of like business project. There's a very direct organization use case, potentially even stronger than the single player use case, really... And there's a lot more scope to do really interesting and flexible things with it.
+
+It is open source, it's Apache 2... What is not yet open source is the hub for syncing. So we have two types of workspace. You can have an offline workspace, that gives you local files on your disk. You can put them in Git... Whatever you want, basically. But the hub is the synchronizing real time backend, and it's not open source right now. Basically, we haven't figured out what to do with it yet, and I would very much rather have something closed source, and then six months from now decide to open source it, than open source something, realize I've backed myself into a corner, and then have to do a horrible license change and it's a nightmare.
+
+**Jerod Santo:** Right. What we say around these parts is "Rug pull, not cool."
+
+**Ellie Huxtable:** Exactly.
+
+**Jerod Santo:** So definitely keep your cards close to your chest, and then if you want to flip one over and show that you've got the ace of spades later, fine. Go ahead and do that.
+
+**Ellie Huxtable:** Yeah. And I feel very strongly that if you're running it on your machine, it should be open source. If it's a desktop app, you should be able to see what's going on, you should be able to change it...
+
+**Jerod Santo:** Yeah, I think that's a good distinction. That's a good place to draw a line in the sand, I think. And I think you're right, because when it comes to runbooks, I just feel like anybody who's playing single player don't care that much about something like this, because they've got their scripts... They might have a Notion or Obsidian, and they just have their docs in there if they write down -- or it's all living in their head. Usually, when you need a runbook, it's like "Hey, I put together how we do this. Now we can all run it when I'm on vacation", or whatever happens when it's your turn to be working and not my turn.
+
+**Ellie Huxtable:** Exactly.
+
+**Jerod Santo:** And so it's almost de facto team, right? Like, it's going to be collaborative... Obviously, there's the outliers who will be using it because they just like your software or whatever, or it's a prettier way to manage their own docs and scripts... But for the most part, I do think it's going to be something that organizations will adopt. Whereas the CLI is just like hackers and nerds... Sure, they work in an organization, but it's gonna be harder for them to get their org to sponsor or buy the thing.
+
+**Ellie Huxtable:** Exactly. And the only real way I could see the CLI applying to an org is like "Share your shell history with your team", which I don't think anyone's ever going to want to do, because it's weird.
+
+**Jerod Santo:** Yeah, it is kind of weird. It's like, at first you're like "That'd be great." And then you realize all the things you type in and you're like "No."
+
+**Ellie Huxtable:** I don't know, it's like my team seeing my browser history. There's not anything weird there, I just don't want to.
+
+**Jerod Santo:** Yeah. It just doesn't -- it's just useless, and maybe embarrassing at times, you know? You're like "Sometimes I type things that is wrong", you know? Or "I feel like a command is going to work and it doesn't." And it's like "There it is. That's my shell history."
+
+**Ellie Huxtable:** Absolutely. We have had a lot of people ask about self-hosting it, which is something I really want to support, but there's also the other angle of supporting it, and I've found with the CLI that it's self-hostable, and has been for forever, and it's tricky to support people. Generally, the Docker Compose we provide works, but a lot of the time people want to run it differently, they want to do their own thing with it, and supporting that is super-hard... So we're now kind of at the point where unless you're using our Docker Compose and it doesn't work, I probably can't help you. Mostly from just a time and energy perspective. With the desktop app, I wouldn't really want to be in the same place, and it is a lot more complicated to run for the backend.
+
+**Jerod Santo:** Sure.
+
+**Ellie Huxtable:** So if we do open source it, and I would like to, it would definitely be like "You can run it... It's on you."
+
+**Jerod Santo:** Yeah. Yeah, that makes a lot of sense... Versus like a GitHub enterprise thing, where you're going to go install it on their premises and operate it for them. Yeah, that makes sense. Going back to the CLI and potential monetization - people are desiring to give you money, and that's because they love the work that you do... And I've seen that online, the response to it is almost universally positive. It's very good software, people love it, they seem to love you, and they just want to give you money. Did you try even just turning that lever on? Like, even though it might not be your end goal, or something that you want to work on --
+
+**Ellie Huxtable:** \[32:23\] So we had GitHub Sponsors for a long time, and --
+
+**Jerod Santo:** Well, what about the hosted sync? Like "Let me give you two bucks a month to do this."
+
+**Ellie Huxtable:** I think that changes what it is. Maybe someday I will, but I think realistically, it doesn't cost me very much to run, and I'm quite happy providing it as just a thing people can use.
+
+**Jerod Santo:** Okay.
+
+**Ellie Huxtable:** The two bucks a month - I mean, I guess... But then it becomes a -- like, supporting people, because there's suddenly... Like, they're paying, they're actual customers, and it's a strong expectation there. But we haven't had any downtime in like a year, so probably we'd probably be fine, but...
+
+**Jerod Santo:** Probably. It sounds like you designed it pretty well, and it's relatively efficient.
+
+**Ellie Huxtable:** I'm happy with it.
+
+**Jerod Santo:** You're not having to constantly be working on it, and maybe upgrading, whether it's vertical, or scaling the other way... So that's nice. Remember when Radiohead came out with their In Rainbows album, and it's like "Pay what you want"? And I don't think free was an option on that one, but basically -- I mean, you could do something like that, where it's between zero and N dollars, where N is whatever you want to give me, but zero is what it costs... I think you have a not insignificant amount, which could also be called a significant amount of money... You know, maybe put some petroleum in your motorcycle, or something like that.
+
+**Ellie Huxtable:** It's true. It's quite expensive these days.
+
+**Jerod Santo:** It is. And it's not getting cheaper.
+
+**Ellie Huxtable:** Yeah. I think honestly, for me, I want to explore what we can do with the desktop app, and then maybe the future of the CLI might change... But where it is now, I'm super-happy with how it's going. I'm super-happy to let people just have it and use it. Realistically, my thoughts on monetizing open source software are more around like "Individuals just use it for free. Go have fun", and then it's the companies that should be paying. They get a lot out of open source, so...
+
+**Jerod Santo:** Right. So how has the reception then been to the desktop release? Because it's in beta, it's out there, it's open source... And it hasn't been very long, maybe a month or so that it's been out.
+
+**Ellie Huxtable:** Yeah. It's been really good. We have emphasis on opt-in telemetry.
+
+**Jerod Santo:** Nice.
+
+**Ellie Huxtable:** I split that up. It's actually opt-out, but you can't opt out on the launcher.
+
+**Jerod Santo:** Okay. So it's opt-out, but you just click a button right there.
+
+**Ellie Huxtable:** Yeah. It's right in your face when you start it up, so it's easy. And it's pretty much - we gained about a thousand signups on the hub, which was really cool to see. We've got people making hundreds and hundreds of runbooks, which is also really cool to see... I think the biggest thing for me though, rather than that telemetry, is more the community response. So in our Discord we have a separate channel for the desktop app, and people have been in there really active using it.
+
+GitHub issues usually imply there's an issue, but it also implies that people care enough to actually tell you that something's not working for their use case, or something's going wrong, and provide like a bunch of steps... We've had people provide some actually really thoughtful GitHub issues as well, which has been nice to see and nice to work on. I think for me it's also good validation... So again, I've been working on this for around a year, in one form or another, and we did release in a closed beta back in April, and that had a nice reception... But it's the validation of "You can go download it right now. It's on the website, go try it and let us know what you think", and that going successfully as well, which is really nice validation that the problem that I see as a problem and that I have built a solution to is a problem that other people see, too.
+
+**Jerod Santo:** \[35:54\] Yeah. Well, I did download it, and I'm staring at it currently... It looks very nice. I like that there's some examples, because I'm a guy with zero runbooks, and... I can put my deploy script into here and write about it, but just trying to kick the tires. You have the Atuin desktop release... This is actually how you guys release a new version of it?
+
+**Ellie Huxtable:** Yup. We run that all the time.
+
+**Jerod Santo:** You run that one all the time... Which is maybe like seven or eight steps... Updated 23 hours ago, so you're still currently working on it. That's always awesome. And then there's also one --
+
+**Ellie Huxtable:** Yeah, we've also got one for the CLI as well. That one's not yet open source, but it's mostly automated; the first 80% of it just runs, and then there's a nice little step of "Go look at this, make sure it's correct, merge the pull request, and then come back to the runbook."
+
+**Jerod Santo:** Gotcha. So I'm looking at the Atuin Desktop release runbook, and in the upper right there is a green Play button, and in front of me is "Notes. This runbook is a copy of the runbook that we use at Atuin", blah, blah, blah. "Here's dependencies, the release process, version bumps..." It has a typical kind of a Markdown-looking documentation, with executable blocks in there, that will run, I assume, on my local machine if I hit the Execute button?
+
+**Ellie Huxtable:** That's right, yeah.
+
+**Jerod Santo:** Okay. The upper right hand corner, with the Play - is that going to run the entire script, top to bottom kind of a thing?
+
+**Ellie Huxtable:** Yeah, that's what it does. So again, it's meant to be flexible enough that you can pick an individual block to run if you need to, or rerun, or whatever... But then if you've built it to be sequential, you can just run the whole thing, end to end.
+
+**Jerod Santo:** Nice. And then I assume it just like stops when certain parts fail... Because inevitably, with somebody else's runbook, it's going to be like "No, you don't have this thing installed."
+
+**Ellie Huxtable:** Yeah, it does. Terminals are the only exception. So terminals need to exit in some way, and they won't exit themselves. So because it's an interactive terminal, we don't know when the command you've run is done, or when what you're trying to do has finished... So you either need to exit the terminal within your script, or you need to press Stop.
+
+**Jerod Santo:** And when you say terminals, you're referring to like an interactive thing that's going to go on in here?
+
+**Ellie Huxtable:** Yes. So we have terminals and scripts. Terminals, it will run like your terminal emulator, just in your docs.
+
+**Jerod Santo:** Okay, so it's going to like block and wait for a response.
+
+**Ellie Huxtable:** Yeah, exactly. Whereas scripts - it's like running Bash on something, basically.
+
+**Jerod Santo:** Right. So in this desktop release, the first thing I see is just a curl command. That's a script, right?
+
+**Ellie Huxtable:** Yeah.
+
+**Jerod Santo:** And then this says "Terminal, install Cargo Bump." That looks like it's also a script.
+
+**Ellie Huxtable:** Oh, those are terminals. So if you run one of them, it'll open a terminal locally and run the thing.
+
+**Jerod Santo:** Gotcha. I'm afraid to do that. Oh, look at this... Oh, because I can actually stop and I can say no, which I'm glad I can, because I'm going to stop it right there... Okay, so terminals are interactive, and then scripts are just going to run, and whatever the response of the script the exit code or whatever.
+
+**Ellie Huxtable:** Yeah. The cool thing with scripts is we can capture the output. So if we want to capture the output of a terminal, there's a ton of control codes and various other things going on... But with a script, we can capture the output, and that we save in our local template system. So we support Ginger pretty much, and you can use the output of blocks to feed in as the input of other blocks.
+
+As an example, you might have a Postgres block, and the input for that - you might need a database URI that you can only get by running an AWS secrets manager command, or like a 1Password thing. So you can have all of the setup to get you to that point in your own book, and then you can run the query with the output and the script.
+
+**Jerod Santo:** Gotcha. Super-cool. So these examples are very useful, especially if I'm going to start writing my own, because they show you kind of how somebody is really using it... Is there going to be, or is there currently some sort of like compendium of example? Like, beyond your guys' official examples. I assume some sort of hub that's like shared coding thing...
+
+**Ellie Huxtable:** \[40:00\] So if you were to go to hub.atuin.sh/ellie, you can see my profile...
+
+**Jerod Santo:** Okay, so this is a thing.
+
+**Ellie Huxtable:** Yeah...
+
+**Jerod Santo:** For the listener, she didn't ask me to ask this question. I didn't tee her up on purpose. I really did not know.
+
+**Ellie Huxtable:** \[laughs\] Yeah. You can currently share runbooks only.
+
+**Jerod Santo:** Okay.
+
+**Ellie Huxtable:** And they only work for using an online workspace. You can mark them as private, public, or unlisted... I might share a Getting Started guide, or a migration guide, or a self-hosting guide for something I'm sharing with the world... Maybe I'll do a build guide for the desktop app.
+
+One of the other things we want to share on the hub is actual blocks. Currently, you can have saved blocks. Maybe you have something for setting up your dependencies that you do across like 10 different runbooks... You can just save that locally and then import it into various other runbooks. But in the future, you'll also be able to share those via the hub, and access other people's blocks... To try and kind of just speed up what everyone's doing.
+
+**Jerod Santo:** This is very cool. This seems like fertile ground for early adopters to get out there and build out their profile, you know? Is it accessible to everybody at this point, it's just not like indexed?
+
+**Ellie Huxtable:** Yeah, go ahead. It's not indexed... Go ahead. \[laughs\]
+
+**Jerod Santo:** Yeah, go ahead. How do you do it? Like, if I wanted to start publishing, what would I do?
+
+**Ellie Huxtable:** In the desktop app, when you start it up, it'll offer to make an account. If you said no, just click your profile on the bottom left and you can register there, and that will get you set up with an account. Connect you to the desktop app... You should be good to go.
+
+**Jerod Santo:** Gotcha. Yeah, so I did appreciate that... One of the things I don't appreciate about desktop apps, that should just be usable local without any sort of account, is when they still don't tell me I can use them locally without an account... And yours does say "Hey, sign up." And then it's like "But if you don't want to, click right here and just use it." And so that shows me that you're a person that I, because I do appreciate that. Just let me use a thing. And then if I want to sign up for a reason, like having a profile, make it valuable, then I'll go ahead and sign up when I want to.
+
+**Ellie Huxtable:** Exactly. I want it to be useful offline, I want it to be useful open source... And I think one of the big things we're aiming for is, you know, we might not be around in three years time. I would very much like to be... However, if we're not, and someone has built out their whole entire workflow on this, it would really suck if they couldn't use it anymore. So it's important that they still can.
+
+**Jerod Santo:** Yeah. How portable do you think it would be if I was building my runbooks, assuming the worst you? Like, you disappear, the company disappears, the GitHub repo's gone... Like, I can't run it, I never got the self-hosted thing... Could I get them out relatively easily? It's gotta be close to text at the end of the day, right Yeah. So if you're using an offline workspace, it's just YAML on your disk...
+
+**Jerod Santo:** There you go.
+
+**Ellie Huxtable:** ...and the format is quite verbose, but a bunch of scripts and you can pull it all out and it's fine. What you can also do is you can export Markdown. So it is -- we don't really like the fact that there's this YAML as our file format... So the end goal will be to use Markdown as the file format. But what you've probably noticed is that the editor is richer than Markdown provides sort of Markdown for.
+
+**Jerod Santo:** For sure.
+
+**Ellie Huxtable:** So figuring out a very light extension we can do that lets us encode everything we need to encode, but also makes it so that you can still actually edit it in a text editor, and read it, and various other things. So that's not done yet, and that's something I want to do correctly, and would actually appreciate input on if anyone has any thoughts. But if you export Markdown, you'll see that there is some front matter in the code blocks, which is how we're currently seeing things.
+
+**Jerod Santo:** Gotcha. Yeah, I think that would definitely be a huge upgrade, the way to get that done. What else are you working on? What else is in the pipeline?
+
+**Ellie Huxtable:** Yeah, so coming really soon is a new runtime. The current runtime is very much like "Let's get something that works and see if people like this", and the new runtime is going to be let's make sure things are really, really robust, and flexible, and will allow you to run runbooks from the CLI... Which is in my mind very important, because then it allows you to run them in the CI systems as well. So maybe you have a readme that's getting set up and you could actually test your readme, and that would be really cool.
+
+\[44:18\] Maybe you have a bunch of disaster response runbooks, and it's really important that they actually work, and you don't find out that they don't work at 3 a.m. when you need them... And if you get things set up right, maybe you could test that on a schedule in some automated system somewhere. I think it's also important just to be able to execute these things without a desktop app if you really don't want to do it, so that's what's coming soon.
+
+**Jerod Santo:** Mm-hm. So when you say a new runtime, can you unpack that phrase for me, and tell me more about what that means?
+
+**Ellie Huxtable:** Yeah. So when you click Play on a terminal, there's a bunch that goes on behind the scenes. Currently, the frontend makes a call to the backend to ask it to open a pseudo terminal, and then it gets a handle back, and then it starts writing to that with the input you've provided it... That's like a special case. So the frontend knows that it's a terminal, it knows that it has to open a PTY, and do X, Y and Z. What we're trying to do is get it to the point where the frontend just says "Run this block", and that's it. And that means that the frontend doesn't have to be a desktop app, it could be anything... And it also opens the door to a lot more flexibility in how blocks are defined. So we're not there right now. However, I would love for people to be able to write their own blocks, their own integrations, whether these are backed by a shell command, or some set of code; it doesn't really matter. But right now it's very much like every block works, but it's like a nice sort of combined effort between the frontend and the backend, and we'd like it to be the frontend as minimal as possible.
+
+**Jerod Santo:** It's more minimal, does less heavy lifting, so that a different terminal could call a runbook and run it from there. Is that going to require you to first figure out a more permanent serialization format, or you already have that?
+
+**Ellie Huxtable:** No... So the serialization is -- there's the in-memory format and then the on-disk format. The on-disk format is the one that's up for change, the in-memory one we're happy with.
+
+**Jerod Santo:** Okay. Right on. It is open source... Is it open contribution? How do you, why do you, when do you, how do you...? Tell me more.
+
+**Ellie Huxtable:** \[laughs\] So we accept contributions from anyone. We accept fixes especially quickly. Features - this is something I never really planned out with the CLI, because I didn't think anyone would care when I first released that. We never really thought about how to accept feature requests from people... And in the early days, I was very much like "Oh my God, someone's built something cool. Let's merge it and let's go." And towards sort of more recently, it's been "Is this something I want to maintain long-term? Is this something that people actually want, or does this just sort of bloat the code for no real reason?"
+
+For instance, there's been a pull request open for a really long time that I feel kind of bad about, but someone's added PowerShell support, which is super-cool. However, when he first opened the pull request, I haven't seen very many pull requests from the guy. And if he's listening, I really appreciate your work and thank you very much... However, when he first opened it, I didn't know if he'd stick around, and I made it very clear that I could not maintain a PowerShell integration. It's not something I have any familiarity with. And the dude stuck around for like six months, making fixes to various things, making other PRs... And now it's like "Well, I have trust in your code, and I also have trust that when you say you're going to stick around and keep it going, you probably will." So that's also a consideration I never really had originally.
+
+But jumping back to your original question - open contribution, fixes... Like, anyone fix anything and open a pull request, I'm super-grateful and we'll review it as quickly as possible. Features - if there's an open issue for something and someone really wants it and wants to make a PR for it, that's awesome. But I do generally prefer an issue before someone makes a PR, purely because it really sucks to say no to something someone's put a lot of effort into.
+
+**Jerod Santo:** Right. It's a lot easier if it's just an issue. If there's code behind it, you're like "Oh, you spent all this time..."
+
+**Ellie Huxtable:** \[48:20\] Yeah, you spent all this time on it, and then that's not something we're going to merge, and I'm really sorry.
+
+**Jerod Santo:** Do you communicate a roadmap, or like a place where you're headed? Because sometimes it's hard to know... Like, I have a feature request. Obviously, I could just ask... But if I knew "Here's here's our direction", I could at least say if I'm tangential to it or not.
+
+**Ellie Huxtable:** So a roadmap is on the roadmap. \[laughs\]
+
+**Jerod Santo:** Okay... \[laughs\]
+
+**Ellie Huxtable:** I'm planning on doing it soon.
+
+**Jerod Santo:** You have a roadmap that includes to create a roadmap. I like that.
+
+**Ellie Huxtable:** Yeah. So soon, we'll be writing a ton of issues. Realistically, I've got internal notes and various other things on what needs doing, and now that we're open source, that needs to be made public. The format of the roadmap will probably be mostly GitHub issues, just because that's what's accessible to most people, and pretty much everyone has a GitHub account... But we will be using the forum and the blog more, too. So a lot of the development so far has been on the forum; it's generally where we've found discussions live best. GitHub issues, I tend to prefer keeping actionable. There's a clear reason to close one because a PR has been merged, or something like that. GitHub discussions never really did it for us. They felt like an afterthought from GitHub, to be honest...
+**Jerod Santo:** Well, it came years later.
+
+**Ellie Huxtable:** Exactly. So Discourse - we host a Discourse forum, and that's been a nice place to host discussions.
+
+**Jerod Santo:** What about the CLI? Do you consider it mostly done? Does it have a major feature requests? Is there a roadmap for that, or what's the status of the CLI tool?
+
+**Ellie Huxtable:** Yeah, there's less of a roadmap, more of a bug fixes I will try and get to as soon as possible. Features - I tend to prefer features a lot of people ask for, or that I personally really feel the need for... Which is a little bit selfish, but you know... But it's very much -- I would rather see enough demand for a feature before bringing it in, unless it's very, very obvious that it's needed. There's been a lot of -- I think recently I saw someone talking about being able to add comments and tags and various other things to commands, and I think that is useful... However, my thoughts there are that the interface for actually searching them needs to be really good. Because there's no point in adding organizational metadata and not being able to search through it effectively. So I think that's quite a large change, and therefore one that unless a lot of our users want it, it's not necessarily worth doing, because then we'd have to maintain it, too.
+
+**Jerod Santo:** And are those conversations going on in the GitHub issues, or is there a separate place for those?
+
+**Ellie Huxtable:** Usually GitHub issues, people will make a feature request and then it can be discussed. But yeah, I need to get better at managing those, because it's a bit of a mess.
+
+**Jerod Santo:** \[51:01\] Fair. Such is life, it gets messy, and sometimes it stays messy... Especially when you have a shiny new object that you're paying all your attention to, right? I mean, this thing is a big lift; I can tell that you put a lot of work into this. And you can't spend all your time on the CLI when you're building a desktop app, right?
+
+**Ellie Huxtable:** Yeah, I think it's definitely -- I don't want to neglect the CLI. We had one of our users thank us for not neglecting it, which was nice, because I'm glad people feel that way. But the CLI is definitely much more mature... We've noticed -- you know, when I first released it, a lot of the early issues were very "This does not work at all on my system because of X, Y, Z." And now it's very "This does not work with a very, very specific setup and very specific set of use cases", which is sort of for us more of a signal of maturity, and less of a "S\*\*t, I just broke everything", and more of a "We need to figure out how to integrate with something I'd never considered."
+
+**Jerod Santo:** Something new, or something obscure. Yeah, from my perspective - I mean, I've just been using it kind of how you built it on day one, originally, just ever since. And it just does what it does, and it's mostly invisible. I don't think about it very often. I use it all the time, but without thinking about it, like you do good shell tools. You're like "Well, I don't think about ls very often either, but I'm going to use that one all the time as well."
+
+**Ellie Huxtable:** We have a lot of people tell us that they didn't realize how valuable it was until they didn't have it.
+
+**Jerod Santo:** Oh, yeah. That's gotta feel good.
+
+**Ellie Huxtable:** Yeah, that's cool.
+
+**Jerod Santo:** And who knows, when I get a new machine, maybe I'll finally get the sync to be useful for me... I'm like the only person that doesn't care about syncing things.
+
+**Ellie Huxtable:** Let me know if you do, it'd be cool. \[laughter\]
+
+**Jerod Santo:** And I'm going to try to find a way to use this desktop app. I love your work, and I think you put a lot of thought into the way you go about building software, and I think that that's rare, and just refreshing... And it just makes for tools that are delightful to use. And I can tell you put a lot of love into this one, even though it's not exactly right in my daily use wheelhouse. I'll still look for a reason to go beyond kicking the tires.
+
+**Ellie Huxtable:** Yeah. I hope you do, but I think in terms of users, it's like, the CLI has a very broad -- you know, if you use a terminal, it's probably useful. Whereas the runbooks app is like maybe less broad use case, but in terms of the actual use, it's much deeper.
+
+**Jerod Santo:** For sure. If you're a user of this, it's going to be part of your daily job, and your team will rely upon it. It'll be very valuable for the people who do use it. So I agree, it'll go deep. And these are people who are generally well-employed at places that have complicated infrastructure, and they need tools like these to make their job not miserable all the time, you know?
+
+**Ellie Huxtable:** Exactly. Yeah. That's what we've been seeing.
+
+**Jerod Santo:** Very cool. It was awesome to catch up with you. You've been doing a lot... Anything that we didn't talk about with regard to Atuin, the CLI, the desktop, San Francisco, motorcycles...? Anything else that you think we should discuss?
+
+**Ellie Huxtable:** think we covered pretty much everything. It's been a good chat.
+
+**Jerod Santo:** Absolutely. Well, the website atuin.sh - there you'll find both the desktop app and the CLI. Maybe you'll use them both, maybe you'll use one of them... Maybe you won't use either one, but you can check them out. They're open source. If you want a good Rust codebase to check out, and maybe contribute to our listener, I would submit that to you because she makes good software, folks, and she pays close attention to the details that matter. So check that stuff out.
+
+Ellie, thank you so much. I wish you all the best with this new endeavor... And definitely keep us in the loop, especially if the self-hosted thing comes out. As more things are released and open sourced, don't be a stranger.
+
+**Ellie Huxtable:** Thank you very much. It's been great talking to you.

Bringing Vitess to Postgres (Interview)_transcript.txt

@@ -0,0 +1,402 @@
+**Jerod Santo:** Today we're joined by Sugu Sougoumarane, co-creator of Vitess, co-founder of PlanetScale, and now the head of Multigres at Supabase. Sugu, we are excited to have you here.
+
+**Sugu Sougoumarane:** I'm very excited to be here, too.
+
+**Jerod Santo:** I'm excited to see you back at work... You've been off work. You said -- you just told me you took a three-year sabbatical? Is that accurate? Is that true?
+
+**Sugu Sougoumarane:** Yeah, yeah. I even wondered if I was going to come back, until, you know, I started feeling the itch... And yeah, that's what brought me back in.
+
+**Adam Stacoviak:** Were you gardening? You know, ripe tomatoes... What kind of vegetables did you start growing?
+
+**Sugu Sougoumarane:** \[laughs\] I actually went on a self-reinvention of some sort. I kind of did a full debug of myself of some sort.
+
+**Jerod Santo:** Did you climb a tall mountain, or how did you do that?
+
+**Sugu Sougoumarane:** Oh, no, just sit at home, think, walks... And a lot of thinking. And I even got into doing some voluntary work, because I realized as a human -- basically, I thought I should be more human, and that led me into a bunch of voluntary work, which I actually started to do. And I even thought of going full-time. Then I realized it's not something you can do full-time... And that's when I realized that I'm starting to miss the fun of doing technical work. So I started doing some advisory work and stuff, and slowly I got pulled back in and realized "Oh my God, we've got to do this project!", which is how I am where I am today.
+
+**Jerod Santo:** Did it take -- how long did it take before you started getting that itch? Because three years is a pretty long one. I think by that point I might be detached enough to not get the itch again.
+
+**Adam Stacoviak:** I agree.
+
+**Sugu Sougoumarane:** It was strange... I think last year, sometime late last year. So for about six to eight months I've been feeling the itch. I think it started because I'm still in touch with friends, and stuff. They used to ask me for questions and help about things... And then as you answer, you kind of enjoy that you did it, and you say "Okay, maybe I can--" Then I talked to some people and they said "Oh, someone like you is still useful." And as you realize, the last three years everything changed in the industry. AI actually -- the first GPT came out just when I went on sabbatical, around mid 2022, I think. And when I thought of coming back, I thought "Maybe I'll be obsolete, or something. Because I don't know anything about AI..." But then it turns out that infra is still way behind.
+
+**Jerod Santo:** Yeah.
+
+**Sugu Sougoumarane:** And they say "Oh, what do people desperately need?"
+
+**Jerod Santo:** Databases still matter, it turns out.
+
+**Sugu Sougoumarane:** Databases matter, infrastructure matters... Nothing much has changed in that area. So I came back and slowly got back into things.
+
+**Jerod Santo:** What pushed you away in the first place? Why did you go on that sabbatical? Had you accomplished all your goals? Were you burning out? What was the situation that made you \[unintelligible 00:07:08.26\]
+
+**Sugu Sougoumarane:** Both of those, actually.
+
+**Jerod Santo:** Both. Okay.
+
+**Sugu Sougoumarane:** So this was 2022... I had been working on Vitess for 12 years straight. We started Vitess in 2010 at YouTube, and... Vitess is an enormous project. And for a while, I was the sole maintainer of Vitess. And it was -- it's a pretty big load on your brain. And after I founded PlanetScale, hired a team and trained them, and finally, the team was outproducing what I could do... And that's when I realized "Okay, it's now time to recover from this huge load that I've been putting myself through." And that's what led to my taking the sabbatical.
+
+**Adam Stacoviak:** \[08:07\] What year was that again? Was it around the time that you -- I'm not familiar with the story behind the scenes really with Sam Lambert, but I guess I'm tangential friends by having him on the show, and hearing his backstory, and PlanetScale's early story, and mainly taking over the CEO role and what that did for PlanetScale at that time. What time was that you stepped away, and was that the time that you brought on Sam as a CEO?
+
+**Sugu Sougoumarane:** So I think my memory may be hazy here... I think he came on board around 2020. Sam has been a fan of PlanetScale from the beginning. He even did a Sequoia Scout investment when we launched. And so he has been following us, he's been actually advising us also in the early couple of years... And then in 2020 he came on board as I think CPO, Chief Product Officer. And he changed the whole image of PlanetScale. Until then, we were very - what I would call a serious backend database, but in some respect boring of some sort... We were pretty serious, we would do things, but now we were -- he brought the excitement in when he came on board.
+
+The first feature he launched was the schema branching and merging, which the dev community completely loved. And that was around '22. Jiten at the time was CEO, and Sam talked... And they came to an agreement that if Sam became the CEO, that would be really effective, because then he could control the whole story. He became the CEO probably -- much earlier. Probably 2021 would be my guess... I'll have to look it up. It was probably announcements. So he had been the CEO for a while, and then in '22 is when I left also. By that time, the team was all wrapped up. Vitess had a tech lead, so everything was looking great. I almost felt like I was a warm body... \[laughter\]
+
+**Jerod Santo:** Well, that's a good feeling. I mean, you've done a lot of things you set out to do. For those of us playing catch up, can you describe Vitess, what it is and how it helps scale databases?
+
+**Sugu Sougoumarane:** Yeah. Yeah, definitely. Vitess, actually -- when we started Vitess, we thought it was a six-month project. \[laughs\] So YouTube was actually in deep trouble scalability-wise. There were more than a handful of outages every day. Sometimes 10, 12 outages per day, and they were all caused by the database. I would say outages 10-12 pages per day. Many of them would lead to outages, and they were all caused by the database. That's the time when my co-creator, Mike Solomon and me, kind of decided to solve this problem once and for all. So we decided to take ourselves out of this firefighting mode and think of a solution that will leap us ahead of where we are today. Because it felt like what we were doing now was a losing battle. It was like whatever we were doing, things were getting worse by the day. And that's how Vitess was born.
+
+\[12:01\] And we wrote code for about a year, actually. It took us about a year to launch the first version. And the first version that we launched was just a connection pooler, which was actually the biggest problem we were having with MySQL at that time. And that immediately gave some breathing room for YouTube. And the way we solved it was we didn't solve it like how you would solve a problem, which is "Let's solve connection pooling." Because we sat and brainstormed about all future problems that we're going to have, we actually built an intelligent proxy that would not just connection pool, but would also understand your queries. So we actually built a parser in that connection pooler.
+
+And once this got launched, people started noticing that this so-called connection pooler is intelligent, and it can do other things beyond just taking care of connections... Like for example fielding bad queries, or watching queries and killing them if they run for too long. And so we slowly started adding those features, and it became an intelligent connection pooler that would protect the database completely.
+
+And that eventually evolved into -- then we added sharding solutions to it, like how do you re-shard... And then we added one more layer, which is the routing layer... And then we realized "Oh my God, this is now almost a fully distributed database. Why don't we make it a fully distributed database?" Which was the final step, where Vitess became a fully sharded solution.
+
+By the time -- we started in 2010. I think what we call as the v3 was around 2014 or so. About four years later, Vitess could be adopted by anyone outside of YouTube. And then the rest is history.
+
+**Jerod Santo:** The rest is history. Open source, mass adoption, PlanetScale, built a business around it... It's a very successful business to this day... Really cool stuff. Kind of an open source success story. I was gonna say "prototypical", or "stereotypical", but not necessarily. But just a good one. Just a good open source success story, with money involved in the business that could be built... Kind of what everybody wants, is build some cool tech, open source it, help companies scale their businesses, provide massive values once and for all, like you said... And then also build a business around it, be able to be successful... And then take a three-year sabbatical when you're all said and done. You're like "Alright, I'm done... For now."
+
+**Sugu Sougoumarane:** \[laughs\]
+
+**Adam Stacoviak:** Yeah. To extend from what you're saying there, Jerod, I think it's wild... Sugu, one thing you mentioned was the nine month or six month you thought that the project would take in terms of like its doneness... But then it turned into a company, and then obviously, PlanetScale has been able to -- and I think others too, have been able to hire and maintain full-time open source maintainers on the Vitess project, not just to support PlanetScale, but to support the project. I think that's just so wild, how you think it's just this little thing, or it's a certain amount of time and solving the scale problems, and the next thing you know it's something much bigger than that.
+
+**Sugu Sougoumarane:** Yeah, yeah. The first adopter was actually a company called Flipkart. I believe they still use the older version of Vitess... Then HubSpot came on board. They still use Vitess, and they have their own operator. They were among the first ones to launch Vitess in Kubernetes. And then later Slack came on board, and GitHub came on board... Now they are still pretty substantial contributors to the project at this point.
+
+**Jerod Santo:** \[16:07\] Well, you said once and for all, and I said once and for all, but the technical facts about Vitess is it's not for all, right? It's for all who are running specific database systems. And one of those alls who aren't part of the all is the Postgres ecosystem... Which has been awesome for many years, but has been exploding, I would say, in the last 5 or 10 years; in usage, in features, in capital support... I mean, there's just lots of excitement and growth in Postgres, none of which Vitess could provide for, right? It was just MySQL-specific, right?
+
+**Sugu Sougoumarane:** Yeah, yeah. I think the Postgres community took note of Vitess a long time ago, as far back as 2018, was the first conversation I had with someone from the Postgres side and they said "Hey, we need something like Vitess for Postgres." And there have been a few false starts on this project. I think the one where we got pretty serious about it was actually in 2020. Someone called -- I don't know, he's pretty well known in the Postgres community, Nicolay. So he and I talked, and we felt that we should do this thing, port Vitess for Postgres. We even formed a sub team and started talking about it.
+
+But you know, 2020, PlanetScale was on the rise... It was just too busy, and unfortunately, I had to make a call, that I have this company to take care of, I have this company to build... I can't afford this distraction at this time. I actually kind of had to back out, saying that I don't think I can do this at this time. I still feel bad about having done this... And this -- so Postgres has been in the back of my mind all along. Even after I went on a sabbatical, I reached out to some of the older contributors and told them "Maybe you can start this project." But at that time they felt like they all had their jobs, and this would all be something on the side... Until this year, when I said "It's now or never." And that's when I said, "Okay, I'm going to take this plunge."
+
+**Jerod Santo:** I think mimicry, or ports, or whatever you want to call it, is kind of a tried and true part of the open source culture. You know, take this good idea from this ecosystem which I appreciate, but I'm not day to day user of, and bring it to my piece of the world, and reimplement it over here, or steal the ideas... And it's kind of -- Vitess has been so successful and so helpful, especially to large companies who are well funded, at helping them scale... I'm surprised that all these years there is not already a Vitess for Postgres, community built, or coming out of a startup or something... Lik, no, Sugu has to do it, right? The Vitess people are going to do it.
+
+**Sugu Sougoumarane:** I don't get this -- yeah, I don't know why I had to come and start this... \[laughs\]
+
+**Jerod Santo:** Is it that hard? Because usually, you look at something and you think, "It can't be that hard. I'll do it over here." Is it that hard of a technical problem? Or is MySQL and Postgres implementation-wise so dramatically different that even the concepts don't necessarily come across? What's that chasm look like?
+
+**Sugu Sougoumarane:** I think it is more of the first problem. The fact is sharding, if you go and ask in a university, they'll say sharding is one of the oldest concepts in databases. It's over 30 years old. But the issue with sharding is that nobody ever worked on it, nobody ever developed it into something practical, that will actually work in production at scale with performance...
+
+\[20:22\] Which is actually strenuously hard. For me to make the leap from an application sharded system to a system where the application does not care about sharding, and generically implement it in a database, it took me about -- I literally sat and thought for three months straight, nothing other than just pure thinking. Understanding SQL, understanding query complexity, understanding relational algebra, going back all the way... Because, okay, how do I shard it? You go hit the books, you go look at databases. Nobody has ever done this before. And so then I realized, "This needs to be basically reinvented from the ground up."
+
+So that was one of the hardest, one of the most strainful thinking things that I have ever done in my life. There were times when I used to get headaches thinking about it, it was that hard. And finally, one day I had the a-ha moment. "It looks like I have an answer to all the questions to make this thing work as a sharded environment, as a sharded system." And it's possibly that is a typical -- when you get a query, the typical way you break it down is different from how you would break it down in a sharded system... Because you get a select statement, you pass it, and then you identify each of those nodes, and you just assign operations to it in a traditional database. And then you run the optimizer that rearranges these things. But in a sharded system, you have to identify groups of things that actually don't need to be changed at all, because there is an underlying database that can execute that entire query with no changes, because what's underneath is also a database.
+So being able to analyze queries and identify these subparts and be able to outsource it to a database, and then do the rest of it on your proxy that's actually sending the queries required a different kind of thinking. I think that is one reason that I could think of why sharding is hard.
+
+**Jerod Santo:** Can you describe your a-ha moment, like what your realization was? Or did you just try to describe it in other words and I missed it? Because it sounds like you're describing the problem and the system with the proxy in there. But what was it that cracked the nut that said "A-ha! We can do it this way"?
+
+**Sugu Sougoumarane:** Yeah, so how did I do it? The a-ha moment was -- actually, there was an early a-ha moment, an early a-ha question, which is... So before sharding, before this generic sharding was there, the way the application worked was it would actually send a query, and along with the query it'll tell you where to find the rows for this query. We called it the keyspace ID, which basically helped you -- there's a lookup where you look up the keyspace ID, it tells you "Oh, this keyspace ID can be found in this shard." The application had to guarantee that all the rows for that query lived in that shard. So that made it extremely simple. And keyspace ID was also a value, a column in every table that we had, which made it safe and simple, which means that you can't go wrong.
+
+\[24:24\] And the question I asked myself is "This keyspace ID is a crutch. Can we move without it?" And it was scary, because the application knows the keyspace ID. But how would this system know the keyspace IDs? Vitess does not know the keyspace ID... So if you just sent me a query, what do I do with it? I don't know where to find this. And so I went to the application and realized "The application is figuring out what this keyspace ID is. How is it figuring it out?" And that's when I realized "Oh, it's figuring it out based on the user ID." So there is a source. And it always turns out that that source is also in the where class of every query that it sent. And that's when I realized "Oh, well, if the application is doing it, what if I did it for you?" You say \[unintelligible 00:25:24.04\] and I will do this computation, which is a hash function, and figure out where it is and send it.
+
+So then the question is "Oh, then in that case, how about the other queries?" There were other complex ways by which the application was computing the keyspace ID... And that's when what if we internalized each and every one of those things, then the application does not need to ever compute the keyspace ID. We will do it. So that was the first a-ha moment. And then the second a-ha moment was -- the second question is "But in that case, only these queries." So that limited the number of queries the application could send.
+
+The next question was "But then if we took this over, what if there's a query that came where a keyspace ID could not be computed? What would we do?" That's when I realized, "Okay, then I have to understand the query." What if it's a join? What if it's a correlated sub query? What if it is a non-correlated -- so that's when I had to go into relational algebra and find a way to map any random query to a query plan where I would say "Oh, this query, I need to send it to all shards." Or "This query, I can send it to just this one shard." Or "This query is in multiple shards, but I cannot send it to all shards, because I need to take data from here and then data from somewhere else..." So the a-ha moment is when I identified all these primitives, and I could prove to myself that using these primitives - I call them the route primitives or the sharded primitives. If I used these, I can satisfy any query.
+
+**Jerod Santo:** Gotcha.
+
+**Sugu Sougoumarane:** The full SQL language can be supported, eventually. We just need to do a -- so the first implementation only supported some constructs, and then we slowly started adding more and more. But we knew that theoretically, all SQL can be supported in this sharded system.
+
+**Jerod Santo:** \[27:41\] So to the Keyspace ID point - so I understand the relational algebra and the entire SQL dialect, supporting all the keywords over time. That makes sense to me. The keyspace ID - as a non-Vitess user, I'm not sure how it works. Does the infra person or the team that sets this up, do they have to go in and somehow configure it to know how you should shard against this particular application, so that it can do the keyspace ID thing? Or is it just -- because different databases have different, as you know...
+
+**Sugu Sougoumarane:** Correct, yeah. There's a cool story there... The first keyspace ID was actually a random assignment.
+
+**Jerod Santo:** Okay.
+
+**Sugu Sougoumarane:** So we would take a user and say -- I don't know if it \[unintelligible 00:28:29.25\] I actually don't remember how we initially computed. We used to randomly assign them to shards. Actually -- oh, now I remember. A user was assigned a shard, and we would put the shard number in a lookup table. It's random. So when a user gets created, they are assigned a shard, and the shard number is put in a random table.
+
+**Jerod Santo:** Sure.
+
+**Sugu Sougoumarane:** And then came the time when we had to reshard it. And we realized "Oh my God, how are you going to reshard this, because now there is this table?" That's when we transitioned to keyspace ID, where we assigned a unique ID to a user, and then we actually assigned that keyspace ID -- we changed actually this lookup table to replace the shard number by the keyspace ID. And at that time, we populated every table with the keyspace ID. And then we realized that we are wasting this lookup for no reason at all. Then we did a full resharding, where we did ranges based on keyspace IDs... In which case you just hash the user - you have the keyspace ID - and then that goes to that shard. So that's how we evolved at YouTube, from where we started. But at that time, the keyspace ID was a column. And the reason the keyspace ID was important there was because we were using MySQL replication to do the resharding, and we had to use that value to decide where the row should go when we resharded. And that was actually another hurdle I had to cross, which is "Can you reshard if you get rid of that row?" In which case then we'd have to make MySQL do that computation to do the replication, which means that it had to be a function that MySQL supported.
+
+The interesting story here is at that time there were a large amount of arguments about "What is the best sharding scheme? Is range-based sharding good, or is a hash-based sharding good?" We spent hours fighting on it. There was no -- or a lookup-based sharding, right? There's so many sharding schemes... And that's when I realized that there is no correct answer. And anything you choose, somebody is going to hate it. And that's when I came up with this idea of no first-class citizens in sharding, which means that Vitess will actually -- every sharding scheme is a plugin, is an extension.
+
+**Jerod Santo:** Gotcha. So you pick the one you want to use.
+
+**Sugu Sougoumarane:** You pick the one you want to use.
+
+**Jerod Santo:** And you don't have any opinions on that, as Vitess.
+
+**Sugu Sougoumarane:** Exactly. Exactly. That goes back to actually a story from -- I don't know if you've heard of Illustra.
+
+**Jerod Santo:** No.
+
+**Sugu Sougoumarane:** \[31:54\] So this is 1990s... Michael Stonebraker had, I think, just launched Postgres. And from that, he -- he founded a company called Illustra, which is a commercialization of Postgres. And the claim to fame for Illustra is that indexes need not be defined by the database. You can define your own index and do it as a plugin. And at that time, I was at Informix, and Illustra was actually winning deals against Informix as a startup. And Informix actually ended up acquiring Illustra... And I loved that concept so much that I actually transferred myself into the Illustra team, to develop pluggable indexes. And that inspiration remained with me all these years. And when I said "We have to develop indexing for sharding schemes", I said "Well, you know what? I'm going to use this idea", where all indexes in the sharding scheme are externally defined.
+
+**Jerod Santo:** Yeah. It makes sense, because -- I mean, oftentimes opinionated software is better. Like, "We do have a golden path, or we do have a way that you should go." You can swap it out if you want, but this is what the framework or the tool thinks you should do. But when it comes to this -- like I said, and like you know, the shape of people's data is so different and custom that there isn't necessarily a best way of sharding, as you guys found. So it's what's best for your circumstance. And so why have a first class citizen when you can just say "No, you're going to figure this one out yourself, because we can't actually figure it out for you"? I think that makes a lot of sense.
+
+**Sugu Sougoumarane:** Yeah. The thing that I'm now starting to realize -- I'm now learning Postgres a little bit more deeply... I find the same ideas still there in Postgres, where indexes are pluggable, you can plug in data types, you can define operators... So I'm kind of feeling like "Oh my God, there's a lot of similarity between--"
+
+**Jerod Santo:** You've gone full circle.
+
+**Sugu Sougoumarane:** Yeah, so it's pretty exciting to see this.
+
+**Break**: \[34:22\]
+
+**Jerod Santo:** So what's different then? It seems like the concepts, the breakthroughs you had with Vitess, the proxy with the parser and the smart way of handling these key spaces, and the bring your own sharding... All this stuff seems like it would apply, at least conceptually, across to Postgres. But I guess at the nitty gritty it's just wildly different, or...?
+
+**Sugu Sougoumarane:** So the part that I'm still trying to figure out is -- so in other words, every idea that exists in Vitess will just smoothly port over to Postgres. But as you said, there may be details where it may not port as well. So there's a few -- at this point, we have made a few policy decisions, let's put it this way. The biggest policy decision is this has to be Postgres-native. So that part we do not want to compromise... Which means that the end product should feel like it should be built for Postgres. So that is one policy decision that we have made. I don't know how far we are going to get there, but that's our North Star of some sort... Which means that from Vitess, there are some parts we will definitely leave behind. For example, anything that's MySQL specific is not coming over.
+
+**Jerod Santo:** Sure.
+
+**Sugu Sougoumarane:** What else will we leave behind? We will leave behind anything that is legacy in Vitess. Anything that we built in Vitess but continue to support because there are users using it, but if we built it from scratch, we wouldn't do it that way... So that we want to leave behind. And also, anything that's not well implemented. If something is -- it works, so we left it alone, but it was not well implemented, given a choice, I would do it differently. So that we are not going to bring over. And everything else, we want to bring over, for sure.
+
+So there are actually -- and there's a fourth category, which is if there are things we can improve, we should. More specifically, I do want to improve the high availability story in Postgres, because that is currently a big challenge that users are facing. There is no good high availability solution, so that is something that we definitely want to improve on the Postgres side of things. So we'll bring some components over from Vitess, because Vitess already has some HA components... So we'll bring some of those components, but we'll also build something that is very specific to Postgres.
+
+And the place where I'm still confused about how to solve the problem - but we will figure it out - is "How do we solve this sharding and query processing"? Because Vitess can do query routing, Vitess has the concept of pluggability... The part that I'm thinking about is the Postgres pluggability is more at the binary level, where you actually build extensions and link them into Postgres... Whereas Vitess is built in Go, which if we brought that in, how do we make these plugins work? Do we have to translate them? Or do we figure out a way to make them still work if we ported the Go code over, is the part that I'm working through right now. Actually, we now have some people hired, so I'm working with some of those people, too.
+
+**Jerod Santo:** Nice.
+
+**Adam Stacoviak:** Does that compromise your desire to be Postgres-native, considering what you learn with Vitess, being written in Go, et cetera? How does that compromise your maybe "Postgres for life" - I might be putting words in your mouth, but like this native Postgres way?
+
+**Sugu Sougoumarane:** So if you look at the Vitess design, it was designed to be -- for the longest time, we didn't want to implement anything specific to MySQL. We thought this should be its own database system, and we actually restricted ourselves to just SQL 92. Because SQL 92 is generic, all databases must support it... Therefore, if we did SQL 92, then it'll work for anything. So we could even swap in Postgres underneath.
+
+\[42:01\] But what we found out is that people were hesitant to migrate over. They said "Oh, I want my application to run as is when I move over", which is what forced us to add MySQL extensions. So in other words, we could port the SQL 92 part of Vitess directly over into Postgres and it'll work. The problem is it's a subset. So if you do anything Postgres specific, it's not going to work.
+
+So I'm beginning to think that porting that means that yes, we can port it, but then we have a lot to build, a lot to catch up to everything in Postgres. And looking at 10 years from now or 5 years from now, when let's say we managed to get full compatibility of Postgres, we will still face the issue of these plugins, these extensions. So that's the problem that I want to think about right now, to make sure that when we get there, there is a solution for these extensions.
+
+**Adam Stacoviak:** Mm-hm. How important is it to reuse a lot of what you learned with Vitess in this new world? Because I've gotta imagine all this work you've put in, and countless hours of open sourceness etc. - like, there's a lot in there. How important is it to bring a lot of it over?
+
+**Sugu Sougoumarane:** The learning is totally valuable. I mean, without the learning, we wouldn't be able to build it. If you say "rebuild Vitess to what it is" - it is a lot of code, but it won't take the 15 years that it took us to build it. \[laughter\]
+
+**Adam Stacoviak:** Probably nine months, six months...
+
+**Sugu Sougoumarane:** Even if we had to rebuild from scratch, it may take us a year, maybe.
+
+**Adam Stacoviak:** Especially with AI, right? I mean, that might be a big help there. I mean, Go is really useful in code gen tooling, and stuff like that, so I've gotta imagine that that might even be a leg up for you to speed up the process, potentially.
+
+**Sugu Sougoumarane:** Yeah, we have been using Claude to analyze Postgres... Yeah, so you're probably right, we can probably do this much faster than even what I'm talking about, if we had to rebuild this whole thing.
+
+**Jerod Santo:** So there's this common movie trope, the retired badass... I'm not sure what this trope is called.
+
+**Sugu Sougoumarane:** \[laughs\]
+
+**Adam Stacoviak:** \[unintelligible 00:44:32.26\] retirement.
+
+**Jerod Santo:** Exactly. For one last --
+
+**Adam Stacoviak:** John Wick comes to mind, Jerod. I'm sorry.
+
+**Jerod Santo:** There's John Wick, Bruce Willis often plays this role...
+**Adam Stacoviak:** It's true, yeah.
+
+**Sugu Sougoumarane:** Or the Tom Cruise -- Top Gun...
+
+**Jerod Santo:** Or Tom Cruise, yeah. I mean, pick -- I'm not sure who would play you, Sugu, but you're coming out of retirement for this, because like "We've got to get --" What is it, like Armageddon? ...where like "It has to be these people." Which was funny, because those were like oil drillers, weren't they?
+
+**Sugu Sougoumarane:** Yeah, yeah, yeah.
+
+**Jerod Santo:** It was an interesting setup. But you know, they come out of retirement for that one last mission... You know, "I'm getting too old for this" is a common thing they say... And here you are. But you're coming out with Supabase. I mean, it's a different story. Supabase came from where? Obviously, the Postgres maxis were fans of Supabase, and I know Paul Copplestone, had him on the show a few times... And so we know about them, we know what they're up to. Obviously, PlanetScale was on the MySQL side... But why Supabase? Why come out and hook up with this ragtag group of Postgres maxis to build this thing?
+
+**Sugu Sougoumarane:** So I actually engaged with Copple on a whim, saying \[unintelligible 00:45:41.29\] So when I said "This problem needs to be solved, and needs to be solved now", my first thought was "I should start a company me. Let me do a startup." I actually started looking for co-founders, you know... I was even talking to Jiten, my PlanetScale co-founder. And I was kind of starting to engage this.
+
+\[46:16\] Then I thought, "What am I trying to achieve this time? I'm not trying to maximize profitability. What I want to achieve is give this project the best chance for it to succeed." And I was thinking, "If I do it as a startup, I'm bringing so much risk and so much distraction." Because then I will have to make sure that I raise money, I have to keep people happy... And instead, what if I found a place where I don't have to worry about these things, where I don't have to worry about raising money, where I don't have to worry about... And even -- if you're a seed stage startup, it's also hard to hire people. There are people who are really, really happy where they are, they are being very highly paid... How do you convince them to come and pile here, and take the risk to -- we had this struggle at PlanetScale, by the way, when we left Google and we thought "Oh, we know so many engineers, so hiring is not going to be a problem." And then we start the company and nobody wants to join us. \[laughs\]
+
+So I thought of that, I said "Maybe there is value in not going all the way to a big company, but at least a company that can pay people well, still give them a good upside..." And so I thought "I'll start talking to some people who are in the startup stage and see where they are." And that's how I engaged with Copple. And I was pleasantly surprised to find out he's almost a perfect fit for what I was looking for. And the biggest one is the open source part.
+
+**Jerod Santo:** Yeah. \[unintelligible 00:48:17.14\] open source.
+
+**Sugu Sougoumarane:** Even in YouTube, we kind of had to push hard to open source Vitess. For me, the success of an open source project is adoption. Open source and adoption is true success. And that's exactly what Paul believed in, Copple. So it was so perfect... And talking to him about all the other values, it was like -- at that time, it was almost like a slam dunk obvious that "This is a place where I should be building this."
+
+**Jerod Santo:** Nice. So it's not exactly like the movies... Because in the movies, he would have sent out his henchman to find you, and you would have been like --
+
+**Sugu Sougoumarane:** "Oh, I don't want to come back." \[laughs\]
+
+**Jerod Santo:** Yeah. You would have been like fixing up some old boat, or practicing Tai Chi, and they would have found you... And you wouldn't have wanted to, but you had to to save the planet, you know... But in this case, you actually reached out.
+
+You were ready to solve this problem once and for all Postgres users. And so that does make sense. I think that jives, Adam, with my impression of Paul and Supabase. Does that make sense to you?
+
+**Adam Stacoviak:** Yeah, I like the fact that it's Postgres-native, open source... I mean, that's the roots that Paul shared here. I think he said "The board of directors will have to fire me or pry this role from my cold dead hands --" One of those -- something elaborate like that to express his love for open source and his love for Postgres, Postgres-native.
+
+**Jerod Santo:** So how did it go from there? So he hired you, he's given you a team now, and he says "Have this on my desk in six months", or how does it work?
+
+**Sugu Sougoumarane:** I'm giving myself about three to six months to come up with an MVP.
+
+**Jerod Santo:** Okay...
+
+**Sugu Sougoumarane:** \[50:06\] It may vary. The reason why it may vary is because of my -- I don't want to compromise on the long-term plan. So if for the sake of delivery I feel like I'm deviating too much from where I want to go, I would rather delay the delivery. But I don't think it'll take us longer than three to six months to come up with an MVP. We have a very well-defined set of objectives to reach for the MVP, and I think we'll hit it all in about three to six months. And then iterate from there.
+
+The idea is that from there, it should only be iterations, no major changes. Like, that's the reason why I want to make the right decisions today, so that it becomes iterative from then on. So that is the oversimplified plan, but I'm pretty sure we'll have -- there are about five big features that we want to launch in this MVP.
+
+**Jerod Santo:** So is Go still on the menu then? Or does that not quite get you Postgres-native enough? I'm not sure if you can write --
+
+**Sugu Sougoumarane:** Which one?
+
+**Jerod Santo:** Golang. Can you write Postgres extensions in Go, or do you have to use the Cgo...? Or how does that work?
+
+**Sugu Sougoumarane:** So that's the part that we are debating about. I don't want to start a brainstorm here, but I can tell you what we are talking about... \[laughter\]
+
+**Adam Stacoviak:** Sometimes that happens...
+
+**Sugu Sougoumarane:** Yeah, yeah. So one option we are looking at is now Cgo. Like, if you have an extension, can we link to call you through Cgo, is one thought that we are having. The other thought we are having is, well, AI is so good nowadays... Can we like translate this for you, and run it as an extension within Vitess? That's the other thing that we are debating.
+
+**Jerod Santo:** Oh.
+
+**Sugu Sougoumarane:** There's the other crazy idea, which is not translate; leave Go and do it in C, directly. So all cards are on the table.
+
+**Jerod Santo:** Gotcha.
+
+**Sugu Sougoumarane:** The last one is the least liked one at this point, but it doesn't matter. Whatever is the right solution.
+
+**Jerod Santo:** Whatever gets the job done.
+
+**Sugu Sougoumarane:** Yeah.
+
+**Jerod Santo:** So that's fun. It's gotta be fun to be back at square one, to a certain extent, with a technical problem in front of you now. Like, a big, gnarly technical problem that's going to take three to six months for an MVP, and much longer if it's successful... Just to be back where you started, but in a different context, in a different part of your life.
+
+**Sugu Sougoumarane:** Yeah, and the best part is I get to fix all the mistakes I made in the previous project.
+
+**Jerod Santo:** Oh, yeah...!
+
+**Sugu Sougoumarane:** \[laughs\]
+
+**Jerod Santo:** What about the mistakes you're going to make this time? That's the problem. \[laughter\]
+
+**Sugu Sougoumarane:** You know, v2 is perfect always, right?
+
+**Jerod Santo:** That's right. You're not going to make any bad decisions this time around. Yeah, that's gotta feel good, because there's so many landmines that you just can avoid this time around.
+
+**Sugu Sougoumarane:** Yeah, yeah.
+
+**Jerod Santo:** That's awesome. What's your team looking like? Who are you looking for? Are you looking for hardcore Postgres people, or did you hire from Vitess?
+
+**Sugu Sougoumarane:** There are some -- so I cannot name names yet, because some of them are still in the pipeline... But it's a pretty exciting, pretty high-caliber team. Some are older Vitess contributors. There's a diaspora of Vitess contributors. Some are from there. And some are actually people that have handled this type of scaling problem before. So for whom Vitess is an obvious thing. So it's a pretty high-caliber team. I think we'll be announcing the team very soon. They're going to be kind of the founding team of this Multigres project, so we'll probably be announcing it in maybe about a month or so.
+
+**Jerod Santo:** Cool.
+
+**Sugu Sougoumarane:** But the hiring is looking very good at this point. I'm pretty optimistic.
+
+**Adam Stacoviak:** \[54:09\] What is implementation like in this new world? Is that still to be found out as you make these choices that have long-term permutations into maintainability, et cetera? Like, how will -- like, I know for the most part how Supabase works as a service. But if I wanted to use what you're calling Multigres, Vitess for Postgres essentially - if I wanted to use that on my own, outside of a Supabase context, what do you know about the future thus far to kind of showcase implementation details for end users?
+
+**Sugu Sougoumarane:** So in this, I should talk about something that we did in Vitess, which we are not going to do in Multigres. Vitess actually is what you would categorize as one of the most flexible enterprise solutions you can ever think of. If you went and asked a question to Vitess, "I want to change this very specific behavior of Vitess", there's probably a command line option for it. So that made Vitess extremely powerful and flexible. That's the reason why Slack could adopt it and completely deploy it using their rules, GitHub could use a completely different approach... And I believe Etsy has an even crazier approach. And Vitess works for all of them. But the problem it created by solving these things is approachability. If I came in and looked at Vitess, I wouldn't know as -- a layperson would just know "Oh, my MySQL is getting too big. I want to use Vitess." It'll take you a few months of learning how Vitess works before you can start to use it confidently. So I realized that's a problem. And so what I want to do in the Multigres is make it a lot more approachable. Actually, at the expense of even reducing those flexibilities. Because I think approachability is more important than flexibility. And if we add flexibilities, we want to hide them away from the user; they shouldn't have to find it until they need it. So that is kind of the approach we are going to use. And what that means is that it should be easy for a user to just take Multigres and just deploy it in their environment and it should just work for them.
+
+**Jerod Santo:** Do you know how that will manifest inside of Supabase? Or is that more on the Supabase product team to decide? Will that be invisible to Supabase customers?
+
+**Sugu Sougoumarane:** Yes, absolutely.
+
+**Jerod Santo:** Yeah. It makes sense.
+
+**Sugu Sougoumarane:** The idea is what we are going to be building is a Kubernetes operator. So if you want to deploy it yourself without Kubernetes, you may have to do a few things for yourself. But if you just took Multigres and use the operator, you specify "This is what I want", and then the entire cluster comes alive. And we plan to do that within Supabase also. So this operator will be deployed within Supabase. And all Supabase has to do is "Hey, bring up a cluster and we'll just spin it up for you."
+
+**Jerod Santo:** \[57:52\] Right. And they have to be operating at a significant enough scale that they can act as a test ground for you, I would think... Or maybe they can't do that with their customers, but... I'm just thinking, you had YouTube to build Vitess with. And that was awesome. You had a real scaled database, that had consistent users etc, and you could build the system. I guess in that case you built it incrementally, over time, as you figured it out. But building this in a vacuum - you've got to have some sort of real scaled things to test against. Maybe these are some of your core team they'll be announcing later, that I'm sure they're going to come from scaled up companies...
+
+**Sugu Sougoumarane:** Yes, yes. So there is a philosophy in -- so we had that luxury, and that's the reason... I think that's the main reason why Vitess succeeded where others have failed, is because nothing that would not scale could be even written in Vitess, because the next day YouTube would go down.
+
+**Jerod Santo:** Right.
+
+**Sugu Sougoumarane:** We have created a few outages due to those things. So if something was not going to scale, we would know right away. So we had that luxury. So everything that went into Vitess had to work at scale. So that was one thing. And the advantage we have is we have those learnings. So we know what not to do.
+
+**Jerod Santo:** Sure. But don't you want to test it as you go, and make sure that you're doing it right?
+
+**Sugu Sougoumarane:** On the testing side, what we over time developed - that is something we we had the luxury of not having to do in the early days at YouTube - is we had to develop really, really, really, really thorough tests. The reason was because at some point of time we would be making changes in the Vitess source code, and we had to be confident that people are going to take this. Slack is going to take this and run it. GitHub is going to take this and run it. It's going to come back into PlanetScale. And there is there is no waiting until it goes live to find out what what is going wrong... So actually, in Vitess, we had an extremely rigorous testing policy; extremely strict, with 100 percent code coverage, performance tests, backward compatibility tests... So that's actually one of the -- it also makes it hard for an engineer to contribute, because you change five lines of code... Now you have to write all these tests to make sure that you won't break anybody. But the confidence is people now have developed the trust that if Vitess releases this code, I can just take it and run and it will not fail. I'm hoping to bring that same philosophy here, which is to test thoroughly, to the extent that we are confident that this will work. And we'll obviously still ease that into production.
+
+**Jerod Santo:** Could you actually bring the test suite over, or at least parts of it, from Vitess? I mean, because that's a pretty robust test suite. If you could bring if you could port that somehow relatively easily...
+
+**Sugu Sougoumarane:** Yeah. Anything we can copy from Vitess, we are going to. Definitely, yeah. There are some parts which we'll just copy. There's no reason to change anything there. So those parts we'll just copy, yeah.
+
+**Adam Stacoviak:** Have you looked at deterministic simulation testing, similar to what Turso is doing? And what was the other database, Jerod, that you had \[unintelligible 01:01:27.11\]
+
+**Jerod Santo:** TigerBeetle.
+
+**Adam Stacoviak:** TigerBeetle, yeah. Have you looked into that, and does that translate? Or is the test base --
+
+**Sugu Sougoumarane:** I need to look this up. But in Vitess we probably have the opposite, which is called the fuzzer. Is that the same thing?
+
+**Jerod Santo:** Similar. Yeah, similar.
+
+**Sugu Sougoumarane:** \[01:01:46.24\] So we do run the fuzzer, and the fuzzer does not miss -- as far as I know, it misses nothing. I wasn't there when they implemented it at PlanetScale. It was someone called \[unintelligible 01:02:00.11\] who did that test suite. He said that the fuzzer was so good that it found bugs in MySQL.
+
+**Adam Stacoviak:** That's good.
+
+**Jerod Santo:** That's a pretty good fuzzer.
+
+**Sugu Sougoumarane:** So I have pretty high confidence in that. But yeah, whatever it takes.
+
+**Adam Stacoviak:** Well, we have a couple of past shows we can recommend for you as well that might give you some insights to -- just some of the behind the scenes; not exactly that how to implement it, but more fodder for your request.
+
+**Sugu Sougoumarane:** Yeah, yeah, definitely.
+
+**Jerod Santo:** Well, it's all very exciting. I assume the licensing is going to be similar to Vitess, or similar to what Supabase normally does. Is there any sort of gotchas in there, in terms of the open source side of it?
+
+**Sugu Sougoumarane:** No, no. Apache is what we're planning to do. Vitess was actually BSD once upon a time, until we joined the CNCF, and we changed it to Apache. But yeah, it's going to be Apache.
+
+**Jerod Santo:** Does Multigres make sense in the CNCF as well, or no?
+
+**Sugu Sougoumarane:** It's an option. It's definitely an option. We even thought of either a fork, or you know, it could be a subproject of Vitess... But then I felt that that will restrict its freedom right now. Starting it as an independent project -- these are very difficult decisions, and I think once we make these decisions and bring it to a good shape, we could look at actually moving it back to CNCF as its own project.
+
+**Jerod Santo:** Yeah, I think you want the space to reinvent and rethink without feeling at all burdened by your past decisions with Vitess.
+
+**Sugu Sougoumarane:** Correct, correct.
+
+**Jerod Santo:** And I think -- obviously, copy and paste the stuff that makes sense... But don't make it a fork where you're like basically starting at this foundational, which you may end up wildly different in implementation. Or you might be the same, but at least give yourself room to come there organically, you know?
+
+**Sugu Sougoumarane:** Yeah, that is something that was pretty obvious to me... Although the fundamentals are the same. Even if you don't copy code, you can copy ideas pretty easily.
+
+**Jerod Santo:** Sure.
+
+**Sugu Sougoumarane:** So those don't change. But when you get into the details, there's so many small things... And the thing is, if you don't do them the Postgres way, that will keep nagging you. You know, "I have written this... There's this if statement hanging there, which is there only to differentiate between Postgres and MySQL."
+
+**Jerod Santo:** Right.
+
+**Sugu Sougoumarane:** You don't want any of that.
+
+**Jerod Santo:** None of that.
+
+**Sugu Sougoumarane:** Yeah.
+
+**Jerod Santo:** That kills the joy part, right? It takes away the joy that you're trying to enjoy what you guys are building here. Well, it's very exciting. Obviously, you're just getting started, so there's still a lot of question marks for us Postgres fans... It's just a lot of waiting, watching maybe the GitHub project as you put it out there, and the announcements as the core team comes together... You get the Vitess Avengers, you get the whole team together... But what else? What else is on your mind, or things that are important for folks to know before we let you go?
+
+**Sugu Sougoumarane:** The thing I feel is I feel a bit anxious that there's not much activity in the open source project. We may seem cool, but we are paddling hard underneath. Really, really hard. So I will be actually publishing a few blogs about some of the thoughts... There's one area which we didn't cover much of, which is the high availability and the consensus part. So actually, my initial blogs will be covering those.
+
+\[01:05:58.20\] I feel like this is something -- the database world kind of went a little bit astray with mounted storage, disaggregated... I feel like that needs to be brought back in. I feel that storage should be with the database, because database are IOPS-hungry, and having your data local to your database is very important. And that story can be complete only if you have a good consensus story, where your data will not be lost if you lose your node. So that is something that we want to solve really, really well in Postgres. Because I have seen the good days of the database, where the database goes screaming fast. Slack and GitHub, they run hundreds of terabytes of database size, they run millions of QPS, and their latency is a millisecond; one millisecond, two milliseconds... That's the kind of latency they get. And they can't tolerate any higher latency. And I want to bring those days back into Postgres. So that is kind of what I'm shooting for.
+
+**Jerod Santo:** So are those things that Vitess tackled or did not tackle? High availability...
+
+**Sugu Sougoumarane:** Yeah, those are very core to Vitess.
+
+**Jerod Santo:** Yeah, I thought so.
+
+**Sugu Sougoumarane:** Those parts we are going to bring into Postgres.
+
+**Jerod Santo:** Right. Interesting. Adam, anything on your mind? Any other questions?
+
+**Adam Stacoviak:** I'm thinking about Neon, honestly. I'm thinking about just this kind of bigger picture... I'm curious if this shake-up, this change in the database world -- like, for a while there we had obviously PlanetScale leading, massively scale MySQL, which was great for MySQL, but not so great for the Postgres folks like us... Then we've had Supabase, obviously, we've had Neon... Neon was acquired by Databricks... I don't know how that's going to shake out for a product that's usable by the public, or if it's a Databricks thing... This seems like a good time for Supabase to do this, because there's change amongst the Postgres world. And I feel like this is the clincher, really, for Supabase to finally have this kind of feature set that wasn't available unless they built it. And why build it in a non-Postgres-native way, in a non-open source way, given Paul Copplestone's way of thinking as a leader there. I'm just curious your thoughts on that, Sugu... It's just the thought of like Neon, Postgres, Supabase, this shakeup, you coming back out of retirement, basically, to avenge this scenario and shard for life with Multigres... Help us unpack this shakeup, I suppose. Are you excited? Obviously, you came back, but how does Neon's change change the landscape for Postgres specifically to open the floodgate for Supabase's mission with Multigres?
+**Sugu Sougoumarane:** Yeah, going back to why I started this project, right? The reason why I started this project was because I believed - I still believe, actually - that there is no scalability solution for Postgres. Not Neon, not Aurora. I'm an advisor to Metronome, which is a Postgres shop. They have one of the largest Aurora instances, and they are struggling. They have no way out, beyond where -- they are looking at moving data out of their Postgres instance, because they've hit the limit. And now they're saying "Oh, we need to do something about this scalability thing." And that's the reason why I wanted to start this project.
+
+\[01:09:57.00\] And one of the things I thought with Supabase -- so this is a problem that is meant to solve it for the industry overall. In other words, Neon could use Multigres if they wanted. If they hit the limit of a machine, you could deploy Multigres on Neon and run it... Unless there were incompatibilities, for which you had to make changes. But even in those cases, the changes would be minimal. So that's what I came for.
+
+And for Supabase specifically, is that if Supabase is going in its trajectory, there are going to be users that are going to max out a single instance very, very quickly. As a matter of fact, I won't be surprised if some are already hitting those limits. And their way out is you could go to Aurora maybe, and that will extend your runway by maybe at best 2x. I'd be surprised even if you got like a 2x runway. As soon as your QPS doubles, you pretty much are at your limit.
+
+So the only way out of this is a sharded solution. And that is why I started this project. In MySQL, if you hit your limits, you could go to Vitess, but there is nothing like that in Postgres. And Vitess for Postgres, at even mediocre level of functionality, will allow these people to stay in Postgres. Otherwise, you have to find a way out of Postgres, once you hit that limit.
+
+**Jerod Santo:** Well, after you accomplish Vitess for Postgres, maybe you can build Vitess for SQLite. No, I'm just kidding. \[laughter\]
+
+**Adam Stacoviak:** He almost punched the screen, Jerod. He's like "I'm gonna punch my screen right now."
+
+**Jerod Santo:** After your next sabbatical, you know... The only thing that gets you back is that Dr. Richard Hipp approaches you and says "Sugu, we need you. We need you at SQLite."
+
+**Sugu Sougoumarane:** \[laughs\] That reminds me of -- I think Airplane, was it? ...where there's a background scene with a movie poster which says Rocky 83, and the really, really old guys \[unintelligible 01:12:24.23\]
+
+**Adam Stacoviak:** Airplane. You're going back in the day, into the '80s, to mention a movie...
+
+**Jerod Santo:** I haven't seen Airplane for years.
+
+**Adam Stacoviak:** I mean, don't call me Shirley, okay?
+
+**Jerod Santo:** Exactly. Good stuff.
+
+**Adam Stacoviak:** Yeah, good stuff. Sugu, thanks for coming back from retirement for this. I know that -- I mean, considering what you've just said, that there's no way out, that sharding is the only way... I mean, you're coming back at the right time...
+
+**Jerod Santo:** Saving the Postgres world...
+
+**Adam Stacoviak:** ...I think you're coming back to the right team, with the right motives, at the right time... And somewhat full disclosure, Jerod and I are both angel investors in Supabase, so we kind of have this desire for our own upside, but like technologically, this would be amazing for Supabase to essentially be one of the only leading at scale Postgres flavored, Postgres for life, open source for life, flavors out there to choose from. And so I'm super-excited about that, for you, and for them, and for Postgres, because... We need it. It's amazing.
+
+**Sugu Sougoumarane:** I am excited, too.
+
+**Jerod Santo:** Get building, man. Get building.
+
+**Sugu Sougoumarane:** Get building.
+
+**Jerod Santo:** Alright. Thanks, Sugu.
+
+**Adam Stacoviak:** Bye, Sugu.
+
+**Sugu Sougoumarane:** Thank you.

Build software that lasts! (Interview)_transcript.txt

@@ -0,0 +1,529 @@
+**Jerod Santo:** We are here with Bert Hubert, a geeky entrepreneur from the Netherlands, with a 30-year track record in commercial and open source software. Bert, welcome to The Changelog.
+
+**Bert Hubert:** Yeah, thank you for having me.
+
+**Jerod Santo:** Happy to have you. Excited to talk about what you're up to these days and what you've been up to historically. 30 years - that's a long time. I read that you landed your first job hacking a cable internet provider? Do you want to tell that story?
+
+**Bert Hubert:** Yeah, it was a fun story. So we lived in this university town and we were quite late to get the cable modem. And it may be interesting to know that in Europe, in many places, if you had dial-up internet, it was actually charged by the hour. So we had no local free calls as in other places, so it was a very big deal to get a cable modem. And when I got it, it was very much delayed, and the cable company had issues with it... And when they had it going, I started scanning their infrastructure, and I found out that while they had finished setting up their internet, they had not finished setting up their security. So I could sort of waltz right into their servers, and stuff... And so I sent a message to the system administrators using the famous wall command. I said, "Hello, I'm here." And they said, "Well, if you're so good with security, then come over."
+
+And so I got on my bicycle - this being the Netherlands - and I drove to the cable company. And on my way there, I was like "Yeah, this could end up really well, or I could end up being arrested", because I'd just waltzed into their systems. And it ended up really well. A little bit too well, because that was the dotcom boom times, when there was such a shortage of staff that I came in, they gave me a chair and they said "Sit down." And that actually ruined my studies of physics, because then I had this job over there.
+
+But it was quite cool. It was very nice to scale an internet service provider from like 50 users to 50,000 users. Everyone should maybe one day scale a company by three orders of magnitude. It's very educational. So I had a great time there. And it was a wonderful way of sort of learning how to run an internet service provider from scratch, which is very educational. Because if you mess it up at that stage, it's not so bad. But if you mess it up by this paradigm and you're at Comcast scale, it is pretty bad.
+
+**Adam Stacoviak:** Yeah, you can't mess around at that stage. You've got to have the uptime, right? All the nines.
+
+**Jerod Santo:** So what were some of your learnings as you scaled up? These are kind of like the first time things are scaling up, so you can't exactly go read a book about it, right?
+
+**Bert Hubert:** Yeah. I actually wondered if anyone had written the book about it, but I don't think we had that rate of scaling that we had in 2001, where you would double every month, or something like that. One of the main things I learned there is that it was quite amateur stuff. So you need to professionalize and figure it out.
+
+For example, at one point I did take down the whole internet service provider, because I had fixed a bug in the LDAP server and I was very pleased with myself that I had fixed that bug... So I shut down my laptop and I went home. And I had forgotten that I left the LDAP server running in the foreground. And the moment I shut down my laptop, the whole internet service provider shut down. And on my bicycle back home again, I got already -- I was barely out of the building before they called me and said "What did you do?" And so I had to rush back in. So I learned a lot there, in the sense that you really need to -- it's nice to hack yourself around these bugs and then feel really good, but it's also maybe really good that once you fix the thing, then to also sit down and really tighten things up again, and not just shut down your laptop and shut down the whole internet service provider.
+
+**Jerod Santo:** Well, unintended consequences, right? You just don't know what's going to happen. You shut that laptop, think you're done for the day...
+
+**Bert Hubert:** Yeah. Well, there is one really interesting thing that I learned there, and I've been passing on this message since forever, and I'm also going to pass it on to your listeners.
+
+**Jerod Santo:** Okay.
+
+**Bert Hubert:** \[08:00\] So we had a redundant email setup. And we thought it was good, it was redundant... So there was this guy in charge there and he said "Well, are you sure this is redundant?" I said "Yeah, it's all redundant." And then he pulled the plug on one of our servers. And then everything broke, because it turned out that it was sort of -- yeah, we felt that it was redundant, but we had not actually had someone pull the plug on us, and see if it actually was redundant. And then it turned out that it was not redundant. And since then, every time someone says "Hey, we've built this redundant system", I say "Can I just go in there and start shutting down some random servers now? Is it going to fly?" And they said, "No, no, no. Don't actually be doing that." And I'm like "Well, maybe you have to work a little bit more on your redundancy, until you can live with me going in there and just shutting down some random stuff to see what happens." So that has always stayed with me. If someone says it's redundant, I'm going to ask you "Can I just shut down some stuff for you then?"
+
+**Jerod Santo:** Yeah.
+
+**Adam Stacoviak:** That's not a good thing, you know. Pull the plug, things fall over.
+
+**Bert Hubert:** Yeah, yeah. Well, it's a good thing. I mean, if people say "Look, it's sort of redundant in the sense that we can make it work again", okay. That's fine. That's my level of expectation that something could fail over and you could rapidly make it good again.
+
+But if you're telling me that "Look, this is going to be automatically redundant, and it's going to work anyhow", then I should feel free to just unplug some stuff to see what happens.
+
+**Adam Stacoviak:** We all learn those lessons to some degree, right? Like, whenever you think something is safe and secure, or redundant, or whatever you might want to call it, and then chaos engineering happens, essentially, you know?
+
+**Bert Hubert:** Yeah. And you only learn that, you only learn that -- recently I've been... I mean, we might come to that. Recently, I'm now running some actual services again that people depend on 24 seven, and everyone has to sort of live through that for a bit. It's experience being on the line. Like, people are relying on this stuff, and I should be doing a good job here. And that's a whole new level of thinking that you get, which is very different from a programmer that just programs and never gets those 3am phone calls.
+
+**Jerod Santo:** This is why I've been trying to convince Adam to run our Zulip community from his home office... Like, just host it for us. Self host that sucker. We've only got a few hundred people using it, but they're using it all throughout the day.
+
+**Adam Stacoviak:** For the learnings?
+
+**Jerod Santo:** Yeah. And for that upgrade of yourself, just what it feels like to have people depending on your service. It changes you.
+
+**Adam Stacoviak:** Okay... I'm not against it necessarily. It's just I prefer not to.
+
+**Jerod Santo:** \[laughs\]
+
+**Bert Hubert:** I mean, I get that feeling. I really understand, because it really sucks. Because once you have that stuff running in your house, you have to start planning around things, because you can not just have an electrician come over and redo your wiring and be like "Yeah, we're going to be down for three hours." I mean, it does make you think.
+
+**Adam Stacoviak:** Yeah, for sure.
+
+**Jerod Santo:** So straight out of university, I ran some mail servers on a Linux network for a company, maybe a few hundred people. And there were two servers, and they worked in conjunction, and there was Spam Assassin, and it's all kinds of stuff. There was PostFix... And when we had issues, it would just queue up. As long as the servers were online, even if a mail server is offline, it comes back on, that delivering server, they're going to try again a little bit later. And so it would just queue up. And so there were times where I would come into work and I would look at the mail queues and people are thinking it's a quiet morning, like "I don't have any email this morning." It's like "Oh, you do. I just haven't been putting them in your inbox."
+
+**Adam Stacoviak:** \[laughs\]
+
+**Jerod Santo:** And then I used to love just watching that queue just work its way down back to zero. I definitely know the feeling of other people relying on your network services. It can be stressful.
+
+**Bert Hubert:** It can be. It can be.
+
+**Adam Stacoviak:** Yeah... Unnecessarily, if it's unnecessary.
+
+**Jerod Santo:** Right. If you've got someone else to do it for you, why do it, right? Bert, what are you running these days? You said you're running some stuff now?
+
+**Bert Hubert:** Yeah. So I built a sort of parliamentary monitoring system. So we have also an upper chamber and a lower chamber in the politics house here, and they have a pretty competent website. So it's not that it's bad, but half a year ago at one point there was a new law here in the Netherlands, which was quite impactful for the internet community, and we hadn't noticed. So the law was about to enter into force and we didn't know about it.
+
+**Jerod Santo:** \[12:19\] Okay.
+
+**Bert Hubert:** And luckily, the law was not so bad, but we only had a few days to look at it... And we knew that if it was bad, we would not have been able to do anything in those few days. So back then I said "We can never have this happen again."
+
+And on the Dutch parliamentary website you can put some search terms and it will send you messages if any new documents match those search terms. It's a bit clunky, it's not that great. So I thought let's make some kind of advanced warning monitor that we on the internet can put up searches and say "The moment you have a meeting about internet censorship, I want to know about it." And not only do I want to know when it actually happens, I want to know as many months in advance as possible. And from that, I built this monitoring website, and it grew and grew and grew...
+
+And actually, the sort of fun thing is if you are outside the government, you can innovate rather quickly. And the Dutch parliament is completely open data. So everything they do, you can effectively monitor the state of the parliamentary documentation system to your house, which I do. And that grew and grew and is now quite popular, in the sense that even many politicians are using it, and many civil organizations are using it. And the most interesting thing, and which I find most important, it turns out that my simple parliamentary information website - and it's like really straightforward; plain HTML, almost no JavaScript - works really well with software for blind people. And it turns out there are some government workers that have to work with these parliamentary documents, and they've found out that their screen reader software has a lot of fun with my site.
+
+**Jerod Santo:** Easier to read on your site than the official site.
+
+**Bert Hubert:** So I now have a few low-vision government workers that are actually relying on my site to do their job. And that has added to my sense of that is not just a hobby anymore, it's just that people actually need this stuff. So from that, I learned all these very interesting things that you only learn when you go live with services, with things that you don't know.
+
+One thing, for example, that I learned is that if you have an iPhone and someone enters a search term in your form, and you do that within quotes, because you want to search for a sentence, the iPhone will turn it into smart quotes. And that is not something your search engine is ready for. Your search engine doesn't recognize these smart quotes as smart quotes. So I had this weird situation that iPhone users couldn't find anything. And this is the kind of thing that you only learn in production.
+
+And the other thing I learned - I have these signup emails, passwordless login email, so you can just log in with your email address, which is quite popular these days... And for some people it didn't work. And it turned out that Microsoft now runs a security scanner that will actually attempt to log in for you.
+
+**Jerod Santo:** We know this very well. We were just talking about this topic on the previous episode with somebody... But you have new information now, because we have the same thing, passwordless logins on our website. And what my trick to fix that was, was I switched the actual requests from a get to a post, but I read on your website that that's not going to even work anymore, because Microsoft is now executing JavaScript in their --
+
+**Bert Hubert:** Yeah, they're posting it to your website.
+
+**Jerod Santo:** They're now posting.
+
+**Bert Hubert:** \[15:54\] And the weird thing is -- so the strange thing is they do the post, which is already, I think, violating many people's assumptions. You should not be posting on behalf of anyone else... But the other thing is, when they do that post, my site actually used to return a cookie, a session cookie, which means that Microsoft with this security measure... So the reason they do this is they want to see "Is there malware on this site? And might that malware only pop up after a post?" I see where they're coming from. But when they sent that post to you, my site would use to respond with a session cookie that said "Welcome. You're logged in now." Which means that Microsoft is receiving tons and tons of these session cookies right now.
+
+**Jerod Santo:** Oh, cookie monsters.
+
+**Adam Stacoviak:** Literally.
+
+**Jerod Santo:** \[laughs\]
+
+**Bert Hubert:** These cookies are very valuable, because these are the session cookies that allow you to do stuff. Well, it now appears that the new barrier is they will execute your JavaScript, they will execute your posts, but they will not, for now, click on a button. So you must have a button in there right now, and that button then does the post.
+
+**Jerod Santo:** Okay. So you are actually -- you're adding a step for the real people, because now they have to click on the link in their email, and then come to your site, and then click on a button to sign in.
+
+**Bert Hubert:** Yeah. But Microsoft did not announce that they would be doing this, and they have also not announced that they're not going to click on buttons. So maybe one day they will click on buttons.
+
+**Jerod Santo:** \[laughs\] Well, you need a monitoring website where you can monitor Microsoft's changes...
+
+**Bert Hubert:** Yeah, well... But it is the kind of thing - and I cannot stress this enough... This is the kind of thing that you learn when as a programmer you go into production again.
+
+**Jerod Santo:** A hundred percent.
+
+**Bert Hubert:** And I've since heard many people, they told me that Trend Micro also does this... And I actually ordered some hardware stuff from a store today and they have a link that is vulnerable to this. You have to click, and it says "I'm going to collect my hardware now." And that is already useless for them, because Microsoft is doing all the clicking right now.
+
+**Jerod Santo:** Because Microsoft's already registering everyone's hardware for them.
+
+**Bert Hubert:** Yeah.
+
+**Jerod Santo:** Yeah, it's a crying shame. It's tough out there in the real world of the internet, isn't it? I'm going back to your monitoring site... What formats does the government put stuff out? Are you like diffing HTML, or is it better than that?
+
+**Bert Hubert:** Oh, this is a story... This is a story. So on the one hand, they have a glorious API. And actually, I didn't read the manual. They use this thing called OpenSync, or something like that... And that is apparently a sort of weakly determined standard by which you can replicate a relational database to somewhere else, as a series of XML changes. And you can pull these, you can say, "I want to get all your changes since marker such and such." And that's actually pretty nice.
+
+So it is quite convoluted, because I think it would have been easier if they just said "Look, this is our SQL database and you can query it." But no, you get this stream of XML messages. And that is actually quite glorious and good. Now is where the problem comes. Governments, when they send you documents, they love PDF. PDF has a lot going for them. You have PDF/A, which is an archival form of PDF, which is actually nice, because that's the sort of PDF variant where you can say "I'm sure I'll be able to read this PDF 50 years from now." So it only uses built-in fonts, it has no language, it doesn't do anything... So governments really love their PDF files. The thing is, web users do not love their PDF files. Web users want to see web pages. They want to see HTML, they just want to click... And the thing is, when the Dutch government gets a document, it starts life as a Word document, so docx, which is easily processed. Then they convert it to XML, which is also easily processed. And then finally, they convert it to PDF, which is not so nice to process.
+
+\[20:00\] So in order to make all this stuff work, I've found out that there is a sort of official government archive of government documents, and there they also publish the XML. So in the course of this trajectory, I retrieved the document like four times, and to save it, I have to retrieve it from some kind of official government archival site where there is XML. And that XML actually is glorious. So when someone speaks in Dutch parliament, they make a note at the exact timestamps when they speak. And this is how they can match it up with the video.
+
+But it also allowed me to make these statistics to say "Who are our Congress people that talk the most, or that talk least, or that have the longest sentences?" And because they log it all so well, I could also "Who are the fastest talkers of our Congress? And who are the slowest talkers?" And this is of course not very necessary, but it is a lot of fun. I never used to be a fan of XML, but it actually does get the job done in this case. It's actually not so bad.
+
+**Jerod Santo:** Yeah, relative to PDF, it's better, right?
+
+**Bert Hubert:** Yeah. Well, everything is better than the PDF. I mean, people -- I understand that people sort of love it because it feels like it's a standard format, and you have everything in there that you need... But to process it, it's nasty. For example, if you have a two column PDF file, it's actually not easy to figure out that it is a two-column PDF file. Because if you look at the postscript, it's just a sentence, then a bunch of spaces, and then another sentence, which is the next column. So it's all not that great.
+
+**Jerod Santo:** So do y'all have the filibuster over there?
+
+**Bert Hubert:** Actually, no, we don't. And I try telling people that our politics are also becoming wilder, and less useful, but we haven't yet sunk to the level that we need a filibuster kind of thing. So for now, if you have a simple majority, you can actually get stuff done in Dutch government... But I worry where it's going.
+
+**Jerod Santo:** I always thought the filibuster was just the dumbest thing. It's like, "If we can just talk long enough, we can run the clock out." It's so silly.
+
+**Bert Hubert:** Yeah, yeah...
+
+**Jerod Santo:** But I ask that because - then do you have any blowhards? Do you have a ranking of "Here's our most wordy politicians, versus not"?
+
+**Bert Hubert:** Yeah, we do. We have these people that actually -- well, they will talk your ears off... And actually, it's not always who you think it is. It is sort of funny to -- when you run these numbers, you find out that it's actually quite often that it's different from what you thought it would be.
+
+**Adam Stacoviak:** What we all need in government is a haka.
+
+**Jerod Santo:** Oh, excuse me?
+
+**Adam Stacoviak:** A haka.
+
+**Jerod Santo:** What's this?
+
+**Adam Stacoviak:** I may or may not be pronouncing that properly... Have y'all seen this in New Zealand? This tail end of last year?
+
+**Jerod Santo:** No.
+
+**Bert Hubert:** Yeah, no, that was very impressive. And they should do more of that.
+
+**Adam Stacoviak:** Oh, my gosh... I'm not sure the context of the protest, but very much a spectacle. And I think we all need versions of that in ours. Describe it, Bert. Describe it in your best words.
+
+**Bert Hubert:** Well, that is in New Zealand... That's what you're referring to, probably.
+
+**Adam Stacoviak:** Yes.
+
+**Bert Hubert:** Yeah. So they had a law that was passed which was not good for the indigenous people in New Zealand, and then they got really upset. And then the indigenous people or the people affiliated with them in Parliament actually performed a whole song, a protest song in government, which is really a very important thing in New Zealand. But it was, I think, a great way of making a point. And I cannot imagine US Congress breaking out into a song.
+
+**Jerod Santo:** Oh, goodness gracious. Please, no.
+
+**Adam Stacoviak:** This is a may-be a song, but I would not describe it necessarily as song.
+
+**Jerod Santo:** \[laughs\] It doesn't sound very nice?
+
+**Adam Stacoviak:** It's more like a screaming song, and like a dance, and a performance...
+
+**Jerod Santo:** A chant, maybe?
+
+**Adam Stacoviak:** Like a chant, yeah. It's probably more akin to a chant.
+
+**Jerod Santo:** Right.
+
+**Adam Stacoviak:** Very performative. One person began it...
+
+**Recording:** \[24:00\]
+
+**Adam Stacoviak:** A great demonstration. I think it's turned into a meme in terms of how to protest... She tears up whatever it is they're protesting, and then many people that are in the Parliament space begin to join in with this dance, that turns into song and dance.
+
+**Jerod Santo:** Wow.
+
+**Adam Stacoviak:** And so it's very -- it demonstrates very well. So if you feel very strongly against something, clearly it's got a very visual and a visceral process to like make you pay attention, you know? So it's interesting... You should check it out, Jerod.
+
+**Jerod Santo:** I will. I'll go check that out.
+
+**Break:** \[24:47\]
+
+**Jerod Santo:** Well, one thing about governments, it seems, as they go on, they either incline or decline. And many of them decline. And this is a very poor way of segueing into the topic of long-term software development, which is something you've been interested in, Bert, and I'm also interested in, as we have systems that exist for decades and become quite crucial. Software has never been more crucial in the life of everybody out there in the world than it is today. And a lot of these systems are not built to last, or they weren't built to last, or they thought they were building it to last, but they actually weren't... And you've done a lot of thinking, and even writing and talks about long-term software development, and I think we would all benefit from hearing your learnings and what you've been talking to other folks who've been experienced or believe about this. We all have thoughts about how you actually write for the long-term. I'm curious what you have found.
+
+**Bert Hubert:** Yeah, so I was invited -- so I'm a very part-time technical advisor to the Dutch government also, to what you would call the FEC, the electoral board... And they built this software for tabulating the national elections. And they are revamping it, and they said "Would you take a look at our code? Can you tell us what you think about it?" And I looked into it and I've found that they were doing sort of more or less standard 2025 software development... And then I wondered -- and that means having like 1,600 dependencies. Because if you have a Node.js built project these days, and you install it, it will, for starters, pull in like a thousand dependencies. And then you haven't done anything yet even.
+
+So I looked into that and I thought "Well, this is not going to be something that's viable for the next 10 years", because these dependencies change all the time, and npm might go away for the next 10 years... So I turned that into a study, "What would you do these days if you had a 10-year software project?" And I went to Mastodon; that is a really good social network to ask these kinds of questions, because it's full of nerds, and they're my people... And so I asked them, I said "Look, what are your thoughts, for anyone doing long-term software development?" And we had people weigh in that have a software with an 80 - and not 18... 80-year horizon. And these are people that study, and have to measure the radioactive decay of nuclear waste. And they are tasked with measuring and supporting that for the next 80 years. So they spent quite some time thinking about this stuff. And it turns out, actually, it's sort of modern software development. It's like sort of the complete opposite of what you would do if you want to have software that works 10 years from now... In the sense that we have these huge -- that even building software is fragile right now. So not even that it does the same thing, but if you want to build a sort of standard 2025 software project, you need a working internet connection, and like hundreds of servers around the world that supply you with data, even before your software runs. And that is not something that is a healthy thing to have if you want to promise people that it will work 10 years from now.
+
+So I got a bunch of very strident feedback from people. They said "Well, if you look at the problems that people have with old software, something goes wrong with the software, there is a problem..." Something always goes wrong, of course... And then in your mature software project, you need to figure out what is actually going wrong. And what almost everyone said is keep it simpler. Of course, we all know that we should not write software that is as complicated as we can make it, because that's not going to end well... But everyone said "Look, we've had a very bad experience. We needed to figure out seven years ago why this clever code -- what it actually does." And then often you find that this clever code, there was no need to make it clever. So if you have like 10 political parties, or -- in the Netherlands \[unintelligible 00:31:15.12\] we have 25 active political parties. So that's -- I mean, you have two?
+
+**Jerod Santo:** Yeah, can we borrow some?
+
+**Bert Hubert:** Well, having 25 is not an optimum, I can tell you that. But even if you have 25, there is no need to set up a complicated data structure to hold 25 political affiliation names. But one day you might sit there and say "Hey, wouldn't it be useful, wouldn't it be fun if we had a complicated red-black tree so that we could make really rapid searches of our 25 political parties?" And then you sit there maybe in 2032 trying to debug why it doesn't work, and that's because you tried to be clever. And so many people came to me and they said "If I have one regret, it's that we built things too complicated, and then seven years later no one understands the code anymore."
+
+\[32:06\] And the other thing that people came up with is that people change jobs a lot, for a number of reasons. But if you want to do long-term software development, the easiest way to do that is with long-term employees. Because they build up knowledge, they know how the stuff works... But if you have people that are changing in and out all the time, then you become very reliant on good documentation. And it's, of course, always good to have good documentation. I mean, by all means, do it. But it's very hard to compensate for all that knowledge walking out of the door after two years.
+
+So people also said, "Look, you really need to--" I mean, it's old fashioned and it's almost crazy talk these days, but that you really need to just invest in keeping your developers happy, and wanting them to stay for like maybe seven years, or eight years, or something like that... But again, this is also not how many places work these days anymore, because they have contractors, and they zoom in and they zoom out...
+
+So as a result -- so I wrote the article based on my presentation to the Dutch electoral board, and it hit Hacker News, and other people... And amazingly, almost everyone agreed with it. It's very rare that that happens. But like 30,000 or 50,000 people read it. And based on that, I got a lot more feedback from people that said "Look, we are doing nothing like this here. We are following sort of the Facebook model of software development, where we deploy all the time, and most days of the week it actually works. And sometimes we go through a bad patch." But then there are all these kinds of people that have to make software that opens and closes bridges, and doors, and MRI scanners that could physically tear themselves apart, or pacemakers, or robots that actually do surgery. These people are like "No one understands us anymore. Because if we do the regular software development thing, they start with stuff that is so fragile that we would never entrust that to open or close a bridge, or launch a rocket, or whatever."
+
+So it's actually becoming quite a rare skill to build the kind of software that has like safety of life consequences to a bunch of people that will start with downloading the latest JavaScript framework... Even though that might be very nice for your website.
+
+**Adam Stacoviak:** A lot of great stuff in here. Dependencies, choosing those wisely, keeping it simple, obviously... These are all, I suppose, tried and true; kind of, to some degree, well known. So it's not like you've reinvented the simplistic wheel. You've just sort of collected what might be a good zeitgeist for folks to follow.
+
+**Bert Hubert:** Yeah, I collected very -- it is a list of extremely obvious things. But once sufficient people have sort of forgotten about the obvious things, it starts becoming worth your while to just restate the obvious things, and say "Look, you're not crazy. You're not --" I mean, because many of these things are obvious, but also sort of old-fashioned.
+
+**Jerod Santo:** Right.
+
+**Bert Hubert:** I think it was worth restating. Also, it has strengthened the spine of a number of people; it has reinforced them to say, "Look, let's work on keeping it simple. Let's refactor this stuff, and let us not optimize these things before they need to be optimized."
+
+**Jerod Santo:** \[35:40\] Right. That's a big one. It's easy to get nerd-sniped into premature optimizations. If we just focus on the simplicity part, let's return to dependencies, and maybe talk about that in detail... But I don't think very many people set out to make a complicated solution. Like, it's not -- unless you're just malicious, or like willfully ignorant, or something. You're not like "I'm going to make the worst possible solution ever." We all want to make it simple, and straightforward, and repeatable, and reusable... These are all things that engineers, software developers strive for. And yet, it seems like we all end up making things that are too complicated anyways, right? Nobody's like "I'm going to make a bunch of spaghetti code, and it's going to be a terrible mess four years from now." And yet we end up there. So do you have -- did people come up with ways that you could actually like deploy a strategy or a tactic towards simplicity?
+
+**Bert Hubert:** Yeah. So I would not actually agree that we -- we all say we like simplicity, this is true. But there are, for example, people that take a language standard and see it as a personal challenge to use all of it.
+
+**Jerod Santo:** \[laughs\] Okay... Well, because maybe that demonstrates proficiency.
+
+**Bert Hubert:** Yeah, yeah. It's like "Look, I'm really good at this stuff... See how complicated this code is." And I think the only people that do this correctly are the people that feel it in their bones. That when they're typing in something complicated, they worry already about the 3am debugging session that's going to ensue five years from now. Because they've been through it.
+
+**Jerod Santo:** They've been burnt before.
+
+**Bert Hubert:** Yeah. And so I think there are actually a bunch of people that really -- maybe it's controversial, but for example, programming in Rust is all the hype these days, and it has many things going for it... But when I tell people, I say "Look, it's quite a complicated programming language", they actually like it that way. It's a challenge to write in there, and some people actually sort of enjoy writing challenging code.
+
+So I think the best way of writing simple code is having been on call. That is like the one way of keeping yourself pure, because you know "If I mess this up, if I write this too complicated, it's going to cost me my sleep one day."
+
+**Adam Stacoviak:** Basically, deal with the mess. You do that in product, you say "Hey, do support for a bit. Hey, if you're an engineer, be on call. Deal with the mess that you create." Because it's pretty easy to write something, throw it over the wall and walk away, because you're not on call.
+
+**Bert Hubert:** Yeah. And in many organizations, there is actually a physical wall. It's just the people that write the code will never -- they will only indirectly hear about it if there are any problems with it. And that means that there's a natural tendency to keep it simple, because you know you're going to be stuck with the mess.
+
+If you put an organizational barrier between dealing with the mess and writing the mess, then it maybe becomes a lot more attractive to try this new style of programming that is like impenetrable.
+
+**Jerod Santo:** And that goes back to your churn mention as well. Like, if you're only going to work somewhere for 12 to 18 months, unlikely that you're even going to deal with the mess that you're creating, because you moved on to the next organization by the time the mess rears its ugly head. And so you actually may be thinking that you didn't make a mess, because you're gone.
+
+**Bert Hubert:** Yeah, you never had the opportunity to find out.
+
+**Adam Stacoviak:** You may even gotten a bonus because you delivered in such a time frame.
+
+**Jerod Santo:** Very complicated, yeah. It's so hard.
+
+**Adam Stacoviak:** The solution's out there, solving the problem, but meanwhile there's something percolating behind the scenes.
+
+**Jerod Santo:** So ego kind of gets in the way. Like, we want to be clever.
+
+**Bert Hubert:** Yeah. And sometimes I write code that is so embarrassingly simple, and I'm like "It cannot possibly be enough." And then like five years later, you go "Well, apparently it was simple enough. Apparently, it was good enough."
+
+And there's also this anticipatory thing. For my parliamentary monitoring system, I have a very interesting situation that in parliament are like seven or eight different kind of important entities. There are meetings, there are documents, there are appointments... And seven or eight - the tricky thing about that is can you write a useful abstraction, that is useful even when you only have seven sort of different things?
+
+\[40:07\] And you could write this lovely generic solution where you could have thousands of different kinds of entities in there, but it would be quite complicated. But because there are only seven or eight entities, it is not always necessary to build this sort of generic solution, but it does feel a little bit bad if you have a switch statement with eight kinds of categories in there. It feels a little bit like "No, I should have built an object, and inherit from that, and made it all generic..." And that's also the one thing I see that people are sort of anticipating huge complexity later on, but that might never happen. You might never get to that point where you needed a generic customer service class handler factory.
+
+**Jerod Santo:** Right. YAGNI just rears its ugly head. So I've had this - have you had this before Bert? ...where it's like, I want to create a simple solution, I go ahead and do that, I put it out in the world, and it's too simple, because the world's actually more complicated than my software will handle. And then, as it kind of grows in order to handle these different edge or corner cases that I didn't consider when I was building simply, it ends up all hairy at the end, because it's like "Well, it has to deal with 17 different things that might happen." And it's like "Well, was there ever a simple thing that would have solved this?" Because the real world is complicated sometimes. And then I throw up my hands and think "I don't know... How am I actually going to do this...?"
+
+**Bert Hubert:** Yeah. So there's one sort of telling story... So I also wrote PowerDNS, which is a name server... And we started PowerDNS in 1999. And DNS evolved over time, and it added cryptography to it DNSSEC. And initially, PowerDNS started out really simple, and then the world turned out to be complicated, and we added more and more complexity to support this cryptography in there. And it was very difficult, because DNS itself, for example, is highly case-insensitive. Case really doesn't matter, until you start doing hashing and encryption, because that's sort of -- yeah, it cares about case. And usually, you sort of end up then -- as you described, you end up with a situation that is just a mess.
+
+And it's very tricky as an organization to allow yourself to say "No, we're going to clean up the mess." So we're not going to do this big, hairy -- we're not going to do version two syndrome, where we say "No, we're going to make the best version ever, that will be so generic that it deals with everything", because it doesn't work. But what does work is actually sitting still and saying "Look, we're going from version two of the software to version three, and we're going to give ourselves a whole year to just do spring cleaning, house cleaning. And we're going to make -- the things that were too complicated, we're going to make them simple again..."
+
+Back when we used -- for example, in PowerDNS we used simple strings to deal with domain names. It turns out domain names only look like strings. They actually have more complicated semantics than that. So within PowerDNS, the big thing we did, we said "Look, we're going to rip out all the strings, and we're going to move to a dedicated DNS name class, that's going to deal with all these weird DNS things." And the only way this could happen is by allowing ourselves a year to do a cleanup. I can very much recommend doing that.
+
+**Jerod Santo:** Did this turn into a business? Was this making money at the time?
+
+**Bert Hubert:** Yes, it turned into a good business... Actually, PowerDNS is a rare example of a fully open source piece of software that actually turned into a commercial success, while still remaining mostly open source.
+
+**Jerod Santo:** So were there pressures on the money side to like not do that?
+
+**Bert Hubert:** Yeah, there were... Yeah.
+
+**Adam Stacoviak:** \[43:59\] Of course...
+
+**Bert Hubert:** But you have to sort of have a little holistic view on that, because it turns out that we could have focused on where we were making all the money. But it turns out that the reason we got paying customers was that the technical community liked what we were doing. They would recommend us everywhere. So if we had focused on purely the financial bits, we would probably have had great business for a few more years, but eventually we would not have been able to do proper DNSSEC. No one in business cares about DNSSEC. They really do not care about that. However, the nerds, the technical people that work there, they really do care about that. And we also cared about that.
+
+So it was, in the end, a solid business decision for us, to say "We're going to give ourselves a year to clean this stuff up." And I dare say that if we had not done that, the company would have gone bust later on, because it would have collapsed under its own complexity. But I do not find a lot of other examples, where people gave themselves like a year, put the old software in maintenance mode, and were going to just do cleanup for a solid year, and then come back better. That is actually a very rare thing.
+
+**Adam Stacoviak:** That's rare.
+
+**Jerod Santo:** The only one I could think of is Snow Leopard, which was Apple's version of macOS that was touted as having no new features, even though there were a few features in there... But it was an entire release cycle, all focused on stability, refactoring, performance.
+
+**Bert Hubert:** Yeah, wow. Great.
+
+**Jerod Santo:** And it is, to this day, many people's favorite version of macOS, going back -- everyone's like "Yeah, Snow Leopard was awesome."
+
+**Bert Hubert:** Yeah. Well, I think as a concept it's doing a spring cleaning. We do it for other things. I mean, in equipment you do this kind of maintenance, where you replace the load-bearing parts, and take it out of service for a bit. So yeah... But it's not very popular. I mean, mostly you see people just having version two, second system syndrome, where they say "Look, now the next version is going to be perfect. I'm going to redo everything from scratch." And then they forgot what made the old version great. That's also a thing. But just doing incremental cleanup... Yeah, I very much recommend it.
+
+**Jerod Santo:** It reminds me of something mechanics say, that I learned recently. They say "If you don't schedule your maintenance, then your car will schedule it for you."
+
+**Bert Hubert:** Yeah, exactly. That's actually also part of the blog post I wrote on software maintenance.
+
+**Jerod Santo:** Oh, is that where I read it? Okay.
+
+**Adam Stacoviak:** \[laughs\]
+
+**Bert Hubert:** That's probably where you read it.
+
+**Jerod Santo:** I love it.
+
+**Bert Hubert:** The thing, I put it in there -- I got it from an Australian guy, and he said "If you have dependencies in your software, schedule some time, maybe every quarter, just to take a look at all your dependencies. How are they doing? Are they still being maintained? Are the security releases coming out? Has the new VC started the monetization cycle yet?" And the idea there was that if you do this on your time schedule, you'll find out ahead of time that things are not going well. And the other way is that you figure it out because the stuff doesn't compile anymore.
+
+**Adam Stacoviak:** Yeah.
+
+**Jerod Santo:** Right.
+
+**Adam Stacoviak:** When I was in the military we had this really awesome term that I still live by. It is PMCS. It stands for Preventative Maintenance Checks and Services. And legitimately, every time I got into a vehicle, I could not just get into it and just drive. I had to do this series of checklists every single time... Because if you don't, eventually you get bitten by this non-Preventative Maintenance Checks and Services. I had this thing called a Humvee. It was a very big vehicle. Yeah, I was in the \[unintelligible 00:47:34.16\] around a lot of diesel fuel in the military called a JP8. But it was diesel, basically. It was 2,500 gallons of very expensive fuel that I had to transport safely, securely... And if my vehicle wasn't capable of getting to and from, or there was something wrong with the tank, or just anything, catastrophe could have been right around the corner.
+
+\[47:59\] So we had this PMCS mindset just ingrained in us, from emptying out connexes to just confirm what was in there last week is still in there. Even though everyone knew it never got opened and it never got used, we had to pull it all out, confirm it's there, and put it back. It's just like, the boring stuff - you have to do that mundane checks, boring stuff to make sure that what's on the rails stays on the rails.
+
+**Bert Hubert:** But interestingly -- it's a very interesting, good concept. I sometimes get that reaction when I tell people about unit tests, and regression tests. They say, "Well, look, the software worked the last time I used it, so it's still going to be working." And I'm "Well, just write some unit tests and see what happens", which is basically the sort of same thing, where you just ask the computer to add two and two, and you check if it's still four. But still, so often unit tests will find that you broke stuff, even though you thought it impossible that something got broken. So I do enjoy that mindset where I say "Look, we're still going to test it every time, and do that", but you never know what happens behind your back. But there are people that are like "I'm not going to write unit tests, because my software worked the last time."
+
+**Jerod Santo:** Right. That test suite becomes such an asset down the road. I was just updating our dependencies today, and just having test coverage where I could say "Okay, here's a dependency, run the tests, everything passed, updated to latest, run the test again, everything passed..." It just has a peace of mind that you're like "Okay, updating that dependency did not break everything."
+
+**Bert Hubert:** No, and it's magical. PowerDNS was, of course, very old, and in 2001 we made the first regression tests and unit tests. And that was like an amazing feeling, that your tests discovered that you made a mistake and they told you before you shipped. I mean, it's still a good feeling. And I sometimes give trainings to scientists on programming, and when I show them the power of unit tests, they go "Wow, that's just awesome, that you get an email message that you made a mistake, instead of getting a complaint later on."
+
+**Jerod Santo:** Yeah, it's like, if you don't run your unit tests, then your customers will for you, to reapply the statement.
+
+**Bert Hubert:** That's the same kind of concept, yeah.
+
+**Break**: \[50:15\]
+
+**Jerod Santo:** Here's a question for you, Bert. How do you know how many dependencies is it too many dependencies? Is there a heuristic? Where do I know that I've just jumped the shark?
+
+**Bert Hubert:** I struggle with this myself... Because I do a lot of reinventing the wheel often, because I know that I could spend the time learning about your dependency, and then I would still be disappointed, because two years from now the dependency would be different again. So I do often spend more time than is wise probably re-implementing some simple things.
+
+I draw a very hard line at cryptography, for example. I mean, unless you are a famous cryptographer, you should not be doing cryptography. You should leave that to someone else. But even then, if you look at the state of many cryptographic libraries, OpenSSL has been a sort of disaster zone for years now. But I draw a very strong line there. You do not build your own cryptography.
+
+But on the other hand, if someone asks me to be responsible for a piece of software - and it's an important piece of software; it's not a birthday calendar, it's something that people rely on. I need to know what is in there. I need to know what I'm shipping. And if I run a modern piece of software development, I end up with like 300 dependencies, which are also included dynamically.
+
+At one point, I was in a situation where we said "We have to audit this software." They're like "Okay." And then it turned out that if you would build the software three weeks apart, and you didn't change anything, it still would have changed, because by that time the dependencies had moved around, and shifted their own dependencies. So I'm actually sort of hardcore in the sense that I don't think you can ship software with like 300 dynamic dependencies. And people are still doing it, of course... But I do wonder what they're doing. And people have said, "Look, you C++ people, or C people, you have one huge dependency, which is called your library. And you rely on your operating system for a lot, and that's also like millions of lines of code." But there you have a little bit of a safety blanket, that everyone relies on the C library.
+
+And so there are a lot of eyes on it. And it also doesn't change that much. And we all rely on the operating system, and lots of people are looking at that.
+
+So I'm sort of maybe on the low end of rebuilding my own wheels too often... But on the other hand, I still cannot understand shipping software where you say "Look, I honestly do not know what is in there, because it gets determined at build time what is in there." And maybe you could audit 300 dependencies if you would have enough... But it requires a lot of work to actually sort of even figure out who is writing all these dependencies, who owns them.
+
+So yeah, I struggle with this, because when I tell people that I simply cannot believe that you run npm install get - I counted it, 5 million lines of code come out when you do that, for the most basic product. How can you ever know what is in those 5 million lines? But apparently -- I'm old, apparently, and it is now considered very normal. And I know for a fact that my car has npm in there. And I find that worrying.
+
+**Adam Stacoviak:** \[56:11\] How do you know?
+
+**Bert Hubert:** Because it tells you so. There is this screen that says the copyright and the licenses and stuff, and it says, "Look, React is in there, and npm is in there..." And I hope it's not in the driving parts of the car. I hope this is just the radio, or something like that.
+
+**Adam Stacoviak:** The entertainment sections of it, yeah.
+
+**Bert Hubert:** I hope that is the case. I hope that is the case.
+
+**Jerod Santo:** Yeah, the tough thing about dependencies, even when you can audit, if you can and have audited and checked out the owners and the code, and the whole do your due diligence, is that what we're learning over the course of years and decades is even if you do that, if you're loading your dependencies from the network, you can't trust the network. And what you think you can trust today, you can't actually trust a year from now, because the network changes. And so even if you're doing some due diligence, you can still get bit. These so-called supply chain attacks are happening more and more often, where all of a sudden what you thought was your dependency is replaced with code that is not the same code. And that's incredibly troubling for me.
+
+**Bert Hubert:** Yeah, it is. And I come from a world where people also have to develop on networks with no internet connection. And it's just basically impossible for many people to function there anymore, because you try to do something and - yeah, there's no internet. There is no network. And the people that are in national security software, or encryption software, or that kind of stuff, intelligence agencies, they are actually in trouble, because the art of developing without a fully functioning network connection is dying out, because there are lots of people that need a runtime connected to the internet to get anything going. And even then, your software might be relying on third-party services. CDNs somewhere. And if you're running this nuclear power plant where people say "Look, we're not going to connect you to the internet", it is becoming a vanishingly rare skill, of being able to build software that runs without a network connection.
+
+**Adam Stacoviak:** One thing I appreciate about this part of your post on dependencies was just the ways that you enumerated over what might happen over time with dependencies. I'm going to read ,them if you don't mind. You said just basically scrutinize and limit your dependencies, because all of them might over time - and here's a list: drift away, leading to adjustments in your code, or worse, silent changes within behavior... Jerod, you kind of mentioned this. Shift to a new major version, with cemented changes, requires rewrites on your part... It might get abandoned, or simply disappear, or start to decay. It might get hijacked by a nation state. We've seen this happen in recent past by nation state actors. Think npm, PyPy etc. It might start to get monetized by the new VC owner, who's I guess acquired the open source or the dependency in some way, shape, or form... And it might develop conflicting dependency requirements of their own.
+
+I think just enumerating those different things -- because you think, "Okay, obviously, the simple answer is just simple software. Limit your dependencies." But why? And that's the why of those, I think -- and it's not all the whys, but it's several versions of the why that might help you understand why you shouldn't.
+
+**Bert Hubert:** Yeah. And these things all happen. So I've worked with an impossible Python project that just requires two different versions of the same dependency, and they cannot do it. And you also don't know how to recover from that. But on the other hand, it is of course sometimes great fun that someone just moves ahead so quickly, because they embed all of Chrome in their project. So yeah, boom.
+
+\[01:00:02.05\] And I'm sort of this ascetic person that needs to be really -- yeah, I need to work really hard to display something that looks like a webpage, and someone else has no qualms at all about embedding an entire Electron copy. So it is always a bit of tension that says "Look, I'm telling you about these things that will be bad like five years from now", and these people are like "Hey, but I'm shipping today." So I do really feel that stuff. But especially we've -- lately we had InfluxDB, I think, which all of a sudden... Which is a time series database, and I think that they have now limited you to 500 metrics or so, because they want to make money. And of course, I respect that they want to make money, but if you built your whole project on InfluxDB - yeah, you have an issue right now, because your business model might not support their business model.
+
+**Adam Stacoviak:** Is this a recent change that happened, or is it something that's --
+
+**Bert Hubert:** Yeah, it is. I think it's \[unintelligible 01:00:58.10\] from the past two weeks, or something like that.
+
+**Adam Stacoviak:** Is that right?
+
+**Bert Hubert:** Which is why I tend to do almost everything with SQLite, because we know SQLite is going to be around forever. And we know the people that run it, and they're not going to do anything like that. They don't want to do anything like that... I mean, a hundred years from now people will be able to read SQLite files. It's that abundant right now. So I try to stick with these things, that way you can be sure that 25 years from now they will have this software around. So it's not gonna bite me.
+
+**Adam Stacoviak:** It kind of leans into some of the things you mentioned too down the line, which is more of a decision tree of sorts, which is how to scrutinize these dependencies. What does \[unintelligible 01:01:48.25\] Who else uses it? Who writes it? What are their goals? I'm kind of wondering if there's something out there that -- a piece of software we can sort of throw a dependency at it and it spits back some versions of all these answers, you know?
+
+**Bert Hubert:** Actually, that would be good. Of course, we have these sort of SBOMs right now, the software bill of materials, but that is only like -- yeah, it's this piece of software, and it's that version... But it would indeed be very interesting to say "Look, this software is actually funded by this foundation, or something like that", and it's well funded, and it'll be around forever... Or you find that "Yeah, this is actually one guy in Nebraska, the famous one guy in Nebraska..." That would be very good.
+
+**Adam Stacoviak:** Yeah.
+
+**Bert Hubert:** But I do want to share one other technique I use to figure out how my dependencies are doing, is I just use a dependency for a bit, in an isolated fashion... And I will always run into some kind of issue. There's always something. And then engage with the project. See what happens. So you go to the project and say "Hey, I just put --" Take SQLite, for example. SQLite had a new feature for database replication, and it's really cool. I really recommend it. It was very new. And I was wondering, can I build on this? Is this solid enough? And I tried it, and there was an issue, and I filed an issue with the SQLite project, and they fixed it in 25 minutes. And they immediately believed me that there was an issue, and they were on it, and it was just done. And when you do this with a dependency, you will often find out that no one responds to your worries because they're not paying attention.
+
+So actually, trying something and then just opening a ticket and figuring out how do they respond? This is incredibly telling. But again, this requires work. This will require you a few hours, at least, per dependency to figure that out. And then people say "Look, I have a thousand dependencies. I won't have time to figure that out." And my response would then be "Then you should not be shipping a thousand dependencies."
+
+**Jerod Santo:** If you select a dependency wisely, it can save you hundreds of hours of effort. So maybe one or two hours spent putting the work in is worthwhile, right?
+
+**Bert Hubert:** \[01:04:09.08\] I think it is. And there are things that -- I mean, like I said with cryptography... I mean, the cryptographers out there, they have done a better job than you would ever do. Or otherwise you would be a cryptographer. But I think it's somewhat of a lost art... I mean, I tell people -- you probably remember the npm Leftpad thing, which was where you could left-align a string...
+
+**Jerod Santo:** Certainly.
+
+**Bert Hubert:** ...and someone removed that, or made it malicious, and then the whole tower of dependencies came falling down. And when I tell that these days to people, I say "Look, I worry about Leftpad." Then many people go like "What is Leftpad?" Although that should be part of our sort of engineering lore that says "Look, do not go around depending on single-line files that could be taken over by any random internet person someday."
+
+**Jerod Santo:** Are these youngsters you're talking to?
+
+**Bert Hubert:** Yeah, well, relatively speaking. Relatively.
+
+**Jerod Santo:** Sure.
+
+**Bert Hubert:** I mean, there are many people that are like 30, and I'm starting to find people at 30 to be a bit on the young side... But that's just me.
+
+**Jerod Santo:** There you go. It's all relative.
+
+**Bert Hubert:** It's all relative. But yeah, this is something -- I mean, this is just the way modern software gets developed. And of course, we have the die-hards that do embedded software, or really the safety of life stuff... But we are in a small minority, and the minority is getting smaller.
+
+**Jerod Santo:** There's also new tech coming out all the time... And you've been around long enough, 30 plus years in the business to know that not all that new tech is going to be additive, or useful, or last.
+
+**Bert Hubert:** Yeah, so I did a talk on this for the Dutch networking community NLNOG, which is great... And the thing I tried to get across there was you have these new things coming along all the time. The new and shiny. And there are the people that are just always attracted to the new and shiny, and they have a glorious time discovering all kinds of new things, and fun... And there are, of course, people that are still, as I call it, rocking it with awk. Because you get stuck with "Yeah, this worked for me in 1997, and it will work for me in 2029 as well."
+
+**Jerod Santo:** Right.
+
+**Bert Hubert:** You have to strike this balance to say "I enjoy the new and shiny, but not too much." And you should also let go at some point of Perl, or awk, or whatever.
+
+**Jerod Santo:** Specifically Perl...
+
+**Bert Hubert:** Yeah, it's high time. It's high time.
+
+**Jerod Santo:** It's high time, yeah.
+
+**Bert Hubert:** And for myself, I also sort of -- I'm a bit ambivalent about this, because on one hand I love the new things, and I think everyone should be... So I do a lot of JavaScript, for example, but I've also been doing C++ since like 30 years... And you could on the one hand say that C++ is starting to get old, and maybe you should move on to something new... But there I decide "No, look, I'm like really good with C++, and I'm getting this stuff done. And I know there are issues with it, but hey, it works for me." But on the other hand, I have no problems with doing all kinds of newfangled JavaScript things. And that was sort of the upshot of my talk, where I said "Look, enjoy the new shiny, and do study it, do look into it, but don't just shift whole frameworks every few years because they are just the bee's knees right now... But also do not get attached to your extremely old frameworks." And I think that that's actually -- that's why I wrote this post and did the presentation, I think this is one of the most biggest determinants of your technical career.
+
+If you would take a look at the technical development over like 20 or 30 years, how you look at new things really determines if you're going to be in technology for the long haul, and continue to be relevant and effective, or if you have to bow out and become some kind of manager, because you're no longer with it. And managing that is, I think, a very important part of a programmer's career also.
+
+**Jerod Santo:** \[01:08:17.11\] Yeah, we're certainly in a career where you can't just learn everything you need to know at university and be done. I mean, if you're not interested in change and adaptation and learning new technologies, this isn't really the space for you. You're not going to last.
+
+**Bert Hubert:** No. But it turns out there are many employers that actually also do not believe in change...
+
+**Jerod Santo:** Sure.
+
+**Bert Hubert:** So you can end up in some kind of insurance place, or a government tax agency and work with 25-year-old technology.
+
+**Jerod Santo:** Writing COBOL on a mainframe...
+
+**Bert Hubert:** Yeah, they're still rocking it over there. So it's not necessarily -- you could come at a point in your life where you say, "Look, I don't need all these new things anymore, and I have only so many years left in my career, and I'm going to spend them in a basement, in some kind of insurance place." So it's not all bad, but it is something that I think you would have to just ponder for yourself, and say "Am I going to invest also?" Because there comes a point in your life - and I think we've touched on this in this presentation also, where I'm just saying "Hey, these 2000 dependencies, they're all bad." And there is a risk that at some point in your career you go "Look, things used to be better, and all this newfangled stuff is just wrong." And when you hit that point, you have to have a real good think, and think about "Am I leaving the industry behind? Am I just old now? I'm just sort of a sad person that sits there and say "Well, in my time..." And you should at least -- when that happens, it should be a conscious decision, where you say "Okay, I love this old stuff more, and I'm not going to pretend that my old stuff is better. I'm just going to admit that I'm really good with this old stuff, and I enjoy doing this old stuff. And I'm not going to tell you that Rust is wrong, or that Swift is wrong, or that whatever new thing is wrong. I'm just going to enjoy my old stuff." But it's really sad to see some people see new things happening, and they just get stuck in telling you how great their old stuff is.
+
+**Jerod Santo:** So how many languages have you invested in, seriously?
+
+**Bert Hubert:** Yeah, so this is a terrible story... So I'm terrible at Python. So I write a lot of Python, and mostly for making graphs in Matplotlib and Jupyter... But my Python is really terrible, and it has been terrible for the past 20 years, because I've never written anything. I only write scripts in Python, and I've never invested in really learning that stuff. So in Python I'm really the copy-paste engineer. And then JavaScript, I actually sort of made a weird decision. At one point, PowerDNS was really not going well, my first company... And then I decided I wanted to invest some time in this new web stuff. And every serious programmer hated JavaScript with a passion. And I was like "Well, that probably is the future."
+
+**Jerod Santo:** That was a good call.
+
+**Bert Hubert:** That was a good call. I probably just got lucky at that time. So JavaScript, I'm sort of reasonably proficient in. And in C++, that's a weird thing... That's my workhorse. And the thing is, I use like 20% of C++. Because over the years, I've learned there is a good core in C++, that's like a really solid language, and it's good stuff... And then there is a lot more stuff. And I'm sort of on a sort of diet version of C++, where I'm like "Look, I know the good parts now. I'm not leaving the good parts. I'm just staying there." So that's the other thing.
+
+\[01:11:46.06\] And lately, I've had to admit that I'm sort of skilled at HTML now, which I never saw coming... Apparently, I got a little bit of that going. But that's really it. I got my start with Perl, which was also sort of -- I've found that people sort of didn't take Perl seriously, and I also did not take Perl seriously. And at one point I had to do a lot of data crunching; like millions and millions of lines... And I thought these Perl weenies - I can do this better in C. And then I found out that these Perl weenies, they do know a lot of stuff about parsing files. And so that was a humbling experience. But these days, I sort of have a trifecta going of C++ and JavaScript, and mostly SQLite in the backend, because I've invested a lot of time in understanding its quirks. And it's tough. And it's actually a really sort of a solid data handling framework, to the point that I now sort of trust SQLite files more than text files. Because it's quite easy to mess up a text file. One stray character return and then it's done.
+
+**Jerod Santo:** Sure.
+
+**Bert Hubert:** SQLite will not disappoint you on that front.
+
+**Jerod Santo:** So how do you see promising technology coming?
+
+**Bert Hubert:** I've actually been sort of not that happy with the tech. For example, programming languages. If you look at the development of programming languages, you see that in the past years, we've not gotten closer to the ideal programming language. So there is something wrong with every programming language right now. Some of them, they're like really safe, and some of them are really fast, and some of them are really simple, until you try to do complicated things. And I'm still hoping that someone, someday will square that circle and have the programming language that is sort of simple, and fast, and safe, and maybe fun... But we're not really getting there. And the other thing I, of course, worry about - we have all these frameworks for website development, and these are also not improving. They're just changing.
+
+And one thing that I have a serious -- yeah, one thing I would love to build... If you look at any kind of software, it involves interactions. So you have to do things with the software, and you might have to sign up for a website, and then you get assets that are yours, and you can configure these assets... And that is a whole workflow kind of system. And I would love for us to move into developing workflows, instead of developing webpages. Because if you look at most interactive websites, they're still like "Well, we've built this page for you", and that page is full of conditions. It says, "Well, you have to fill in this, you have to fill in that, and then you can press Send, and then it happens." But I would love for us to move to a higher level description, where you say, "Look, this is a website where you can sign up for college", or whatever, and this is in a domain-specific programming language, where we say "Look, you can only move from this state to that state once you have filled out the following things." And if we ever developed that, we would get so much more robust sign-up flows. Because the system is just aware of what can happen, and what cannot happen...
+
+And I would love one day for someone to develop this higher-level way of interaction developing... And I'm sure that it has been done in some places, but I would love to have this as infrastructure, so that no one has to write a transition handler anymore that decides if you filled in the right zip code, or whatever. I just want a library that already knows what the zip code is, as a primitive. But I don't see any of that happening. I see new iterations of ways of creating HTML pages. And every five years, you get a new sort of wave of these new frameworks, but I don't see any of the high-level stuff happening.
+
+**Jerod Santo:** \[01:15:56.15\] I do see it happening at the super-high level, almost no code... If you look at like form builders, which we give to people, and they can put in their conditionals, and they can say "Give me a zip code widget, and put it on the screen..." But you know that that's like... Not ad hoc. What's it called? The one that's just like specific use for that one particular...
+
+**Adam Stacoviak:** Bespoke.
+
+**Bert Hubert:** Bespoke, yeah.
+
+**Bert Hubert:** Yeah, bespoke. Just the one thing. So I also do presentations, and you can hire me as a speaker, and my speaker agency wrote a whole low-code solution system just to do the logistics of that. And it's sort of wonderful in a way, because I'm really impressed that they managed to build that mostly themselves... But on the other hand, every time I do something with the system, it takes like three seconds to go to the next page. And I'm wondering, "What is it doing in the background?" But it's sort of -- it is impressive, because it is built on this concept where you click this stuff together without having to be a real programmer. But the thing is, I would love to be a real programmer, and have the primitives that make me do this safely, and at a higher level.
+
+**Adam Stacoviak:** Yeah, it's hard to resist the future, really. Like, you've said, you've got to pay attention to it, but not so much that you're like engulfed by it. You have to weigh it against what you know, where you've been, and where it may go... And hopefully, the wisdom you've gained gives you the ability to discern that future and how it applies to your present. Otherwise, you get left.
+
+**Bert Hubert:** Yeah. So in this talk on dealing with change, I also mentioned AI. And then a part of the audience goes like "Yeah, AI! I want to do everything with AI." And part of the audience goes "No, it's unethical, it's bad, it's not as good, it's not reliable..." And of course, it's tricky to find the middle ground in there. The AI is just astoundingly good at transcribing spoken text. It is actually superhuman. So there's this Whisper system from OpenAI... It does 99 languages, and it even does languages where they didn't know that it could do it. And it is superhuman, because it's better than any human being on the planet. So there is actually really good AI.
+
+And then, of course, there are things where we still have to worry, "Yeah, what good is a model that just hallucinates stuff?" and we could find that really bad. But it is extremely tricky for people right now to understand that "Look, we could do some wonderful things with AI." It really could be good, even though there are some stupendously bad things you could do with it. But I would love for AI to also come into our universe as serious programmers to say "Hey, we have this tool for you, and it can recognize --" If a user enters something, the AI could say "Well, this is probably not what they meant." Something like that, that looks over your shoulder. Or something that deals with -- you have a system that has to deal with user complaints, and you have an AI system that automatically does the triage and says "Look, this user has a real problem, and this looks like only a feature request." I would love to have those kinds of primitives. But there are people out there that are just so in love with AI that they're like "We're not going to need programmers anymore", and there are programmers that hate AI so much that they say "We're not ever going to touch it." And Adam, as you said, navigating that... We cannot hold back the future, but it's tricky to sort of navigate the stuff that is good, and that we should have more of, while staying away from the areas where we say "We're going to just replace the whole thing by an AI chatbot."
+
+**Jerod Santo:** Get out of here, AI chatbots... We've had enough of you.
+
+**Adam Stacoviak:** Yeah, well... I think that this decision tree that you've got listed - which is rudimentary; it's not fully featured in terms of every possible way you would decide against choosing a dependency in this post... I think that's a great example of where you'd throw AI at it, or automation, let's just say. The basis of automation, and AI where necessary to sort of investigate a dependency, and whether or not it checks against your balance.
+
+**Bert Hubert:** Yeah.
+
+**Adam Stacoviak:** \[01:20:02.10\] You know, does a tech look good? I mean, maybe that would be a great example of throwing AI at it. It can do an analysis of a codebase and say "Well, what does good look like?" And then the model obviously depicts some of that stuff. Who writes it? That's a pretty easy one to just automate through databasing, but... Yeah, there's a lot of AI you can do here with this decision tree to make choosing dependencies slightly faster and easier.
+
+**Bert Hubert:** Yeah, I have to admit, I had not thought of that yet... But people did mention it. They said, "Look, I have 1,600 dependencies, and I'm going to make the AI look at them."
+
+**Adam Stacoviak:** Yeah.
+
+**Bert Hubert:** It's not necessarily the worst thing to do, because the AI might in fact very well just figure out that "Look, this is just wrong", or "This code is already in North Korea."
+
+**Jerod Santo:** \[laughs\]
+
+**Bert Hubert:** So yeah, it's actually -- as a layer, it is probably an interesting thing to do.
+
+**Adam Stacoviak:** Well, I mean, if you've got that many dependencies, it's literally impossible to investigate your own dependency tree.
+
+**Jerod Santo:** It's better than nothing. It's one of those "What have you got to lose, besides maybe a few cents?" Well, Bert, what else? Anything else you have on your mind that we haven't talked about yet?
+
+**Bert Hubert:** No, no, I think we touched on the -- yeah, we've spoken a lot about dependencies, of course...
+
+**Jerod Santo:** Of course.
+
+**Bert Hubert:** And yeah, the thing I keep harping on about, about how do you select the new things that you, that you don't like? That's something that I would really recommend people to spend a bit more time on, because it's really strange... If you look at how we choose what technology to use, it's actually quite often a sort of random thing. It's like "Hey, I've found this stuff on Hacker News, and I liked it, and I tried it for a bit... I liked it as well." And then accidentally, we pick the future direction of our career, or over a software project. So I would really recommend people sort of make more of a conscious decision on these kinds of things. But we covered that already.
+
+**Jerod Santo:** Yeah, yeah, yeah. I think that's a really good point. I think we do tend to overvalue social proof - you know, what everybody else is doing - and undervalue our own ability to dig in and make a decision based on what we find.
+
+**Bert Hubert:** Yeah, one of the things I stressed in my talk also on selecting these technologies - you can use your coworkers or the people that work with you on the project, and just tell people "Look, I'm gonna present on this new thing. I've found this new magical database, and it does 300 million lookups per second. And these are my experiences, these are the graphs that I've made..." And then your own co-workers could be the filter, because they might not be impressed by your story, for example.
+
+**Jerod Santo:** Right.
+
+**Bert Hubert:** \[01:22:46.23\] Or you might find, because you actually have to get down with it, to actually make the graphs to present to your coworkers, that you find that the graphs are actually not that great. So in this way, you can both entertain the people around you with your findings, and also use them as a way to keep yourself honest. And if you don't have any coworkers around, you can write a blog, or a vlog, or go on a podcast or whatever... But whatever forces you to take a serious look at something and work on "How would I present that to someone else?"
+
+**Adam Stacoviak:** If you do any of those things, I would recommend going to Changelog.com/community. That's a good place to hang. Share your findings there, share your thoughts there... There's a thriving Zulip there, waiting for you to thread, and be a part of... And it's free, too.
+
+**Bert Hubert:** Yeah.
+
+**Jerod Santo:** Bert, what's the best place for folks to connect with you on the internet?
+
+**Bert Hubert:** Yeah, so that's an interesting one. So I've been pretty enjoying my time on Mastodon. I think it's really a geek heaven. So that works really well. But interestingly, people ask me that question quite a lot, and I've found, and I think it's a very weird, strange sort of regression... Many people are now running mailing lists again, and sending out updates, and newsletters, and stuff.
+
+And I've found it to be a bit pompous, to write mail updates on what Bert has been doing... But apparently, there is demand for that. So I am probably gonna restart a small newsletter.
+
+**Jerod Santo:** Nice.
+
+**Bert Hubert:** That kind of stuff. I also used to tell people to just follow me on LinkedIn, but the problem there is that, on one hand, I'm sort of active in Dutch politics, things which are entirely not interesting to the rest of the world...
+
+**Jerod Santo:** To a lot of people, sure.
+
+**Bert Hubert:** Yeah. And the other way around, there are many people there that follow me for the Dutch politics stuff, and they are sort of confused, they are massively confused to suddenly see me talk about --
+
+**Jerod Santo:** \[laughs\] "What's all this nerdy stuff?"
+
+**Bert Hubert:** C++... And then they go "What is this?" So actually, for the newsletter thing I'm building now, it's actually gonna have like four newsletters. It's gonna be one for people that like Dutch news, that like news in English, that like stuff about politics, and that like stuff about technology. And you can pick your own quadrant in that, to see which updates you want to receive. But for now, if you go to my website - and I'm sure it will be in the show notes, or whatever; or you can just google search Bert Hubert. And then I have a website, which is RSS feeds, and stuff.
+
+**Jerod Santo:** Nice. Gotta love RSS. Here's a name you can name your new newsletter... You can call it Bert's Pompous Update. \[unintelligible 01:25:30.16\]
+
+**Bert Hubert:** Yeah. "Bert's self-important things that you actually all asked for." Yeah, just tongue-in-cheek that sucker. It always works. Cool. Love it. We appreciate you coming on the show. Nice meeting you. Adam, anything else you wanted to talk about or ask?
+
+**Adam Stacoviak:** Keep that software simple.
+
+**Bert Hubert:** Yeah, that's the goal. That's the goal.

Building for application developers (Interview)_transcript.txt

@@ -0,0 +1,673 @@
+**Adam Stacoviak:** Well, Anurag, I'm glad to have you here... It's been, I would say, such a journey for this conversation, because I've known of you and I've known you for many years, and just have not had you on this show... Now here you are, $155 million deep raised, $80 million recently announced last month... You must be excited, so welcome, finally, to The Changelog.
+
+**Anurag Goel:** Thank you. Yes, it has been a journey, but I'm really happy that we're finally having this conversation.
+
+**Adam Stacoviak:** You know, I'm such a fan of the boldness it must require, and the courage it must require to take on Goliaths. And not just one, but many. Render is in the shadow, and maybe now above the shadow of the days of Heroku... GCP is obviously still there... We'll talk about your infrastructure. I think you're built on the cloud, so you compete with the clouds you're actually built on too, I imagine, but at a different level. I just admire the courage and required tenacity to do what you've done. And you've obviously done it well.
+
+**Anurag Goel:** Thank you. I really appreciate that. It is not how I think about it.
+
+**Adam Stacoviak:** No? How do you think about it?
+
+**Anurag Goel:** Well, I think this is a large and important problem. And maybe I can give you some background on how I ended up doing what I do now. So before Render - I won't go too far back, but I joined Stripe as the fifth engineer in 2011, and I left in 2016. And the company did well, and so I had this opportunity to really think hard about what I wanted to do for the rest of my career. I could choose to do nothing and sort of chill out... But my wife and I both, we talked about it, and I think we both saw it and see it as a responsibility... Very few people are granted this opportunity to really have the freedom to pursue any goal that they want, and so my goal then became to find a really big problem that I'm also very excited to solve, and that I am good at solving. And as part of that, I ended up building applications in a lot of different domains.
+
+I mean, even at Stripe, I had seen our infrastructure team that was managing AWS grow from when I was there one person out of five was fully focused on AWS... That number, that percentage, about 20%, stayed about the same as Stripe continued to grow, at least when I was there, 2011 through 2016... And all these people were simply managing VMs on AWS. They weren't contributing to the product. They weren't helping the other teams necessarily move faster. So that was a big waste. And then when I built my own applications post-Stripe to try to figure out which domain I was going to be interested in over the long term, it hit home firsthand. It was very obvious to me that getting something up and running in the cloud was still very broken. Ultimately, you end up building out a Kubernetes cluster, and no one wants to do that. I'm sure there are some people in the world who love building out Kubernetes clusters, and bringing all the care and feeding what those clusters require over time, and spending all the money that Kubernetes ends up sucking out of your ops budget... But ultimately, what people want, what I wanted, was I want my application to -- first of all, I want my cloud to allow me to spin up an application really quickly, and get it up and running, and make it reliable. But then I want it to scale, I want a lot of security by default around it, I want all the features that I will need as the application grows...
+
+\[08:22\] And so the dichotomy between Heroku and AWS when I started Render was Heroku was easy to get started on, but it stopped scaling beyond a certain point. And if you needed more than 14 gigs of RAM, you couldn't have it. You just couldn't. Your applications were restarted every -- and a lot of this is still true. Your applications were restarted every 24 hours... And then the price became unsustainable once you went beyond a certain scale.
+
+And then AWS was AWS. Just even logging in and looking at the dropdown of the 300 plus services... They have a search in their dropdown. Who has a search in a dropdown? AWS does.
+
+**Adam Stacoviak:** AWS does. Yeah.
+
+**Anurag Goel:** Yeah, AWS does. And it comes down to the culture of AWS. They were built around two pizza teams, and two pizza teams do really well when they are focused on a small surface area that can be deep, but then it doesn't -- it's not broad. And they end up building really amazing primitives, that are stable, reliable, scalable... But then you have 300 of these primitives on AWS that you then have to go manage and integrate, and then build on top of. And the cloud has only become more complex over time. And I was always drawn to developer productivity, but then when I started spinning up all of my own infrastructure, I continued to feel a draw towards infrastructure as well. I enjoyed spinning up infrastructure, but also I knew that I was doing sort of the same thing over and over again... So when I combined those two things together, that led to Render's genesis in 2018.
+
+And it wasn't about taking on AWS, it was about solving the problem that I had seen, and in a way that made sense to me from a product standpoint. And AWS is always going to be very, very, very, very successful. Even when Render is a large public company, AWS will still be making a ton of money. And that's because the market we operate in is so massive, and you don't have to have a winner take all mentality. I mean, you think about it... You know, AWS had this massive lead, and now they're still the leader in the cloud space, but Azure and I guess GCP have built businesses worth hundreds of billions of dollars coming from behind. So it just comes down to the demand in the market you're in, and certain markets are winner take all, other markets are multiple people can exist and have their own unique angle on the problem... And that's how I think about Render.
+
+**Adam Stacoviak:** Yeah, it makes sense hearing that backstory why you didn't agree with my thing of taking on Goliath. I still think you're facing some sort of imminent threat, at least not so much today, but in the early days... Because the threat isn't necessarily the technology, it's the ability to attract a developer's attention. To give any care to "Why Render? Why try Render?" But I think a lot of developers had this - and still do - affinity towards Heroku. Heroku was the very first platform as a service to care, I would say, about developer experience, and make it literally just too easy to ship. It was just so easy.
+
+**Anurag Goel:** \[12:02\] It was great. For the time, for 2005, it was magic.
+
+**Adam Stacoviak:** And you followed in those footsteps to take that magic to a new level, because as you said, there's limitations. And I know many friends who have launched their applications there successfully, smaller shops, smaller applications for a small business, so to speak... But then once you get to a certain scale, there is this limitation that was hit on Heroku. And that's kind of -- Salesforce, and other things made it not go beyond where it could have gone. There was a lot of potential that wasn't fully realized beyond what they'd actually achieved.
+
+**Anurag Goel:** And at the same time, AWS kept launching service after service... Because AWS launched in some ways after Heroku did, right? Because we're talking about Heroku launching in -- I guess that's not entirely true. Heroku was on the scene in 2007, and then they got acquired in 2010. So they've now been part of Salesforce for 15 years. And AWS was also -- like, S3 came on the scene around 2005-ish... But then EC2, and others... And I think Heroku could have done a better job keeping up to date with how people were building applications. And they really missed the whole trend around static sites, and containers, and all the other ways in which people built applications. And that's really the differentiation, I think... When you think about Render, from the beginning, it was about how people were building applications today, and what they needed from the cloud. And then that continues.
+
+So fundamentally, the way developers build applications changes every few years. And frameworks come and go, databases come and go, the way we do async processing comes and goes... And these days it's all about calling out to ChatGPT or cloud APIs, and then managing all this data that you're feeding to these models, and then you're collecting the results, and doing all of this in discrete steps and combining all of them together at the end...
+
+So I think the cloud - any cloud - should be constantly thinking about the primitives that developers need, that every company ends up building themselves. And when Render started, it wasn't just about Heroku, it was also about Kubernetes... Because every company ends up building a sort of internal Kubernetes cluster that looks very similar. People end up using the same Helm charts, they end up installing the same sort of standard things... And the problem, of course, is that that's a lot of wasted work. And companies do it because they think that's what needs to be done. And Render is saying "No, actually, what you want to get out of Kubernetes, we give you out of the box. If you want private networking, if you want auto scaling, if you want health checks, zero downtime deploys. If you automatically want to compress your responses. If you want your SSL to be completely managed for you. If you want volumes. If you want the ability to store data on a disk. You can do all of that on Render, and you don't really need Kubernetes." And that has been one of the ways in which we have attracted a lot of people over the years. In January we signed up 110,000 developers.
+
+**Adam Stacoviak:** Is that an uncommon thing, that amount? Or is that a high number? I know it's big, but is that a peak for you?
+
+**Anurag Goel:** It's been a steady increase over time... But to give you some sense of where we were in July of 2022, we signed up 5,000 developers.
+
+**Adam Stacoviak:** Oh, yeah. That's some excess.
+
+**Anurag Goel:** So in January 2025 we're at 110,000, so that's a pretty big ramp up.
+
+**Adam Stacoviak:** Wow.
+
+**Anurag Goel:** \[16:13\] And by the way, all of that is through word of mouth. It's all organic. We don't have a marketing team. I'm looking -- by the way, if you're hearing this and you're a great marketing leader, please come and join Render and build out your team to help grow us faster. We're already growing pretty fast, but we could be doing better.
+
+**Adam Stacoviak:** Wow. So no marketing. That's great.
+
+**Anurag Goel:** No marketing. It's all word of mouth. And that's the thing about developers... If you build something that really appeals to them, that really solves the problems that they're running into every day, you get the benefit of developers telling other developers about it. This doesn't happen as much in other markets. But in a developer market, it's massive.
+
+And so we've really focused on the product, really focused on the look and feel, but also the kind of features that people want, especially as they grow. And our whole thing is, look, we're differentiated in two big ways. One, we're not opinionated... Well, obviously, it depends on who you're comparing us to, but we're not opinionated in how you build your app. We want you to pick any architecture you want. We're not going to tell you, "Look, you can only use serverless functions and have a static frontend" or, SSR, ISR, or whatever the fancy buzzwords are of the year or the month. We're not going to tell you everything is a Postgres, and you have to build functions on top of Postgres, and you have triggers that do everything for you. We want you to build your app the way you do it, because we know that that changes over time, and we find ways in our product to support you however you build your app. So that's one thing.
+
+The second way in which I think we're very different is what I was talking about earlier, which is we allow you to scale on the platform. Both in terms of cost, but more importantly, in terms of the features that you're looking for as you grow. And so we've continued to focus on those features as companies on Render have grown from being just three people spending $30 a month on Render, to now we have folks who are spending $2 million a year on Render, and they're very big in terms of their team sizes, and they've continued to grow, and the bulk of their compute - pretty much all their compute, except BigQuery in this one case - is on Render. And we're making sure that we continue to build for them.
+
+**Adam Stacoviak:** One thing you said that I had not actually heard of before is this idea of two pizza teams. Can you explain that? What is a two-pizza team?
+
+**Anurag Goel:** So this is something that came out of Amazon, where back in the day Bezos or someone at Amazon was like "Look, team sizes should not exceed the number of people who can consume two pizzas." And so if you have a team of 50 people, they'll require way more than two pizzas. That's not great. We want your team to be small enough, and we want their scope to be small enough. And the whole thing around service-oriented architecture - so everything is a service, and then all these teams build these well-defined interfaces that talk to each other, or that people have to go assemble themselves... I mean, in some ways, that's also the Unix philosophy; you have tools that do one thing well, and then you find ways to configure them or use all of them together. So that is essentially how AWS - and not just AWS; I'm sure other parts of Amazon - that's how they thought about organizational design.
+
+\[19:45\] And it's true that you ship your org chart, right? You've heard that one before, I'm sure. And when you think about it that way, that's why AWS isn't able to do an end-to-end application stack well. That's why they falter. And it's not something that they're going to be able to change. But when you think about Render, we're always thinking about stuff at the very top, at the very layer where we're focused on what application developers want. We don't actually build for DevOps engineers. We think they're great, they're valiant soldiers in this battle that they rage against complexity every day, but we live in a world, Render lives in a world where application engineers are responsible, and want to control their compute. So that's a very different mindset and philosophy, and it requires thinking about products in a way that is fully integrated. Because as an app engineer, I don't really want to both integrate Lambda, and API gateway, and something else, and then EC2, and then Kubernetes, and install all of this. No. I just want my app to be out there, to be resilient, to just work, and I want reliable, scalable compute.
+
+So that's how we think about it, and that leads fundamentally to very different products. And I think it's pretty evident in how AWS has tried to build these products at the top level. We've all heard of Elastic Beanstalk, we've all heard of other attempts that AWS has made at building these higher-level products... And they fail. They fail -- well, failure is in quotes, because at AWS's scale, any product you launch will get some adoption, but it doesn't become the gold standard for how to do it. And that's because they always have these trapdoors. When you start out with something like EC2, then your incentives for making a high-level product aren't as strong, because you can say "Well, if this product doesn't work for you, you can always use EC2, and you still make money that way." For Render, that's not true. We're not selling VMs. We'll never sell VMs. Our platform is what we are selling. So if this doesn't work for you, then we lose a sale. So we make it work for you. So that's, again, a philosophical difference.
+
+**Adam Stacoviak:** Yeah. You want the entire application to be there, and you want it to use cron jobs, you want it to use managed Postgres...
+
+**Anurag Goel:** Exactly.
+
+**Adam Stacoviak:** ...you want it to use different services to enable that application developer to have the autonomy to deliver their application and not have to go through an intermediary... Which is really the DevOps world, right? You're cutting that out, not so much because you don't like it, but because you're trying to serve a different sliver of the developer marketplace, so to speak. As you said before, was that the cloud is big, there's room for all, basically, and you're going to have AWS winning, you're going to have Azure winning, you're going to have GCP winning, but you're not going to win -- they're winning in different areas. They're winning different types of teams. Maybe larger enterprises that have uniquely different scenarios where they literally have to have a DevOps team because their business requires them to operate a different way. I don't know, that's kind of how I look at it.
+
+One thing you said, actually, in your acquisition -- sorry, not in your acquisition; in your announcement, I should say, of this $80 million funding that you just announced back in January... Congrats, by the way.
+
+**Anurag Goel:** Thank you.
+
+**Adam Stacoviak:** You said "Most struggle to maintain focus today because they face an impossible choice with cloud infrastructure." And you go on to say "They can either use legacy simplistic cloud platforms that fail to scale and constrain their technical choices, or they could dedicate substantial resources", a.k.a. using AWS etc. "to manage unnecessary complex infrastructure on public clouds." And they burn through precious time, resources, and ultimately just delay time to market, delay their application from existing. Render is in that middle ground there, and you said before that you were number 5 to join Stripe, so that means that you probably exited well, with some deep pockets... You didn't have to go back to work. You said you and your wife said "Hey", I'm paraphrasing some version of conversation you probably had, "What should we do here? Should we just hang out on this beach, or should we go and solve some problems?" What made you want to be in that middle ground there, to put Render in that gap that you described, to solve this problem? Why not just go off in the sunset?
+
+**Anurag Goel:** \[24:31\] Chill out?
+
+**Adam Stacoviak:** Yeah, chill out. Why not? Why go back to work, and -- I'm sure you work hard, right? You probably work 50, 60 hours a week, or more, I don't know... Why? You don't have to. You didn't have to.
+
+**Anurag Goel:** So I often ask myself that question...
+
+**Adam Stacoviak:** Is that right? I bet you do. You're like "Gosh, did I really have to do this...?"
+
+**Anurag Goel:** There are days. There are days when I ask myself that question. But no, I think -- first of all, we were both young when this happened.
+
+**Adam Stacoviak:** Eight years ago?
+
+**Anurag Goel:** Yeah, a little over eight years ago.
+
+**Adam Stacoviak:** \[unintelligible 00:25:07.24\]
+
+**Anurag Goel:** Yeah, but -- so we still had our lives ahead of us. And I think for me, at least, it came down to who I am fundamentally. And I am driven by impact. I get bored when I am not doing stuff. And I'm also just fundamentally somewhat boring. And by that I mean --
+
+**Adam Stacoviak:** As a person, you mean?
+
+**Anurag Goel:** As a person, in that I don't have these sort of Renaissance hobbies that I could spend a lot of time on, and I could go sculpt, or be a carpenter... A lot of people try that. That's almost like a meme in the valley, where "Okay, well, now I made this money through this whatever company, and now I'm going to go be a carpenter for a little bit."
+
+**Adam Stacoviak:** Okay, I didn't know this meme. I love it.
+
+**Anurag Goel:** People do it. But I think -- there's also traveling; like, you spend all your life traveling. I don't -- I'm a reluctant traveler.
+
+**Adam Stacoviak:** Is that right?
+
+**Anurag Goel:** Yeah. I don't love traveling.
+
+**Adam Stacoviak:** Is it the plane? Is it the people? What is it?
+
+**Anurag Goel:** You know, I've got a great setup at home.
+
+**Adam Stacoviak:** "I'm comfortable."
+
+**Anurag Goel:** I'm comfortable.
+
+**Adam Stacoviak:** Alright.
+
+**Anurag Goel:** Yeah, exactly.
+
+**Adam Stacoviak:** "I know where the bed is, and I like it."
+
+**Anurag Goel:** I like it. I have all my stuff here... I enjoy it, I have my routine. No, so jokes aside, I think there is something about fundamentally who you are and what drives you... And for me, it was important to, again, spend -- I want to look back at the end of my career or my life, and be like "I contributed to solving this large problem." And I know that that will make me feel good about how I spent my time. So that was what eventually led me to the search.
+
+**Adam Stacoviak:** Can you tell me about, if you don't mind, a layer deeper? Your wife - is your wife in technology? I'm not familiar. What's her background? How could she contribute to "Yes, Anurag, you should go back to work and build this platform that will eventually, potentially", and as you said before, "will" IPO. I mean, that's an audacious goal, and I think you're probably going to get there. Not so much what is her qualifications, but does she have some awareness of the true requirement to do what you've done?
+
+**Anurag Goel:** Oh, she does. She knew what Stripe was.
+
+**Adam Stacoviak:** Okay. Sure.
+
+**Anurag Goel:** \[27:57\] She knew how hard we all worked. She had seen all the early people go through everything that we went through... So it was very clear. We knew what we were getting into. And I think it was -- it's just the nature of our relationship where what's really important to the other person is important to me. Or what's important to me is what is important to her, and vice versa. And we've been together a while. I think it's going to be it's going to be 18 years.
+
+**Adam Stacoviak:** Oh, wow. Congratulations.
+
+**Anurag Goel:** Thank you. Yeah. This year.
+
+**Adam Stacoviak:** You must be doing something right then, because you know, that's a long time.
+
+**Anurag Goel:** It is, and it also doesn't feel like 18 years. It's just -- wow, you look back and now that's 18 years.
+
+**Adam Stacoviak:** Do you like Bob Seger, by any chance?
+
+**Anurag Goel:** I don't. I've heard of him, but I --
+
+**Adam Stacoviak:** There's just one phrase from one song he has, "20 years now, where'd they go? 20 years, I don't know. I still look at myself sometimes, where they've gone." But essentially, this idea that "Wow, 20 years have gone by so quickly", and you're alluding to a version of that.
+
+**Anurag Goel:** Yeah. I still remember the first time I saw her. I very distinctly have that memory in my head. And she thought I was a snob.
+
+**Adam Stacoviak:** Is she listening to this show? She's probably listening to this show now. She's tuning into this chapter at least.
+
+**Anurag Goel:** She's probably going to listen to it, yeah. So I'm probably going to get some brownie points. But no, I do remember the first time I saw her, and she thought I was a snob when she first saw me. For a variety of reasons. But then I fixed it.
+
+It was fine. And here we are, 18 years later.
+
+**Adam Stacoviak:** \[laughs\] Cool. Well, it's good to be equally yoked in a decision like that. Because I can only imagine that you had a choice, to go back to "work", to climb another mountain... And obviously, once you begin to climb, it would not make sense to even start the climb if you're not going to finish the climb, right?
+
+**Anurag Goel:** Oh yeah, of course.
+
+**Adam Stacoviak:** So why would you go and even start if you're going to quit, or decide to change that decision a year or two into that choice? It just doesn't make any sense. You alluded to the potential of an IPO... So your focus is on IPOing Render?
+
+**Anurag Goel:** Well, my focus is on continuing to solve our customer problems...
+
+**Adam Stacoviak:** Of course.
+
+**Anurag Goel:** ...and growing revenue, and doing it in a way that allows me to keep Render an independent entity, and for me to have enough control over what we build and how we build it, and to work with the people I want to work with... And all of that essentially means staying independent over the very long term. We've also raised all this capital, and we have a responsibility to make money for our investors, for our employees... And so all of that, I think, does naturally mean in today's environment an IPO. But that's just like a step in the long-term journey. Because one thing you realize is post-IPO I'll still have the intensity of problems that I have today. They're not going to go away. We'll just have a different set of problems. You just go faster. It never gets easier. This is a famous cycling quote. And it doesn't. It doesn't get easier.
+
+We've raised all this money. In some ways, my life is harder than it was when I wrote the first line of code for Render. And it's harder because I have to constantly reinvent my job every six months. And learn very quickly. And learn a lot of stuff that has nothing to do with engineering, on how to run a company. And that's never going to change. So the IPO is just a step in that journey. We're nowhere close to an IPO, but the way I think about it is long term this company should outlast me. And we talk about it internally as this being a generational effort. And all of that points to staying independent, because we've seen what happens when you get acquired by Salesforce.
+
+**Adam Stacoviak:** Yeah.
+
+**Break**: \[30:30\]
+
+**Adam Stacoviak:** Let's examine that. Can we examine reinventing your job?
+
+**Anurag Goel:** Sure.
+
+**Adam Stacoviak:** What you do. Can you describe -- I imagine early on CEO is just simply a label. Obviously, somebody has to hold that title in an organization. Can you walk me through -- I assume now you're far more CEO than you were first line of code, right? Can you walk me through the steps of reinventing the role that you've fulfilled?
+
+**Anurag Goel:** Yeah, absolutely. And even learning that that is what you need to do is a journey. So I didn't know when I started the company that I would need to reinvent my job every six months. And I think for the first at least two years - actually longer - the bulk of my time was spent on essentially IC work, where I was coding, and the rest of my time was hiring.
+
+I think early on in a startup journey, the CEO has three jobs. They have to raise capital, they have to hire, and then they have to build product. And whether they're technical or not, they have to build product. And that changes over time. And so for me, as you hire more people, it becomes clear that -- especially as you raise more money and you're able to continue to hire people, it becomes clear that the organization needs a different set of things from you. You need to start thinking about planning. You need to start thinking about how all these individuals are going to start working with each other in a way that is driving everyone in the same direction. And you think about management, and you think about hiring managers, but also, well, okay, what are your expectations for the managers? How do you even evaluate managers? And how do you make sure that they aren't actually slowing your engineers down? How do you think about the non-engineering and product side of the org?
+
+So for the longest time, I was the only person responsible for "product" at Render. So the company started in 2018, we launched live on stage at TechCrunch Disrupt in 2019 Battlefield, which we won that year. That was kind of insane.
+
+**Adam Stacoviak:** Do you still have that big check?
+
+**Anurag Goel:** Yeah, it's in here, somewhere in the office.
+
+**Adam Stacoviak:** You don't have to get it, I'm just curious... I'd love to see a picture. If you can, as sort of post-promotion of this show, I'd love to see a picture of that check on social media as part of this. That'd be awesome.
+
+**Anurag Goel:** Yeah, of course. I'll send it to you. That was kind of crazy. You obviously see this stuff on TV...
+
+**Adam Stacoviak:** Silicon Valley, the TV show, man. I mean, come on.
+
+**Anurag Goel:** Yeah, it was --
+
+**Adam Stacoviak:** You were the Pied Piper... In terms of the winner, not so much the actual application. And maybe the exact thing was happening, but you got the check and you won the Battlefield round.
+
+**Anurag Goel:** \[37:54\] It was less dramatic. It was just a lot of hard work, I'll say. And you're on stage, and there's all these incredibly accomplished people, and Ashton Kutcher...
+
+**Adam Stacoviak:** That's hilarious.
+
+**Anurag Goel:** And it was quite an experience. But I think at that point -- Render was actually just four people when we won in 2019. And we launched the thing... And when you launch it, once you have customers, things just feel very different. Once you have people who want more things from you, then you have to get out of build mode and you have to really focus on listening to them. And being my job was much more around "Okay, well, what do people need?" Talking to them, making sure that I was steering the company, the product in the right direction, while also building it on the side... And then we hired our first product person, who is our VP of product now, in December 2022. So for four years we didn't really have a product person. It was just me and the other engineers... Which I think served us well. But at a certain point, you have to start thinking about the other functions you bring in. So then you think about bringing in finance, you think about bringing in someone to manage your people team. You think about bringing in marketing, but that never worked for us.
+
+But that's the other thing, how do you grow this thing? And I think we were lucky in that we kept growing, despite not having any marketing or sales. So especially in the early years, it was an afterthought, and building the product and allowing people to scale on the platform and giving them what they want was front and center.
+
+I think now we've raised all this money and we have 80-ish people in the company, and we're growing fast. We're hiring fast. And my goal now as CEO has switched to repeating myself over and over again to make sure that we're all aligned around the same kind of thinking around what we're doing, why we're doing it, who we're doing it for, and what needs to be done. And the CEO's job becomes just constant messaging, and repeating the same message over and over again, once you get to a certain size. But it also becomes hiring -- I mean, hiring executives is... You talk to a bunch of CEOs in my position, everyone is going to tell you hiring executives is hard, and it is also the highest leverage thing you can do.
+
+**Adam Stacoviak:** Yeah.
+
+**Anurag Goel:** And that is a lot of what I spend my time on - hiring, and making sure that we're all moving in the right direction, that everyone knows why we're doing what we're doing... Because it turns out -- you know, all the people who joined early on, they knew what we were doing. They didn't need to be told. They just went ahead and did it. But then the 80th person you hire has a very narrow view of the organization by default. They're coming in into a specific team, they have a specific project that they're taking on... But the whole architecture, the stack is much bigger. So you still want them to have the autonomy to make the right decisions day to day, and the only way they can make the right decisions is if they have the broader context. So sharing context, sharing everything that they need to do their jobs well - that becomes your primary responsibility.
+
+**Adam Stacoviak:** Wow. Is that -- so it sounds like you must have gone through a period where... I imagine you don't write code too frequently today. You don't write a lot of code, I imagine.
+
+**Anurag Goel:** I wish I did. I wish I did.
+
+**Adam Stacoviak:** If that's true, then why did you do what you've done?
+
+**Anurag Goel:** No, no, no. I wish I did in the sort of --
+
+**Adam Stacoviak:** \[42:11\] I understand what you mean. I'm being facetious, to some degree, and somewhat rhetorical in that response. It's just a joke, but...
+
+**Anurag Goel:** No, it's a good question. I wish I wrote -- so here's the thing. I wish I wrote more code on the weekends. Just on the weekends. I don't wish I wrote code day to day. But my weekends are spent doing other things.
+
+**Adam Stacoviak:** What do you do on the weekends?
+
+**Anurag Goel:** I'm spending time with my wife, man.
+
+**Adam Stacoviak:** Okay. I get that. I mean, you never know, though. You never know. You could be doing something like a side project just to keep you warm, just to keep your skills refreshed and whatnot.
+
+**Anurag Goel:** Yeah, so I need to get into that. For some reason, that hasn't -- I think I'm just in a state of trying to recharge on the weekend. I actually end up playing video games on the weekend, real-time strategy...
+
+**Adam Stacoviak:** Oh, man. Do you build machines? Do you build your own gaming machines?
+
+**Anurag Goel:** I did build my own machine, and I still use it.
+
+**Adam Stacoviak:** Okay. How long ago? Recently?
+
+**Anurag Goel:** No, that was 2016, when I built this startup... Well, if you can call it that. It was a one-person startup that I ended up selling to another company. But I built the first GPU-backed Jupyter Notebook online that you could start with a single click. And it had all these data scientists who were really running into tons of trouble trying to spin up a Jupyter notebook backed by a GPU to do deep learning, which was the craze back in 2015, 2016... And I saw that, and I was taking one of these courses where -- I wanted to learn deep learning, and so this was my exploratory period. And then I realized that this was insanely hard for most people, and I went and built a whole thing on top of Kubernetes, no less, to offer people a one-click Jupyter Notebook.
+
+But then Google came in, and they were like "Hey, we'll give it to everyone for free." And to be completely honest, I didn't care as much about data scientists as I cared about developers, application devs... And so then I started working on Render, and I was like "I don't want to do this thing anymore." So then I sold it through a tweet. I tweeted about how I am now focused on something else, and there's 10,000 data scientists using it... I would love for someone else to take it over. And then there's this company called Dock.ai, and they acquired it for me, for a relatively small sum, and I was happy.
+
+**Adam Stacoviak:** So how does that relate to you building your own gaming PC?
+
+**Anurag Goel:** Oh, so I needed to run deep learning on my own GPUs. And so it wasn't intended to be a gaming machine. It just became a gaming machine later.
+
+**Adam Stacoviak:** Because it had no more use.
+
+**Anurag Goel:** Because - yeah, it had no more use.
+
+**Adam Stacoviak:** You're like "I've got this GPU, this massive GPU..."
+
+**Anurag Goel:** I had two of them. I built a two-GPU machine.
+
+**Adam Stacoviak:** Yeah, it's kind of crazy how that swings back again, because -- and we could probably touch on this to some degree when we get to your focus now with Render on AI, and what that means in your announcements, because I have some questions around that... But there's this new resurgence to -- especially with Ollama, and public and/or open LLMs to be used and consumed in your own private inference, essentially.
+
+Finding a GPU, whether it's an older gen like a 3090, which is still relatively useful, 4090, which is one of the most earliest models, and then now 5090s - those are all at the top tier of NVIDIA's RTX-level GPUs. Good luck finding those things, right? Good luck finding them, and if you do, they cost just as much as the original price tag, if not more, because there's such a demand for them. I can imagine that's -- it's wild.
+
+\[45:57\] It's kind of crazy that it swings back again. You were playing with that, and maybe instead of Render -- I guess maybe now you're best positioned to maybe pick that back up again, because I imagine part of this innovation you're talking about in your announcement post, which I'm sort of paraphrasing... Let me see if I can pull it up real quick, just so I can be more clear about it. You mentioned -- you allude to the new chapter for Render, and then the desire to build for the AI era. I imagine -- when I read that, I think you're going out to buy a bunch of GPUs, and you're taking a lot of that $80 million, or a large chunk of it, and you're like "Okay, this $20 million is for GPUs", considering the price tag I just mentioned for your sort of consumer, prosumer-level GPU. I imagine building for that AI era is tech on top of it, software and orchestration on top of it, but it's ultimately acquire GPUs, plug them into machines, and make them available.
+
+**Anurag Goel:** So it's actually not that.
+
+**Adam Stacoviak:** Okay. My assumption is incorrect.
+
+**Anurag Goel:** Well, it might become that, we don't know... But we don't have enough evidence to say right now that that is what we should do. But I think it goes back to what I was saying earlier about people changing how they're building applications... And we realized early on in 2022 that our customers were not the kind of customers who were going to go train their own models. And we're not selling -- Databricks is not using Render, or Grok 3 is not built on Render. We're selling to application developers. And application developers are not in the business of training their own models. They use ChatGPT, or Claude, or whatever, and they build on top of that.
+
+We did a bunch of customer research, because in 2020 we were like "Well, should we go buy a bunch of GPUs?" And you and I both know of companies that are peers that went and did that, and then realized a bit too late that they shouldn't have done it. But we knew early on that this was not the path for us, and I'm not convinced that that has changed for us.
+
+So when I talk about what we're doing for how people are building apps in the AI era, I'm thinking about the ways in which we can make it easy for people to run a lot of distributed compute, or compute that requires a lot of retries, or compute that is based on a lot of different conditions.
+
+**Adam Stacoviak:** Temporal then. You're trying to build what Temporal's built.
+
+**Anurag Goel:** I can neither confirm nor deny that statement.
+
+**Adam Stacoviak:** Well, that's what you just said. I can confirm that by what you've just said.
+
+**Anurag Goel:** Well, I think there is an opportunity for us to build something that makes that kind of problem a lot easier to solve for application developers.
+
+**Adam Stacoviak:** Yeah, it's necessary.
+
+**Anurag Goel:** Exactly.
+
+**Adam Stacoviak:** Yeah. Well, if you don't, your application falls over. You've got something waiting, so your application is blocked. You need to have that resilience.
+
+**Anurag Goel:** Exactly. Exactly.
+
+**Adam Stacoviak:** And the stack doesn't necessarily -- you have to build it yourself. Or - not an ad. They're a sponsor of ours. I'm aware of how their platform works... But they may actually sponsor this episode. I don't even know, because we don't know which episode is being sponsored when we produce it, but...
+
+**Anurag Goel:** And that's totally fine, because I think Temporal is a great product. But again, it requires a lot of wrangling, and it requires a lot of like figuring out. And Render is really good at making complex things easy, and giving people an interface that they like to use, and that they love talking about... And that's our strength. And when you combine that with our strength around running millions of applications at scale reliably, then you kind of see where this is going. But that doesn't rule out Render offering GPU inference.
+
+\[50:16\] Again, model training is not where we would focus... But you can imagine -- you still can't run serverless inference in a way that is... I mean, there are places where you can go do it, but the other thing we've realized for Render is people generally want everything in the same place, to the extent that they can. And Render is never going to build literally every single service that AWS has built. That's not who we are. That's not the way for us to win. But there is a core stack that we target, and it goes back to the components that everyone needs, or that 80 percent of the market needs. What does that mean for what Render provides for that component? And so one of the things that often comes up, especially for AI companies, is object storage. They want -- they're running all their other compute on Render. They want to run managed inference on Render, but they also want to store their models on Render. And they want to -- like DeepSeek, you store the model on Render, you run managed inference against it... Render doesn't have object storage today. People can use S3 if they want, and they do use S3 while they're using Render. But we consistently hear from people, "Look, S3 is the only part of my stack that is on AWS. And I would love for you guys to fix that." But it's been a while, and we've chosen not to build S3.
+
+**Adam Stacoviak:** It's kind of funny, your next innovation is like the oldest tech, basically...
+
+**Anurag Goel:** Yeah, the thing that AWS built first.
+
+**Adam Stacoviak:** Yeah. I mean, it's what they're most known for. They've wanted so much that S3-compatible is the way. I'm sure that when you build your object storage, you're going to make it S3-compatible, because you have to.
+
+**Anurag Goel:** Absolutely. Yeah. But again, it's not -- so the way we think about product is not to build technology for the sake of it, it's to build technology to solve the very real problems that people have.
+
+**Adam Stacoviak:** Yeah.
+
+**Anurag Goel:** And that might lead to what from the outside might appear to be sort of a boring platform. But our numbers say otherwise, and our revenue says otherwise.
+
+**Adam Stacoviak:** Well, you need to follow that boring tech mentality, too. I think there's a certain recipe of innovation, and there's a certain recipe combining that with boring. Boring is stable.
+
+**Anurag Goel:** Boring is stable. Boring is secure.
+
+**Adam Stacoviak:** Right. Boring is capable.
+
+**Anurag Goel:** Yeah.
+
+**Adam Stacoviak:** It's not going to fall over necessarily in the middle of the night, and cause me and my team headaches because I chose Render's super-innovative next thing. It's more like "Well --" That's kind of funny, that your next innovation is object storage.
+
+Okay, so you've got -- you currently have GPUs running, so you're running inference...
+
+**Anurag Goel:** No, we're not.
+
+**Adam Stacoviak:** Okay, you're not.
+
+**Anurag Goel:** No, no. I was just saying we can't rule that out. So some of that $80 million might go towards figuring out the GPU situation.
+
+**Adam Stacoviak:** Do you think it'd be $20 million? I was just joking about that number, but man, I can imagine if you had to stack some GPUs, it'd probably be about $20 million to outfit a proper setup.
+
+**Anurag Goel:** Yeah, I don't know... I don't want to spend $20 million on GPUs. \[laughter\] I just don't.
+
+**Adam Stacoviak:** Who does, right?
+
+**Anurag Goel:** Yeah, who does?
+
+**Adam Stacoviak:** Because two years from now or three years from now, when the demand curve swings back, you'd have probably paid 3x, 4x more than you should have.
+
+**Anurag Goel:** \[53:59\] Yeah, but the good thing is there's a bunch of people who already spent that money on GPUs, and they are looking to make some money from it, because they're just lying there, being wasted. And so there is a marketplace approach to this, and there are companies that are becoming effectively GPU marketplaces. SF Compute is an interesting company that we've come across recently. So we don't have to go buy our own GPUs, and there is a world in which we test the waters by looking at some sort of marketplace approach.
+
+**Adam Stacoviak:** So without revealing or me asking you to reveal too much of your cards, or even beyond my question mentioning Temporal ever again, in terms of this in particular, is the next thing you're trying to build that retry world where you don't have an application fallover? Because going back to your original state, that I paraphrased from your announcement post, was like you've got this one camp that hangs out in legacy, and this other camp that hangs out in public cloud, that they have to build everything. And there's this middle layer, which is called Render, which we've been talking about, of course, and that's where some of the orchestration is done for on behalf of an application developer. There's application developers that want to utilize APIs, whether they're for artificial intelligence, inference, or whatever it might be. They're using APIs every single day in their application... But they've got this problem. And you essentially watch that with cron, or background jobs, or queues, and you've got to manage those orchestrations... You've probably got dashboards and logs, just like PagerDuty, and whatever you might have. So I imagine -- can you give me a peek into what you're doing to solve that problem?
+
+**Anurag Goel:** Yeah. So the way people manage it today is exactly what you said, which is you run a background worker, and you have a Redis queue, and you run something like Celery, or whatever. And I think, again, that architecture served us well for maybe 10 years, but it's showing its age, based on the new kinds of requirements that a lot of applications have if they're tapping into these long-lived API calls that are streaming data, WebSockets are becoming increasingly more relevant... We have a lot of people move over to Render from Vercel, because we have WebSockets, and they can run Next.js in a way that allows them to call out to these LLM APIs.
+
+And so when we think about not just for AI APIs, but even async tasks, things that need to happen, you need to make sure that they happen once and only once, and then you need to have conditional events happen, and so you basically build out a whole graph of activity... And you want to be able to just express that in your code. So you just want to say "This function - this is my job. Run it once. That's it. I don't want to care about anything else. I don't want to spin up background workers... I don't want to spin up queues. You do all of that for me. Make sure it's reliable. Give me a dashboard that tells me exactly how things are going. Give me observability into it, and charge me only for the time that my task is running. And give me different levels of compute, depending on the task." And I can tell you what level of compute, and maybe you autoscale compute vertically depending on the task.
+
+And so there are a lot of these questions and answers that we are thinking about as we think about this next product...
+
+And I think it's really exciting. I think it's fundamentally changing the abstraction level of how people run a bunch of distributed tasks, and making sure that we give people the same -- again, it's all about what your infra team would do for you if you were on AWS.
+
+\[58:10\] Your infra team would set up Temporal, and they would train you on how to use Temporal, and then developers would have to learn Temporal. We're replacing a lot of that, and we're saying, again, "Application devs, if you want all of this, here it is." And you don't have to worry about how many Temporal workers you have, whether they have enough memory, whether you need to increase the number of Temporal workers you have... So there's a lot of that just unnecessary work, that even we have to do, to deal with something like Temporal. But I think Temporal is a great product, and I think that they serve this growing market, this need that they're never going to be unique in serving if it's a problem that everyone faces.
+
+**Adam Stacoviak:** Are you friendly with other CEOs?
+
+**Anurag Goel:** It depends on the CEO.
+
+**Adam Stacoviak:** Are you friendly with Samar, for example, Temporal's CEO?
+
+**Anurag Goel:** I don't know him. I don't know him.
+
+**Adam Stacoviak:** Okay. I've talked to him before. You guys seem passionate about solving this problem, so I bet you guys would be good friends.
+
+**Anurag Goel:** Sure, yeah. Yeah. I mean, again, I think what they've built is great.
+
+**Adam Stacoviak:** I think -- yeah, when I look at what you should focus on, it seems, in my prior examinations of what I assumed Render was, and then now my examinations of what Render is based on this conversation with you, you look at the marketplace of being a developer, in particular an application developer, because that's who you serve... And you say, "What problems does that kind of developer have deploying applications?" And "Let's just solve their problems." And then you pick one product after the next after the next, and you just keep layering on until you have a full suite of essentially batteries included, configuration not required. It's convention over configuration, here you go. A lot of great choices. Now, you can fine-tune it if you want to, it's off the beaten path, but here's the blessed path, application developer. Go have fun. Deliver your application.
+
+**Anurag Goel:** You just described our business strategy.
+
+**Adam Stacoviak:** There you go.
+
+**Anurag Goel:** Yes, that is exactly what it is.
+
+**Adam Stacoviak:** Okay. So when it comes to, I suppose, this -- if you want to say it, this $80 million that you've got... Did you need that money? Did you need to raise another round, or did you do it because it was time, it was necessary? Why raise a series C? Why that for you all? I know you have to innovate. Do you need that cash to sort of accelerate certain things? Because that's usually why you raise cash, is to accelerate a new plan.
+
+**Anurag Goel:** Yeah. So I think for us there are two big reasons. One, increased legitimacy. I think there are a lot of people, both customers and potential hires.
+
+**Adam Stacoviak:** Perception, essentially.
+
+**Anurag Goel:** A lot of it has to do with perception, because you look at a series A company, and they have this product, and it might be great, but they're like -- series A companies die at a much higher rate than series C companies. Everyone just knows that. And there are people who will join a series C company because it has reached a certain level of product-market fit and has reduced the risk to a large extent. But they won't go and join a series B company.
+
+And so it was important to us to continue to have this public perception of stability and long-term viability, given what we do. And as soon as we announced the round, the number of people building serious things on Render, that has taken off. So there's that.
+
+**Adam Stacoviak:** \[01:01:55.00\] Okay. How do you maintain an awareness to that? Answer the rest of your question, and then come back to that one, because I want to hear what you were saying. I was just thinking about that in the moment... I'm like "Gosh, how do you pay attention to serious applications, or not serious applications?" But anyways.
+
+**Anurag Goel:** Yeah, I'll get to that. But the second reason is - of course, we need the money to hire more people. I need to build out a marketing team.
+
+**Adam Stacoviak:** Marketing...
+
+**Anurag Goel:** Yeah, marketing and sales. We have one -- again, a lot of people might not be interested in Render after hearing this, but we have one account exec. That's it.
+
+**Adam Stacoviak:** If I was a salesperson who had no job or wanted to transition to a new job, I would kill for that job. You know why? I can make all the money. Or a lot of money, because if you've got one across - I don't even know what your revenue rate is, but I'm sure it's going to grow, being a series C... Well, I mean, that's a great place to be in as a salesperson or any account executive, is to be in a position where you can collect.
+
+**Anurag Goel:** Yeah, but the thing is, most people who use Render never need to or want to talk to sales.
+
+**Adam Stacoviak:** Sad. I mean, happy, but sad.
+
+**Anurag Goel:** Sad for the salesperson, yeah. \[laughs\]
+
+**Adam Stacoviak:** Well, I think that's the way to go, right? You want to have that sort of on-demand to the application developer. That's the way they think. An application developer is not going to want to sit down with somebody. However, sometimes they're going to need to convince other people, and that's where the AEs come in. It's like "Hey, I'm an application developer in my company, but I have to have my CTO's approval, and he/she has to meet somebody at Render to legitimize this spend, or my ability to even give it a POC."
+
+**Anurag Goel:** Yeah. And we have a sales engineer. Also just one sales engineer. So I think part of it is just there are people who just want to talk to someone, like you said. And it might not even be to legitimize it to anyone else, it would just be for their own comfort.
+
+**Adam Stacoviak:** A real email address, a real person.
+
+**Anurag Goel:** Yeah, a real person. They want to get on a Zoom call. They want to make sure that they hear from someone how things are going to go... And then they have this person's contact now, and if things don't go well, this is a person they call. And so there is that level of human interaction that a lot of people need to feel comfortable about trusting their entire business over to us... Because we're not building a tiny thing for them, we're running their entire business online. And so that's why I think it's not even sales, it's essentially help and education. And we don't have this, like, "Oh, you need to meet this quota" as a salesperson. We don't have that, because a lot of the self-serve people will reach a point and then they'll be like "Yeah, I need to talk to someone." And do we count that in sales comp? Do we not? That's an interesting question. And a lot of people migrate to us from a places like Heroku, AWS, and they self-serve migrate... But during that process -- they're going to migrate anyway, but during that process, they do feel the need to talk to someone... So, that is probably sales-assisted. And so you count that in their quota... But it's too early for us to set quotas, given that we don't effectively have this.
+
+But to go back to your question about -- so the money that we're raising, we do need to build more things faster. We're going to do that. So it's not just marketing and sales... Actually, a lot of it is going to go towards hiring more engineers, and hiring people who were not accessible to us just because we were an earlier stage company. You also were asking about -- the other question that you said we were going to come back to...
+
+**Adam Stacoviak:** \[01:05:58.08\] Yeah, how do you pay attention to serious application?
+
+**Anurag Goel:** Yeah, yeah. Honestly, the simplest way to tell is they use their work email, and they're not using a Gmail address.
+
+**Adam Stacoviak:** So not a Gmail account, or not an AOL, or ICQ... Something that's just based on that.
+
+**Anurag Goel:** I don't think ICQ ever had email, but yes. AOL, yes. Is ICQ still around?
+
+**Adam Stacoviak:** I don't think so... I fumbled that one. It shouldn't have been ICQ, it should have been like EarthNet, or something like that.
+
+**Anurag Goel:** Or CenturyLink.
+
+**Adam Stacoviak:** Yeah.
+
+**Anurag Goel:** Yeah, CenturyLink.net.
+
+**Adam Stacoviak:** Or SBC.net. Your ISP's email. Who in the world ever uses their --
+
+**Anurag Goel:** Yeah, exactly.
+
+**Adam Stacoviak:** Precisely.
+
+**Anurag Goel:** Exactly. Anyway, slight tangent... So yeah, so not having any of those emails, and having a work email makes sense. But we also have a lot of people using Render for specific things while working at large companies. We have people who have signed up with Meta.com, and we have Alibaba.com, and all these other... Even Amazon and Google... These are paying customers. People have signed up with their work email, who are not actually running a large thing from Google or Amazon on Render. It doesn't make sense to do that. But they're using -- so these might be like data scientists, who want to run something in Python and they just can't get the support they need from their DevOps teams. And this is very common in large companies, where you have enterprises where their main application is handled by the DevOps team, but then people want to build other things, whether it's data scientists, or sales engineers... Even marketing engineers, they want to do stuff that they want to put out in the world, they want it to have a URL that other people can access, and they just don't get the support that they need from DevOps, because DevOps couldn't be bothered, and they're a bottleneck, because they have all these other needs that they need to fulfill from the main application. So we get a lot of people like that.
+
+But fundamentally, the people who we're most excited about today are early-stage startups, because Render is a great fit for them. When you're a startup, you want to move fast, and you want reliability, and Render gives you both of those things.
+
+And then the second sort of serious customer segment is what we call tech-enabled businesses. So tech-enabled businesses are those where the core competency isn't actually software. What they're doing is something like they're building a mortgage CRM, or they're a law firm and they need technology to achieve a specific goal. Or they are in some cases also like a crypto company that needs something specific, and they need to host it online. But they're not in the business of hiring DevOps engineers. They might not even be able to hire DevOps engineers. And so they are very, very happy with what Render and products like Render provide.
+
+And then I think we're increasingly seeing what I was talking about earlier, where there are these teams within larger companies that end up using Render because it gives them the same speed of iteration, velocity. We also just launched things like SSO, and SCIM...
+
+**Adam Stacoviak:** Finally, huh? Wow.
+
+**Anurag Goel:** I know, yeah.
+
+**Adam Stacoviak:** Crossed that chasm, finally.
+
+**Anurag Goel:** \[01:09:56.16\] Yes. So every startup wants to change the world, and we're like "We're going to change the world", and then the customer is like "Do you have SSO?" And everyone ends up building it. So here we are. And we actually built it ourselves. We didn't use one of these companies that everyone ends up using, for a variety of reasons. Also our engineers are great, they want to build this stuff... So it's now available in early access. It'll be fully available in a month or so... But there are these companies that want this level of compliance, and they want this level of security, and so SOC 2, ISO 27001, HIPAA... We're working on HIPAA. It should be available this fall. No, actually this late spring is when we'll be ready to work with companies that require HIPAA. So all of that indicates different levels of seriousness.
+
+**Adam Stacoviak:** Yeah \[unintelligible 01:10:53.04\] Well, you're a serious C company now, you have to have this enterprise level feature set. I mean... Right? You have to. Because now you're going to attract these folks, and it's a no unless it's a yes on those two fronts, or those fronts, you know?
+
+**Anurag Goel:** Yeah, exactly. And it makes sense, because you don't want to go in and remove -- a lot of SSO is really about offboarding. People don't want SSO for actually single sign on data. Security departments don't care if you use Google to sign on, although, I mean, they might... But fundamentally, it's about control of who can access this app, especially with something like your infrastructure... And so as soon as someone leaves the company, you want to make sure they don't have access to your most secret and most critical infrastructure. And so SSO is really about offboarding more than anything else, so you have to build that in.
+
+And then you have SCIM, where you add someone, they automatically get added, and you offboard someone in Okta and they automatically get offboarded from Render, and all of that. And you assign roles and permissions... All of that becomes pretty important as you become a serious company.
+
+**Adam Stacoviak:** So you built it versus bought it. How did you -- were you personally involved in that choice? What made you build versus buy?
+
+**Anurag Goel:** I was personally involved to the extent that when they first said they were going to build it, I was like "Why?" That was my first thought.
+
+**Adam Stacoviak:** Okay... As you should as a CEO, you're like "Why? Because we can buy this right now..."
+
+**Anurag Goel:** Yeah, exactly. And our VP of engineering, Mike, was the same. He was also like "Why? Why are you doing this?" But then, actually, when we dug into it - and I didn't dig into it, but other people did... And part of it is also that we're not one of those B2B SaaS companies where you are signing up a few hundred people every month, or whatever... Like, we're signing up all these people, right? 110,000 people.
+
+**Adam Stacoviak:** A tremendous amount of people, yeah.
+
+**Anurag Goel:** And Render's a kind of thing that will require a lot of SSO seats, so to speak, once you get into an organization. So we wanted to make sure that we weren't spending a ton of money. But also, a lot of these things work well. So if you have like Okta, or Auth0, or Stitch, or WorkOS, they work really well with SSO if you've also integrated your auth with them. And we were not going to move our auth over to them. And part of that was just because we built a strong auth system, and we need to integrate our auth with our own GitHub auth, plus GitHub permissions for your apps... And there's some complication there, so it's not just plain auth.
+
+\[01:13:58.27\] And also, again, we're not going to pay them for 110,000 signups a month, or 120,000 signups a month. And so it just didn't make -- first of all, it didn't make economic sense for us. I'm sure it makes economic sense for most companies, which is why a lot of these companies exist, and are doing well. But for us, it just didn't make sense. Plus, the team that was building it, that built it previously at Figma. So they were just like "Hey, we just did this at Figma. We can go do it again." And they did. They built it really quickly, and it works, and I'm sure that we'll continue to improve it; it's still in early access. But I feel good about it.
+
+**Adam Stacoviak:** Just curious, really... I don't really care about the number necessarily, I'm just sort of curious of - if you saved, which I imagine you did... Was it dramatic? Was it a dramatic savings to build it yourself, and to have control? Because I imagine the reason why you did it was because, as you said, you have people on your team that did do it recently... And so it's fresh to redo again, so why not... I imagine you have some autonomy and control over your choices within implementation that is unique to the platform... So you have a bit more control, whereas maybe in other scenarios you may be a bit more restricted potentially... And then obviously there's the savings because now you don't have the monthly bill based on the user count.
+
+**Anurag Goel:** I think all of those things are true to varying extents. The biggest thing was we had this existing auth system, and we just didn't see enough value in using these things for just SSO... And it was a big risk for us to switch our auth over to the new provider.
+
+**Break**: \[01:15:43.03\]
+
+**Adam Stacoviak:** I guess what's next? I mean, you're spinning up marketing, you've crossed or are in the process of crossing the enterprise chasm, which is to offer SOC-2 compliance etc. Letting this series C be the financial boom, let's just say, or fuel to the fire, as well as the legitimacy to attract newer types of customers... I guess, how do you go from here? Where do you go from here? This raise was just last month, and your job is reinventing itself time and time again. I imagine this raise may give you one more variation to your personal role of reinvention.
+
+**Anurag Goel:** Yeah, I think I expect the next year for me personally -- I'll end up spending more time both externally and internally, sharing vision and strategy, and talking about Render. And like I told you, I'm a reluctant traveler, but I've already got business trips lined up this year.
+
+**Adam Stacoviak:** Okay. Are you coming to Austin at all?
+
+**Anurag Goel:** Not yet, but I'd love to.
+
+**Adam Stacoviak:** If you're coming to Austin, I'd love to meet up, man. That's where I'm at.
+
+**Anurag Goel:** I would love to. I absolutely would love to grab a beer. But yeah, I think business travel is in my future. It's that fortune cookie that I got...
+
+**Adam Stacoviak:** There you go. Oh, yes.
+
+**Anurag Goel:** Yeah, I know. But it's also about, as you hire more people, I think talent density is just so critical. And that is how we built the company so far... And we want to make sure that we we keep talent density high. In fact, if anything, our bar should be higher for the next stage. And so that is going to be critical for the company. I also think it's going to be really important for us, for me personally, and our leaders and the engineers who are working on these new things that we talked about - they're all great, but I personally feel a need to stay informed, and close to these things at a high level... And so part of my job would be to kind of calibrate my level of involvement in these things, in these large investments. And I have previously been very involved, but I'm not sure that that is not the right thing to do going forward as well.
+
+So I think it's going to be thinking hard about that, it's going to be talking to other founders and CEOs who have gone through these stages. That's always incredibly helpful. And yeah, just getting better at my job as a CEO. So nothing changes, right? It's all the same. This is just a milestone, and my goal is to continue to do what our customers ask of us, and want, and need.
+
+Continuing to stay close to the customers, continuing to stay close to all the new people we hire... Because I don't want to become this sort of "Oh, he's the CEO, and you'd never get to talk to him." That's not who I am. I still interview every person we make an offer to.
+
+**Adam Stacoviak:** Is that right? At what point? Early or late, usually?
+
+**Anurag Goel:** I interview them when we're very close to making an offer, because otherwise I would not get anything done.
+
+**Adam Stacoviak:** Okay.
+
+**Anurag Goel:** \[01:21:54.09\] Yeah. And I don't know how long I will continue --
+
+**Adam Stacoviak:** Is that your own personal role? Where did that --
+
+**Anurag Goel:** Yes, it is.
+
+**Adam Stacoviak:** Okay.
+
+**Anurag Goel:** Yeah. I love doing it because I want to -- it's very personal to me who we bring onto this journey... So I want to talk to them before we bring them on. And I don't go and tell the team to not hire them. There has been maybe a couple of times when I have done that, but I rarely do it. Almost never. But I want to make sure that I know their name before they join the company, that I put a face to the name, so when I see them next time, it's not like "Hey, who are you? Where did you come from?" It's not going to be possible forever, of course, but I love doing it. And I also think it helps them to see - again, going back to my role as CEO, to align the company around where we're going - for them to hear that from me. And I actually do think it helps us in closing the best candidates.
+
+**Adam Stacoviak:** Yeah, for sure. I want to dig in a little bit, because I'm looking at your -- at least what I can tell your org chart based on your About page... I don't see a COO, I don't see a CTO; those two particular roles in particular. I do see VPs of people, product, finance, engineering, and maybe a VP of engineering is the CTO acting role, but they're not literally the CTO. And if you're -- because sometimes, to be a great leader, you have to have great leadership beneath you, and those two roles... Like, a COO is such a critical role to a well-producing CEO. And I would even say the CTO as well. Do you act in those roles, essentially? Do you sort of bridge the gap between what seems to be missing based on just titles?
+
+**Anurag Goel:** That's a great question. So I don't know if we need either of those roles right now. I'm not acting COO, for what it's worth. And I am also not necessarily playing CTO. And I'll tell you -- the reasons are different in each case.
+
+For the COO role, I think typically what ends up happening is marketing, finance, sales... Everything that is not EPD - engineering, product and design - they end up reporting to a COO. And I am very comfortable, and would actually prefer to have the leaders of these functions report directly to me, because I want to be that level of involved in what's going on. And I don't want to create yet another layer between me and the leader of marketing and the leader of sales and the leader of finance. So I enjoy that level of working closely with them.
+
+I also think -- it might not always be true, but I also think that we can attract more senior executives that way, because they're reporting directly to the CEO, and they're part of the day to day executive decision-making. And I haven't felt the need for it. So yes, I'm hiring a marketing leader right now, but then they'll report directly to me, and that's totally fine, and I don't need a COO to do that. So that's how I think about the COO role. I'm not sure it'll be necessary at any point in our future. There are a lot of very successful companies that don't actually have CEOs. So there are different models for running these things.
+
+On the CTO side, I think it does come down to the kind of engineers you hire, and then obviously your VP of engineering, of course... But we have generally hired engineers who have very strong points of view on what technology can do, and what else is out there in the market, and where technology is going, and it's a team effort for us. So even our VP of product is very involved in the technology choices that our customers are making, and understanding those choices. It's fundamentally a very, very, very highly technical company.
+
+\[01:26:24.16\] Our VP of product was previously an engineer at Heroku back in the day, a very long time ago. And then she was at Slack, and she was at Figma, and then she was at Webflow, and then she joined Render. And so she's -- but then our VP of engineering was previously the VP of product engineering at Discord. And so we have these people around the table, and you know, I'm an engineer by trade... So we don't see that CTO-shaped hole yet. And again, there are a lot of companies, at our stage especially, that don't have CTOs. And so we're okay. I don't want to hire people for the sake of it, just to fill out an org chart. And I actually think -- I mean, the VP title is whatever; they could have the CXO title also.
+
+**Adam Stacoviak:** Sure.
+
+**Anurag Goel:** So it's just a thing. And so that's how I think about both of those roles.
+
+**Adam Stacoviak:** Gotcha. I imagine as you keep reinventing -- like you had said, you want to maintain... You're calibrating, to use your words back at you. You're calibrating your involvement in particulars; in particular innovations, products etc. and how you hire, and whatnot. I think that's so cool that you meet with everyone that you're going to give an offer to. I think that's -- it's grounding, is what it is. It's grounding for you, one, to see every person who's coming into the organization, and two, removing that stigma of your potential level of importance to the organization... Which is crucial, obviously, but your accessibility to folks. I think it's probably so rewarding for them and for you to know that "Okay, Anurag gave me time; sat down and just talked with me about what my role is, where this company is going to go, and how thankful he is to have me as a part of this." And just show that -- I imagine that's what you do, right? To show your gratitude.
+
+**Anurag Goel:** Yeah. And people have lots of choices. Great people have lots of choices. And when you want to hire great people, it's more them doing you a favor than anything else.
+
+**Adam Stacoviak:** That's right.
+
+**Anurag Goel:** Yeah. So it just becomes really important for me to know that I can talk to them directly, and that we can have a one-on-one conversation that feels genuine, natural, and honest. And again, we want to hire people who are low ego, no drama, want to do great work, are ambitious, but not like blindly so, and are excited about what we're doing, are passionate about what we're doing here. And all of that shines through in these conversations. And sometimes I have to explain to them what we do, which is also something that, again, keeps me fresh, right? Like, what do we do? I get asked that question several times a week.
+
+**Adam Stacoviak:** Tell them to listen to this podcast. They'll learn, right?
+
+**Anurag Goel:** Yeah, exactly. Exactly.
+
+**Adam Stacoviak:** For the next little bit, this is your introduction to who we are, is listen to this podcast.
+
+**Anurag Goel:** Do you want me to talk about -- sorry.
+
+**Adam Stacoviak:** I mean, I was joking, but I think it's true, because...
+
+**Anurag Goel:** Oh, you're joking. Okay. Yeah, yeah.
+
+**Adam Stacoviak:** \[01:29:54.20\] No, no, no. I was joking but being serious at the same time. I don't know how you describe that... Actually, I'll give you the frame of reference, because a while back I had interviewed Sid Sijbrandij, who's famous for founding GitLab. And he and his team loved the way I conducted the conversation. I wouldn't call it an interview. Obviously, I'm asking you questions, but it's more conversation, you know? You know, I'm not \[unintelligible 01:30:18.28\] And they had their open documentation for when they hired new people... And it was essentially if you wanted to get to know Sid, you would listen to or pay attention to these pieces of content/media that was out there. And I had gotten serious praise from within GitLab about my ability to conduct conversations with Sid over the years. And so that's kind of where that joke, but seriousness came from, was that -- it was sort of insider baseball, that's why I explained it... And honestly, it was an honor to be told that, because like -- they didn't even have to tell me that. I don't know who links up my stuff or links up our stuff, and shares it and says "This is a reference point." But I think we encapsulated your focus on application developers.
+
+So I think clarity would be just simply like "Wow, Render's focus is on app developers. Not DevOps, not this particular discipline... It's application developers, people who are building and shipping the actual application. And if it doesn't serve them, is it truly something we should be doing?" So I think that's why I say that as joke, but seriousness.
+
+**Anurag Goel:** Obviously, I've really enjoyed the conversation so far, and I think you're at the top of your game. You know what you're doing, so I'm not surprised.
+
+**Adam Stacoviak:** I try. It's just talking... It's just talking.
+
+**Anurag Goel:** Yeah. I love -- I was listening to the chat you had with Kurt at Fly, and you're like "Dude, how are you?"
+
+**Adam Stacoviak:** Yeah, man. How's life? I almost pulled that part out.
+
+**Anurag Goel:** No, I thought it was great. I'm glad that you left it in.
+
+**Adam Stacoviak:** Yeah. Well, because it was -- to give some context to those who might be catching up (we'll link it up in the show notes, of course), I asked Kurt how his life is, and he recently went through a divorce that sort of changed his entire life. And he was happy to share that. And so in the moment, I was like -- in my brain, I'm thinking "How deep should we go here? Because this is a deviation from the topic, obviously..." We'd been talking about reinventing the cloud, this new public cloud kind of idea. And now here we are talking about divorce, and I don't want to get melodramatic, and get sad, but I also want to examine this... Because I feel like part of what we do here at Changelog is just – it's peel back the layers of the realness of what is what. Why did you not go and sip Mai Tais on the beach? Why did you go back to work? Why do you care so deeply about interviewing, or sitting down with every person who comes in? Why are you so passionate about application developers? And it's beyond the technology. It's beyond even your drive as a human being. It's like "What makes you tick?" That shared decision between you and your wife to say "You know what? Let's do this." And it's not just "You go do the work." It's us making the choice of allowing one part of the whole to go and do one more - maybe more - adventures; one big adventure, one more big adventure. And I think that's what it's all about, really. You can't just simply say, "Tell me why you think Kubernetes sucks. Let's talk about that for 40 minutes." Or it doesn't suck. Or let's just say, why do you think it sucks for application developers, and it doesn't suck for DevOps. I've got a great friend of mine who is a DevOps, I would just say borderline genius... He loves Kubernetes.
+
+**Anurag Goel:** No, I think Kubernetes is great for a lot of reasons.
+
+**Adam Stacoviak:** Yes, but that's because he's a DevOps engineer. He's not an application developer. Application developers don't really care that much about orchestration. They care about applications in production. So yeah, I appreciate that. I appreciate that.
+
+**Anurag Goel:** And I think that's interesting that you bring up Kubernetes, because another way to think about Render is often Kubernetes, the good parts.
+
+**Adam Stacoviak:** Okay.
+
+**Anurag Goel:** Yeah, because you get the things that --
+
+**Adam Stacoviak:** \[01:34:09.03\] The small book versus the big book.
+
+**Anurag Goel:** Yeah, exactly. The stuff that you actually want from Kubernetes, without ever having to learn Kubernetes... But I think Kubernetes actually is a great piece of technology. Obviously, massive ecosystems, so much value that has been created... Also, just unbearable complexity. And there's all these vendors upon vendors upon vendors that -- I don't know if you've seen this, but the CNCF landscape chart... Yeah, it's massive, and it keeps growing every year.
+
+**Adam Stacoviak:** And it will. I mean, that's just -- I mean, it's entropy. That's how it works for them. I mean, it makes sense to do so.
+
+**Anurag Goel:** Yeah, it makes sense for a certain market. But actually, going back to your interview with Kurt, I'm glad that you left it in, and also that Kurt was open to talking about it... Because I like Kurt. Yeah, Kurt and I have chatted in the past. I like him. I think he's a great human being. And I think, again, we're serving different parts of sort of the same market, but I think it's -- again, like I say, it's big enough for... That's the good thing about the cloud, it's very big, so people can do their own thing and still be successful.
+
+**Adam Stacoviak:** You know, knowing what I know about both of you as individuals, and then also Render and Fly, I would say you're very similar, but also very, very different. Very, very different. When I play with Render versus Fly -- you're solving similar problems for the same developer type, application developers, but you're doing them in uniquely different ways. And that's what the market needs, because where there's push, there's pull. And it's not a choice of Fly versus Render, or either/or. I think there's really just -- it comes down to the kind of developer you are. There's still fractures and fragments within the application developer sliver of the market. You're going to have some who care more deeply about unique things they can abstract away or up into the Fly platform from Linux, and all the different things they're doing there. I think there's so much room in the market. That's why I'll never -- we've never been a... We've tried to resist. I mean, because sometimes you do pick a favorite, but we've really tried to resist picking a favorite. And that's hard. It's really hard when you have that responsibility... Because when you kind of have a favorite, it's like "Well, you know what? I do have a favorite. Can I tell you my favorite?" Maybe I can, maybe I can't.
+
+But we really try to keep an open mind, because similar to your ethos, our ethos is that we want to share what could be and should be given attention to to developers across the globe. So it's not like "Oh, this exists and nothing else exists, so use only that." It's more like "If you're a developer, whatever your challenge is, here's the menu", so to speak, "of what you can use to solve your problems." Because at the end of the day, in our Zulip chat room - which is not Slack anymore. We use Zulip now. Not Slack. I don't know if you know about Zulip.
+
+**Anurag Goel:** Oh, I know the founder... Yeah.
+
+**Adam Stacoviak:** Fantastic. I mean, I always -- I come to that conversation sometimes, I'm like "Do you know what Zulip is? Because if you don't, let me tell you. But if not, now you know." But in our Zulip, we're talking to individuals. We're talking to developers. And we have friends and loyal fans and listeners over many, many, many years. And we're here to serve them, to showcase what's cool, what's important, what's changing, what's not, and to give that option to them, and not pick favorites, really. That's where we're at with that.
+
+**Anurag Goel:** \[01:37:59.00\] That's awesome.
+
+**Adam Stacoviak:** I'm pretty excited to have you on the show, I'm pretty excited about this new raise for you... I'm excited for this legitimacy., that I don't think you necessarily needed... But obviously, a new label, series C, is going to change some people's minds, so to speak.
+
+**Anurag Goel:** Yeah. I remember back in -- I think it was 2020... We emailed some folks who were on Heroku. And we were like "Hey, you should maybe consider Render. And one of them was like "Dude, I'm not going to switch over to this startup."
+
+**Adam Stacoviak:** They were honest with you. Yeah.
+
+**Anurag Goel:** Yeah, he was pretty honest with me, which I appreciated. And I think it's true. I wouldn't do it if I were in their position. I'd be like "I don't know if Render is going to be around. They're just like a series A company", or whatever. And back then we also didn't have all the features that we do... And that's always going to be true; we're always going to have fewer features today than we will two years from now, or even a year from now, or even a month from now. But I think back then we were missing more core things, and we've basically built out all the core stuff now, except S3. We'll go build that.
+
+**Adam Stacoviak:** Coming up.
+
+**Anurag Goel:** Yeah, coming up. I don't know if -- here's the other thing about figuring out what to build and what not to build. The way we think about product is that a no is temporary, and a yes is forever. And you can keep saying no to stuff, but as soon as you launch something, now you're stuck with it. As soon as you launch something that people are using, you're stuck with it.
+
+**Adam Stacoviak:** Good point, yeah.
+
+**Anurag Goel:** Yeah. You have to support it, you have to figure out how to deal with pricing, you have to make sure it works with the rest of your stack, your API surface area, your infrastructure as code, your whatever, your SDKs... You just have to keep doing that work over and over again as soon as you build something. And so we're very careful about what we take on. But I think one of the things that we're going to be able to do this year is take on more interesting stuff that I talked about, and even more than that. And that's what I'm excited about, just allowing people to use Render in new ways, and also continuing to help people scale.
+
+Our managed Postgres has come a long way... So this was another thing that we were pretty far behind Heroku on, and now we're actually better than them, and they're switching over to Aurora, saying "Oh, we're not going to build our own Postgres anymore." It's just like "Alright... Well, that's your call."
+
+And I think for us, it goes back to the experience, right? So if you try to build a wrapper over Aurora, then you're a wrapper over Aurora, and you have to deal with their vagaries. And Aurora is not even Postgres-Postgres. It's Postgres-compatible, but there will be stuff in your function code or your triggers that just won't work the same way. And all the documentation online is for Postgres. It's not necessarily for how Postgres on Aurora works.
+
+And so I think the other thing that we've been lucky about is Heroku has made our job easier. It has helped drive people into our arms. And that was the other thing that happened when they killed their free tier back in 2022... That was when a lot of people first discovered Render. And it was not just that our free tier took off after that, it was also that our revenue took off after Heroku killed off their free tier... Which is just a \[unintelligible 01:41:55.11\] I'd say.
+
+**Adam Stacoviak:** Right place, right time... I think it's prudent for you to showcase that your demeanor might be -- or your point might be "We're the better new Heroku."
+
+**Anurag Goel:** \[01:42:11.13\] Well, but we're not. That's the other thing. Because I don't think about Render as being a better Heroku, or even a better AWS. I think when you think about a company as like a better X, then it constrains the possibilities of what's possible. Only back in the day, Stripe benefited from PayPal's mistakes.
+
+**Adam Stacoviak:** For sure.
+
+**Anurag Goel:** And from the mistakes of everyone else. So I think in some ways the story is similar. But I think with us, people come to us -- people who have used Heroku before and who we know are using Heroku look at us and they want details on how Render might be better. And so maybe that's what we were trying here. But surprisingly -- I was surprised by this. So when we ask people where they're moving from - and they can choose to not tell us, but if they do tell us, when we look at where people are moving from, more people move to Render from AWS than from Heroku. Yeah. And many more people move to us from AWS than from Heroku... Which to me says we're doing something right, because Render is not the next Heroku. Render is actually a replacement for people for both Heroku and AWS. And for a certain category of people, application engineers, like you were saying, like I was saying, who want to get started quickly, but also scale, and not have to worry about whether they need to spin up Kubernetes... Like, that's who Render is great for. And in the future, there will still be people who are doing DevOps. And there will still be people who are running stuff on bare metal, as there are today.
+
+**Adam Stacoviak:** Preach, yeah.
+
+**Anurag Goel:** In fact, there are people who've gone back to bare metal, as we all know...
+
+**Adam Stacoviak:** We all know that, yeah. We didn't talk about that, but yes, for sure.
+
+**Anurag Goel:** Exactly.
+
+**Adam Stacoviak:** Anti-cloud.
+
+**Anurag Goel:** Yeah, there's this whole thing. I mean, it's a bit -- I think DHH does a great job marketing it. It's not clear to me if the big move is happening, because the clouds keep making more money every quarter.
+
+**Adam Stacoviak:** Yeah. Well, I think just what you said before, which is there's room for that, too.
+
+**Anurag Goel:** Yeah, absolutely.
+
+**Adam Stacoviak:** There's room for bare metal, and there's room for those who are like "You know what, it's not so much anti-cloud, it's more like you're a build versus buy. We could, so we should." Right? You were able to, you had the talent, you had the control, it benefited you to have the control, so you did. Same thing here. I think it's less -- but I think DHH's persona is countercultural. It is provocative, in a way; in a lot of ways. So I think it's natural for his opinions, because he holds his opinions very hard. He's hard on his opinions. And he's very articulate and clear, which I love about him. I mean, whether you like his opinions or not, that's a fact I love about him, which is no matter if I like his opinion or not, he will have clarity and conviction behind his opinion.
+
+**Anurag Goel:** Yes, he's incredibly smart.
+
+**Adam Stacoviak:** Right. And so I think just the same, there's room for those who are like "You know what? I can put this server rack in co-location, and we can do this, and all this different stuff..." And that might serve them and their purpose for then, and for whatever. But I think you're building for a different application developer that thrives in speed and autonomy, not in standing up servers and managing servers, whether it's easy or not.
+
+**Anurag Goel:** Yeah. And we use -- behind the scenes, we use both AWS and GCP. And that doesn't mean that that is what we're always going to use. In fact, again, this year, that might change.
+
+**Adam Stacoviak:** \[01:46:07.20\] Is that right? Some bare metal?
+
+**Anurag Goel:** So we're not going to get rid of AWS or GCP, and we're not going to ask customers who are on AWS or GCP to move. In fact, we'll still continue to add people there. But I think now it's incumbent on us as a business to also look at lower cost compute, whether that's bare metal, or some dedicated hosting in Hetzner. I've heard a lot of scary stories about Hetzner servers going down, but a lot of people use them, and they're cheap, if nothing else.
+
+So I think we need to explore our options, because it's not clear to me that AWS is the best long term partner for us. And I do think that AWS is a better partner than GCP, though. And we can talk about that some other time.
+
+**Adam Stacoviak:** Yeah. Well, I think the way to get there is really "What do application developers want?"
+
+**Anurag Goel:** Yes, that is the mantra.
+
+**Adam Stacoviak:** Right? If they want a diversity of choice - well, then you're going to offer them a diversity of choice. If they want a diversity of choice because of cost - well, then you're going to offer a diversity of choice because of cost.
+
+**Anurag Goel:** Exactly.
+
+**Adam Stacoviak:** And if that doesn't make sense, then you won't do it. Because if your mission is to serve app developers, then you're going to figure out what the 80% of the 100% of the market that's an app developer wants... And Render is, I assume, attempting to solve for 80% of the challenge that app developers have, right? Not 100%, because that's not Pareto's principle, right? 80/20.
+
+**Anurag Goel:** Yeah. And I think people on Render will always -- they should always be able to use specialized compute or specialized services on other cloud providers, right? And so we actually -- if you have some compute running on AWS, then Render can set a private link from Render to AWS, so you can always connect your services on AWS securely and privately. And a lot of our customers really appreciate that.
+
+**Adam Stacoviak:** Yeah, for sure. Well, I'm excited for you. I'm excited that we've got you on the show, so we can check that box... I'm excited that potentially this could become the new -- if you want to learn about Anurag and what Render is doing in the next year or so, this could be a new link for you to share with folks.
+
+**Anurag Goel:** Oh, for sure.
+
+**Adam Stacoviak:** What else is left? What else can we share in terms of anything we haven't covered, that you want to?
+
+**Anurag Goel:** No, I think this was pretty far-ranging. I enjoyed the conversation. There was a lot of switching between topics and things, but that's what a conversation is. It's not a straight line.
+
+**Adam Stacoviak:** We'll get you back on. We'll make you a staple. Once a year, let's say. Something like that. Let's get you back on.
+
+**Anurag Goel:** I would love that.
+
+**Adam Stacoviak:** I'm a fan of what you've done, I'm a fan of innovating for developers, and I think if that's what you're doing, which I think you are, I'm happy you're doing that. And stoked.
+
+**Anurag Goel:** Thank you.
+
+**Adam Stacoviak:** Stoked for Render.
+
+**Anurag Goel:** Thank you, Adam. I really appreciate it.
+
+**Adam Stacoviak:** Alright, Anurag. Thank you.

Chasing that next BIG thing (Interview)_transcript.txt

@@ -0,0 +1,567 @@
+**Adam Stacoviak:** Well, Drew Wilson is back in the house... It's been a while since you've been here.
+
+**Drew Wilson:** It has.
+
+**Adam Stacoviak:** I want to say -- gosh, I didn't even do my math here. Let me go search quickly, Drew Wilson...
+
+**Drew Wilson:** I think maybe 15 years, 12 years... I don't know, something like that.
+
+**Adam Stacoviak:** Okay, so we had you on Founders Talk a couple of times. Founders Talk was a show that I did solo, where I talked to founders, makers, builders, things like that... And it was Valio. Remember this name, Valio?
+
+**Drew Wilson:** Yeah. ValioCon, man.
+
+**Adam Stacoviak:** And part one and part two, for whatever reason. That was the name of the show. It was Drew Wilson / Valio - Part 1 and Valio - Part two.
+
+**Drew Wilson:** Oh, wow. Maybe we're talking pre-ValioCon, and we're talking back when I was making stuff under the name Valio, which would have been like 2009 or something.
+
+**Adam Stacoviak:** You know, I don't know. ValioCon was mentioned... You were at \[unintelligible 00:06:15.07\]
+
+**Drew Wilson:** Oh, dang. So that's probably like 2011.
+
+**Adam Stacoviak:** That's 2011, man. Yeah, that's right. Exactly. 2011. Pegged it. September 2011.
+
+**Drew Wilson:** Wow.
+
+**Adam Stacoviak:** Okay, so that was that show. That was Founders Talk. I think we had you on this show too at some point, but I don't know when.
+
+**Drew Wilson:** Yeah, I've been on here before, and I think -- it was a while ago. It was a couple years after that one, for sure. So maybe 2014 or 2013, something like that.
+
+**Adam Stacoviak:** Well, listen, I should have did my better research on when Drew was last on this podcast. It was ages ago. It was basically the beginning, it was basically the earliest of days... And it's been a lifetime. I think I literally haven't talked to you in a decade.
+
+**Drew Wilson:** It has been a while, man.
+
+**Adam Stacoviak:** At least a decade. I would say at least a decade.
+
+**Drew Wilson:** I think maybe one or two tweets, or something like that...
+
+**Adam Stacoviak:** Yeah, here and there. Nothing special. You got busy, though.
+
+**Drew Wilson:** Yeah, I got busy.
+
+**Adam Stacoviak:** You've always been the busiest person I know of, but you got really busy.
+
+**Drew Wilson:** Yeah, yeah.
+
+**Adam Stacoviak:** Like, you built a company and sold it busy.
+
+**Drew Wilson:** Yeah, yeah. That was fun.
+
+**Adam Stacoviak:** How does that feel?
+
+**Drew Wilson:** It was -- dude, it was incredible. It's like one of those -- I mean, it's not to the scale of "You're a billionaire now. You never have to work again." But it was one of those, like, you come up with an idea, you build it yourself, you're able to build a little company around it, you raise some money, you do the whole investor thing, and then you sell it and you make money. And that's kind of like the dream, right? It's not necessarily like noteworthy scale size dream, but it has the blueprint.
+
+**Adam Stacoviak:** Noteworthy scale size... Well, I think you may say that because you've always been a big dreamer. And it's easy, to some degree, when you're in that position to downplay your accomplishments. Would you agree with that?
+
+**Drew Wilson:** I think so. Yeah. I think the way I see success is different than the way maybe my extended family sees success. And that's kind of like how I can gauge regular people versus people in tech. The things you're used to in regular people land in terms of success is way smaller than what folks in tech are used to.
+
+**Adam Stacoviak:** I don't even think you were a dad. When was your kid born? What year?
+
+**Drew Wilson:** My first was 2010. I have three kids. So 2010, 2012, and 2013. Yeah.
+
+**Adam Stacoviak:** Okay, so you were a dad, barely. Did you hide it? Was it a secret birth?
+
+**Drew Wilson:** No, no, no.
+
+**Adam Stacoviak:** \[laughs\] It's a joke, obviously...
+
+**Drew Wilson:** Yeah, yeah. We've just never wanted to like put our kids on social media. So up until recently, I think I had some tweets like earlier this year -- or no, I guess later last year, about like "Hey, I want to start recording videos for fun, of my personal life and doing stuff." And so I did some, but then I just got too busy to keep doing it. But yeah, it's always been like a conscious thing to like not post about our family or our kids or anything like that. We're weirdos.
+
+**Adam Stacoviak:** I see you have a famous or an infamous Apple computer behind you... What is that computer?
+
+**Drew Wilson:** Yes. So this is the Macintosh Portable. It's got a track pad and a little click button, and it fully, fully, fully works.
+
+**Adam Stacoviak:** Wow.
+
+**Drew Wilson:** \[09:52\] I got that thing on eBay. You can boot it up. It came with a -- it's a 16-pound laptop. So this is a first ever laptop, and look at that. They look almost identical to the way they do today. So they got it right on their first try.
+
+**Adam Stacoviak:** Look at that...
+
+**Drew Wilson:** Yeah. It's 16 pounds, okay? And it had a lead acid battery in it, a car battery, for reals. And so that sort of tech doesn't even last to today. No matter what you do, that sort of tech will break down. And so you can't actually have a working battery in one of these. It's impossible. And so this one, the modification that was made on it was putting in a lithium-ion battery and a 3D printing case to make it the same size. So that's what this has, and that's why it can work. And even then, the little pieces of metal that conduct electricity - and I'm forgetting the name - they don't work so well.
+
+So you've kind of always got to leave it plugged in, running just off battery. Even though it's fully charged, it will only last like a minute or two. But you don't want to be carrying that thing around. It's way too heavy.
+
+**Adam Stacoviak:** Does that run Figma, or the web, or what does it run? Anything?
+
+**Drew Wilson:** It does. It does. It actually even has Wi-Fi.
+
+**Adam Stacoviak:** What?!
+
+**Drew Wilson:** That was like one of the first computers with Wi-Fi.
+
+**Adam Stacoviak:** Oh, my gosh. What spec?
+
+**Drew Wilson:** Wi-Fi 1.0, I think.
+
+**Adam Stacoviak:** Right. The earliest of Wi-Fi's...
+
+**Drew Wilson:** Yeah. It's wild. It doesn't -- I don't know if it can connect to my Wi-Fi. Did I try it? I can't remember if I tried it. But I've got to, at some point. I have this one, I have like an old Apple II, and I have like the old, like the square tall Macintosh... Yeah, it's fun. I like collecting older computers.
+
+**Adam Stacoviak:** You know, off camera I have an Apple II I think it is, or a Mac classic...
+
+**Drew Wilson:** Yeah.
+
+**Adam Stacoviak:** I think it's an Apple II, if I recall correctly. It's just a prop. It's just for - I call it art. It's just art. It doesn't work.
+
+**Drew Wilson:** Yeah. Apple II is the flat piece with no monitor. And then you had to get an individual monitor. The classic is the one that's like the tall \[unintelligible 00:12:03.18\]
+
+**Adam Stacoviak:** Yeah. Okay, so it's a Classic, not a two. I'm looking at the front of it and it just says Macintosh Classic on it. It's got a little floppy drive on there, and whatnot.
+
+**Drew Wilson:** Yeah.
+
+**Adam Stacoviak:** You know, all the good stuff. It's got a keyboard and a mouse, but I don't even keep those things. It's just there, over there, just being art, basically. Just looking like trash, basically. As a matter of fact - not that we should go on this rant much further... On the left - and I'm pointing over here off camera - is the Macintosh Classic. On the right is the 2013... What do you think it is?
+
+**Drew Wilson:** 2013?
+
+**Adam Stacoviak:** The trash can.
+
+**Drew Wilson:** Oh, wow. Yeah.
+
+**Adam Stacoviak:** The trash can, right next to it. It's kind of funny, because you see this -- I suppose maybe the undercurrent to this conversation could be iteration. It's this iteration of monitor, compute, floppy, keyboard or external devices... Same idea pretty much, except for the monitor, with the 2013 Macintosh trash can, which was the Mac Pro... And it's like the same size, basically. It's basically the same size.
+
+**Drew Wilson:** Yeah. Yeah. And just a very different form factor.
+
+**Adam Stacoviak:** Very different.
+
+**Drew Wilson:** Yeah, it's wild too, like - this laptop is huge, but if you take the little back portion out of consideration... Because that's where a lot of the larger motherboard and everything is, and the battery mainly... And you just collapse that down and you just take the clamshell piece, it's essentially the same size as a modern day laptop. It's a little bit thicker, but it's not like insanely thicker.
+
+**Adam Stacoviak:** Yeah.
+
+**Drew Wilson:** So it's wild how \[unintelligible 00:13:49.08\] stay the same.
+
+**Adam Stacoviak:** I heard recently that it was Steve Jobs that actually coined the term laptop. Do you know about that?
+
+**Drew Wilson:** I don't know, but I do know that that term came after this... Because this is called a Portable. Macintosh Portable. They even have a commercial... You can look it up on YouTube, you can find the Macintosh Portable commercial, and they've got like this guy in a business suit, 1980s looking, and he's got this massive case - which I have the case that came with it - carrying it around. I'm like "Yeah, right. No one's ever going to do that." And they didn't. That's why these things didn't sell well.
+
+**Adam Stacoviak:** \[14:26\] And they didn't. \[laughter\] Oh, man...
+
+**Drew Wilson:** Because they're so heavy.
+
+**Adam Stacoviak:** I missed you, man. You know, what you have is this -- you have this unusual take, I suppose, on technology... Like you just said there, "And no one's ever going to do that. And they never did." Like, you have this sort of like uncanny instant intuition of what will and what will not persevere through the tech space. And maybe I'm pushing that boundary a little too far, but that's my perception.
+
+**Drew Wilson:** Crypto... Crypto... Crypto won't.
+
+**Adam Stacoviak:** You've got a pretty good intuition.
+
+**Drew Wilson:** \[laughs\]
+
+**Adam Stacoviak:** What'd you say?
+
+**Drew Wilson:** I'm trying to be all hot-takey. I'm saying "Crypto won't."
+
+**Adam Stacoviak:** Oh, yes. Crypto won't. Are you against crypto?
+
+**Drew Wilson:** I'm not against crypto, I'm just against the idea that crypto is the savior of the world, which I think is now falling out of mainstream think, which is cool... Because -- I mean, the technology existed before, it's called SMTP. It's called email, distributed servers serving up the same set of data. And now it's just on a different protocol. And it was never used for any -- SMTP was never used for anything beyond email, and crypto will never be used for anything beyond collectibles and scams. That's where it fits. And this idea -- I always thought it was weird, this idea of like a distributed database being such a big deal. I'm like "Why is it such a big deal?" Like, every company you look at that has done anything great owns the technology, and it's all centralized. No one wants to go to a distributed bank. I'm sorry, but they don't. No one wants to go to a distributed almost anything. They want the centralized version. They want the company that's there, who's incentivized to make money, and who will build up a scale of employees to service you, the customer. That's what they want. They don't want a distributed do-it-yourself sort of thing. I mean, there are people that do, but it's less than a fraction of a percent.
+
+**Adam Stacoviak:** Yeah. What are some hard beliefs that you would say you have?
+
+**Drew Wilson:** Crypto will be great for collectibles, and hopefully also another spot, like through Bitcoin, the name brand crypto, will be another spot for super-wealthy people to move money, and so that way they don't have to move it into real estate, and harm the real estate industry... In that by doing so, by buying up a bunch of properties, it makes property go up everywhere; property values go up everywhere, and it makes it harder for regular people to buy homes at an affordable price... That's one, if we're staying on the crypto train. If we're not staying on the crypto train, another hard belief - let's see... Maybe a hard observation, how about that? I still think it's interesting that design and engineering are now a regular job. I was always the young guy in tech, and so that will never kind of go away from me. But this idea of getting into tech and it being a job, and not like a hobby and something that you want to do is still so strange to me... Especially when you get into running companies and having employees or having teams, and you start to realize that "Okay, this person does one thing, and they're probably only ever going to do one thing." And maybe they have aspirations of being really good at this one thing, but they don't really want to spread out, and like think of tech as one thing, and there's many aspects to it. They think of "I'm a designer" or "I'm an engineer" or "I'm a product manager." So that's the segue into my hard belief. Here comes my hard belief.
+
+**Adam Stacoviak:** \[18:15\] Your hard observation.
+
+**Drew Wilson:** That was my hard observation, so now my hard belief...
+
+**Adam Stacoviak:** Okay. I'm down. Take me there.
+
+**Drew Wilson:** I don't know how to phrase this, so I'm just going to think of how I could phrase it... It's something that I'd always thought about, but I've never had to put it into a concise set of words. So I think product managers who are not really good at sales, or really good at engineering, or really good at design are useless in tech. It is so disheartening to be on a team where there's a product manager - and that has nothing to do with them personally. It's just, it brings the capabilities of the team so far down when you have a product manager who's never been an engineer, never been a designer, never been good at sales... What the crap are you good at? Why are you at a tech company? What are you doing here running this --what are you doing running this department? You have nothing to contribute except ideas. But guess what? Everyone here has got ideas, and they're vastly more informed than your ideas. So what the crap are you doing here? And why are CEOs allowing this to take place? Oh, my gosh. Like, if you're going to be someone in product, you should know product. And that doesn't mean "I have a good idea and a good intuition on what users want." If you do have a good intuition on what users want, you should be super-good at sales, and you should have some experience there, and you can bring that to the table. Because for the most part, we're all building something that at some point has to be sold.
+
+**Adam Stacoviak:** Right.
+
+**Drew Wilson:** But if you're not that, you're not an engineer. You can't understand how to build something. And you're not a designer, and you can't understand how to build something in that aspect... It's like, what are you doing here? I would say most product managers fit that mold of not being able to do any of those three. And I think that's why you see these large layoffs in product. Those people are usually the first to go, because they don't really have any skills to contribute other than organization. And if you're not even that good at organization, which a lot of them aren't... That's why some people come over, they take over a team, they're like "What the crap? Where's all the documentation of what you guys are doing, or where you're going, or how this works?" No one ever made it. So what was the purpose of this product manager, besides receiving a paycheck for doing almost nothing?
+
+**Adam Stacoviak:** Wow... Okay, so --
+
+**Drew Wilson:** There's my hard belief. \[laughs\]
+
+**Adam Stacoviak:** ...the way you're describing this makes me feel as if you've been, you know, bitten twice, let's just say. Have you been bitten twice by this product manager who has no skills bug?
+
+**Drew Wilson:** I haven't really been bitten, in that I haven't been on a team where I had to be beholden to somebody like that. It's just that I've been in companies that have that. So when I sold my last -- not my last startup; one of my last startups to GoDaddy... I was a part of GoDaddy, which is a great organization, great people... But they are a 10,000 plus person company. They're a public company. They have a lot of these product managers that - they don't have any skills beyond just thinking they have skills. That's their skill, is they think they have skills, and they're able to convince themselves and an employer that they do. So you look at what these other teams are doing and you're like "My gosh, you're just running in circles. You're never going to \[unintelligible 00:21:36.19\] I came into GoDaddy as an acquisition, and we moved our tech Plasso, which was a payments platform, into GoDaddy, and it became their payments platform. It became the way that folks transact using GoDaddy's websites. And when I did that, we integrated in six months, and I won the Getting S\*\*t Done Award at GoDaddy. They have an award that they give out, which is really cool.
+
+**Adam Stacoviak:** Hah!
+
+**Drew Wilson:** \[22:08\] And it was because I did it in six months and we've never had an acquisition happen in less than three years, or something. I'm like \[unintelligible 00:22:13.08\] And then I found out why. Product leads that company. Every product person there is essentially the mini CEO of whatever they're working on. And that is the standard tech model, and that is how most tech companies work. And the problem with that is if you have product people that really don't have any idea how to build a product, you're going to turn out crap.
+
+**Adam Stacoviak:** Yeah...
+
+**Drew Wilson:** I think you see that happen. There are also other companies that I've been a part of that suffer from the same things. It's not just them.
+
+**Break**: \[22:52\]
+
+**Adam Stacoviak:** So we have some history, and I'm happy to keep talking as if it's just you and I. And that's cool with me. But it is not just you and I here, Drew... There are people listening. You know that, right? This is a podcast. It's a podcast, man.
+
+**Drew Wilson:** Yeah, let's take a look at these people. \[laughs\]
+
+**Adam Stacoviak:** Just because - and I never asked anybody to do this, but just because I haven't talked to you in 10 years, you haven't been on the podcast in a decade, we have some history... I know who you are, but just give a slight primer, like maybe a two-minute TL;DR on Drew Wilson.
+
+**Drew Wilson:** Hey... Alright, so I'm pretty early to computers and the web, built my first website when I was 13 in '96, taught myself how to design on Photoshop too, and then never wanted to, but didn't know any engineers, so needed to learn how to code in order to build the things I was designing, so I taught myself how to do PHP in 2000, and built my first backend code and database back then. And then loved it. And so I've always been a designer and engineer working on the web. I was early on the iOS and Mac app stores, built apps for Apple natively... Not at Apple, just as an individual. I had like a number one best-selling Mac app. It was called Screeny, and it did screen recording before you could do that on an Apple device using QuickTime. I built these icons called Pictos, which tons of people still use... It's still the default icon as the shopping icon in Shopify. That went off like hotcakes. And then I built a couple of different e-commerce platforms, one of which was called Plasso. And like I said, I sold that to GoDaddy in 2018.
+
+I also have a conference that I run called ValioCon, valiocon.com. I ran that for six years, bringing it back again this year, later in September... It's been seven years since we've done one, but that was always fun. A conference for designers and engineers.
+
+And then after Plasso, I wanted to take on something that I thought would be a really tough challenge, and it really was, and that was a bank. So I started a bank and ran a bank for four years called Letter, letter.co. Got into YC with that one, raised money, did the whole thing, partnered with a billionaire, lent out millions of dollars to folks. It was great, it was fun, learned a lot. I ultimately couldn't make it work, so I had to shut that thing down.
+
+Then ended up at AngelList for a little bit, and then left there and jumped over to Clerk, clerk.com, which is where I'm at now, running the upcoming commerce and billing product that we'll be launching. And then also working on some stuff on the side, too.
+
+**Adam Stacoviak:** A lot to cover in there. What I've always appreciated about your history was that you were - driven, of course, but self-driven and self-taught. Learning to design, designing things, and then getting frustrated not having people around to - or not wanting to have people around to build the things that you're designing... And then being like "I'm going to figure out this PHP stuff." And you just figured it out. Next thing you're building stuff, you're building iOS stuff... And you've always had this -- it's almost as if you've got this Drew Wilson style, sort of like stuck in there. And whatever comes out, it just oozes this certain feel, this certain -- even when I go to valiocon.com right now, this feels like a Drew Wilson design. This doesn't feel like -- the lettering is cool... Did you do this lettering?
+
+**Drew Wilson:** Yeah.
+
+**Adam Stacoviak:** I mean, everything is -- how do you just do everything, Drew? How are you a renaissance person?
+
+**Drew Wilson:** Well, you just -- I don't know, it's many, many years. It's like the overnight success. Well, it was actually like 7,000 years in the making. That's my case, right? I've been literally working in Photoshop since 1994. I don't know, something like that. Before there was layers.
+
+**Adam Stacoviak:** \[unintelligible 00:29:28.16\]
+
+**Drew Wilson:** Before there was layers.
+
+**Adam Stacoviak:** Oh, before there was layers... Oh, my gosh.
+
+**Drew Wilson:** Yeah. Can you imagine that? \[laughs\]
+
+**Adam Stacoviak:** No, I cannot imagine that. Because I don't think I've ever worked in Photoshop before layers... So I've only ever known layers, Drew.
+
+**Drew Wilson:** Yeah.
+
+**Adam Stacoviak:** You're blowing my mind here.
+
+**Drew Wilson:** \[29:48\] So it's just like a bunch of years -- for example, I wanted to start a bank, and I didn't know anything about banking. I didn't know anything about finance. At that point, I think I had just started investing in stocks. I didn't really -- can you imagine that? In 2018 I was barely aware of stocks. \[laughs\] So I started a bank, and I was like "Well, I guess I'm going to be a banker." And I was like "I guess I need to go for these--" There's these accreditation tests, and these different certificates you can get. So I started studying for that, thinking I needed this, but it turns out "Oh, just one person on my team needs to have that." My CFO already had that, so I was like "Okay, I don't have to go take this test." My co-founder and I were all set to go down and take this test, that we just like crammed for in like a week. Normally, people spend like months on this thing. We just went through it... Crazy. We're going to go take the test. But when was this? This was right around COVID. And so the testing -- the government testing facility was shut down because of COVID. They're like "We're not going to open for like months." I'm like "Oh, crap." And that's when I found out I actually don't have to have this certificate.
+
+But anyways, so then I learned a ton about banking in general. I just went on LinkedIn when I knew I wanted to start this company, and I hit up all the banking CEOs in the San Diego area. And one of them got back to me about like "Hey, can we get lunch? You don't know who I am? Would you just get lunch with a stranger?" And he's like "Sure." \[laughs\] So we went and had lunch for like two hours and chatted about all things, and I learned a ton about banking through him... And then we just went around to different banks around the country, and tried to find a partner. We ended up finding one through a banking platform. Went through like three different banking platforms, helped a new banking platform, it's called Una.co. We were like their first big customer. Built a lot of security stuff ourselves, because they didn't have any at the time. They modeled it after what we did. And yeah, now I know a ton about banking, in case that's ever useful. And about finance. So that's how it happens. You just go on these tangents, here and there, and the other place.
+
+**Adam Stacoviak:** I'm not sure that's how it happens, Drew. I'm really just not sure that's how it happens. \[laughter\] Okay? Let me clue you into something that's a truth; that I believe is a truth. I think these things happen to you. I don't think these things happen to other people, generally. Can they? Yes. Is it possible? Yes. I'm not denying that. But I think only someone like you - and what I know of you at least - is to say, "You know what? I've got an idea for wealth management for high net worth owners, folks out there... And I've got this idea for a bank for them. And I'm just going to start designing screens, and flows", and whatever it is. That only happens to you, Drew. And only to you do you have to learn banking, having never even touched or knew much about stocks in 2018, to then create Letter. Even if defunct, I'm really curious about this - what did you have to learn about banking to make Letter possible? Where did you begin?
+
+**Drew Wilson:** It blows, dude. It sucks. Banking sucks, big time. Fintech sucks, big time. Fintech seems like this place that would be an absolute gold mine. But it just isn't. It just absolutely is not. The way that banking is set up is that - in the US and in most places, it's a government-regulated business type. And the reason they regulate it - and banking super-heavily - is because they cannot afford to have banks go under. You can't just fail. Like, a startup can just fail and no one gives a crap. Maybe some investor who overextended themselves on your round will care. But nobody else will care. A bank can't fail. And it costs a lot of money, to a lot of different entities if a bank fails. And so they highly regulate what you can do, what products you can offer, what prices you can offer them at, what sorts of verticals you can go into so that you will not fail. As a bank, you can't just be like "Oh, wouldn't it be cool if we could offer this sort of loan?" and just offer it. No. You have to get approval for that kind of stuff.
+
+**Adam Stacoviak:** \[34:20\] Really?
+
+**Drew Wilson:** Yeah.
+
+**Adam Stacoviak:** How do you get approval for -- give me an idea for a loan you may have had; this idea for a loan. Did you have this idea? Was this your thing, or...?
+
+**Drew Wilson:** Yeah. A new type of lending product, which would originate loans. To do a new type of lending product - if you're a federally chartered bank, meaning you can operate in any state in the US, the OCC is going to be your regulator, and you have to go through them. Now, this is if you're a bank. To get to be a bank is a huge deal. The US stopped issuing bank licenses from the year 2007 to the year 2017 because of the banking crisis. And a bunch of the little banks got bought up by larger banks. Just to give you an idea, your average mom and pop bank, which barely exists anymore... I mean, they're technically considered, financially considered a small business. They're making like a million dollars a year. It's nothing big. Yet you would think, "Oh, it's a bank. They're making tons of money." No, they're not. If they're one of the big ones, they absolutely are. Because with banking, you can't make big revenues because you're regulated against this idea of loan sharking or charging too much money. You literally cannot make a lot of money in banking unless you have an enormous amount of customers. That's just how it works. And it will never change.
+
+**Adam Stacoviak:** What's enormous?
+
+**Drew Wilson:** Millions and millions of customers. We're talking like Chase size, Wells Fargo size, or some of the new startups that have gotten there, who can now become profitable. What's the one with the gold card? Robinhood's getting into banking... And what's the other R one? Revolut. Even Monzo, who was doing really well in the UK, because they hand out banking licenses called these e-bank licenses really easily, was doing really good over there. Couldn't make it in the US. These regulations are just too stiff. And you've got to have way more money than what they have.
+
+So banking is a very difficult... I mean, even Silicon Valley Bank failed. Banking is a difficult industry to be in, especially the one I jumped in. I jumped in thinking "Wow, there's nothing like new and modern for high net worth individuals. Let's do something there." Well, it turns out if you want to get into that space, you have to go through money managers, wealth managers, because the wealthy people don't give a crap about where their money is. They just want their manager to deal with it. Like "Yeah, you deal with it. You deal with it. You figure out what's best, I don't really care." And so it's very difficult to do a B2C play in that arena.
+
+**Adam Stacoviak:** Yeah, definitely challenging. So what made you -- where were you at, I suppose, in your life to make Letter even an idea? Not even possible, but an idea.
+
+**Drew Wilson:** Yeah. So what happened to me was I had sold my company, Plasso, to GoDaddy. And it was literally a dream scenario. An all cash deal, with the option to jump on as an employee, and get a good paycheck and additional payouts if I stay on. So I stayed on there for two years. But with that initial cash sale, I paid back my investors plus some, because they made money on the transaction... And then paid some of my employees that had equity. I made sure everyone who worked with me had equity, so folks made money. And then I had money. And I had never had money before. As a matter of fact, in order to get Plasso to where it was, I had to take out a bunch of loans on the same day, so they would not know about each other, and they wouldn't show up on the FICO score thing, the credit report. So I took out a bunch of loans on the same day just to keep the company going.
+
+\[38:24\] I was making enough for myself, but that's it. And I needed these other employees to help sell the thing. And so I took on massive debt. The amount of debt that would take me my entire lifetime to get out of if things didn't work out. And so yeah, I took on that debt... But then, just as I had to fire my last employee, being like "Sorry, I don't have any money", literally, that was on a Friday, that was their last day... I came into the office on a Monday, and there's an email in my inbox from GoDaddy saying "Hey, we'd love to talk about partnering." And acquisitions always happen under the guise of partnerships. I didn't know that. And then three months later, I had sold my company. A little over three months. But to me it was like 10 years of my life. It was just so exciting, and at the same time -- it's like a rollercoaster of emotions, because things are going your way, not going your way, is it going to happen, it's not going to happen... It's going to happen... Oh, man.
+
+And I didn't have anybody else. It was just me. I was the only person doing negotiations. There was no advice giver on my end. No lawyers, nothing. It was just me doing it all. And so it was a great experience. It was a great experience. The company acquisition team at GoDaddy is super-nice.
+
+So it was cool, got through it all, made some money... Then I was like "Well, I've never had money before. What the crap do I do with money?" And I'm like "There's got to be a site somewhere --"
+
+**Adam Stacoviak:** "I've never had money before..." \[laughter\]
+
+**Drew Wilson:** I was like "There's got to be a site that says what to do." I'm looking around... I'm like "Oh, there's no site." The closest you'll get is The Points Guy kind of stuff, but that's a totally different scale. And so I'm like "What the heck do I do? I have no idea." I'm like "I'm sure other people are in this situation. Maybe I should build an app for them, to help them manage their money and have this advisor here with you all the time, that you don't have to interact with personally." Because no one wants to pick up the phone and call someone. So that was kind of the idea behind Letter, and that's how I arrived at that idea. And I was like "I should make a bank. No one's making banks. I could make a bank."
+
+**Adam Stacoviak:** "I could make a bank..."
+
+**Drew Wilson:** "There's a bunch of old people making banks... I could do it."
+
+**Adam Stacoviak:** Is this literally right after the acquisition time frame? The three months are over, the acquisition goes through, all cash, you're an employee, then Letter? Or is it like a year down the line? Help me understand that compressed timeline bit.
+
+**Drew Wilson:** It's all at the same time.
+
+**Adam Stacoviak:** Really?
+
+**Drew Wilson:** So I could be very optimistic... As an example, when I was going through the transaction, this was not a done deal. But in my mind, I'm like "I could see this totally being a done deal." And we're still like a month and a half out of what would be the end date. The thing is, you don't start a process, like an acquisition process, knowing when the end date is going to be. And it's not till like a week or so out that you know like "Okay, this is when it's happening." So this whole time it just could go either way. But I'm like "Yeah, it's gonna go my way." And so I'm looking at this house that I live in now, and I'm like... \[laughs\] I have no money, right? I have no money. And my wife and I are going around, we're looking at homes, like "Yeah, let's take this one." This one's like a very expensive house. And they're like "Okay, cool, if you want this house, put in the application" or whatever. So I put it in and they're like "We need a good faith deposit." It's typical in America to put down a little bit of money... And the deposit they wanted, I just flat out didn't have. So I'm like -- they wanted, basically, I don't remember, it was like $20,000 or $40,000. And I'm like "Will you take $1,000?" \[laughter\] I'm like "I promise I have the money to buy this house." And I have to show them financial statements, and it shows no money.
+
+**Adam Stacoviak:** \[laughs\]
+
+**Drew Wilson:** \[42:22\] And "I promise, I promise it's there." I have no idea why they just were like "Okay, sure." But they took the thousand bucks, and they took that, and I was able to get the house. And it came down to the point where, "Okay, now you've got to pay for the house." And I'm like "I just need like another week." And they're like "Okay." And that week's coming, and I emailed them like "I know I said another week, but I'm going to need like one more week." And they're like "Okay, fine..." And then that week happens, and literally like a day or two before I needed the money, I got the money myself, and I was able to give it to them. So that's how optimistic I can be.
+
+I'm trying to remember your original question... But I think it was around - yeah, Letter, when I started. So that kind of shows you my mindset. So I got the house, I'm an employee of the company now, I'm working at GoDaddy... I think it was probably six months or so before I came up with the idea for Letter. And I wanted to just chill out and not like work on anything other than this one thing I was doing. And so I got through that initial launch, and it took me like six months to get the product out the door at GoDaddy. We got that out, and I think it was right around then when I was talking to another buddy of mine who worked at GoDaddy, that I'd met there, who had sold his company to GoDaddy the year before me... And - yeah, I was just thinking more about finance and about what I should do with the money, and there was really nowhere to turn. And I didn't want some 20-year-old kid working at Wells Fargo deciding what to do with my money, who's never had money himself... No. So I had a financial advisor who was my CFO, who ended up being my CFO at the bank... And even still, it's like, this is so manual. Everything's a phone call. Everything takes forever. It's like, there's got to be a faster way. So that's how I came up with the idea for Letter. And what's funny is I started the company, and Letter ended up -- originally, we were going to be banking as a platform.
+
+So like purely a code thing. And then we turned into this more like branded bank sort of thing. And I applied to YC. It took me four tries to get in. YC calls me and is like "Hey, we think we want you in, but we're not sure, so we're going to have another partner talk to you. Are you available in the next two minutes?" "Yes." They call me. I pitch them. And they're like "Okay, okay, okay. I'm not sure." And then like 10 minutes later, "We're going to have you talk to another partner. They're not sure." Talk to another. This was four times. Four times I had to talk. And finally, they're like "Okay, okay, we'll let you in."
+
+**Adam Stacoviak:** What was the pitch? I don't want to interrupt you, but I do want to hear this pitch. If you can like maybe even earmark that... I want to hear whatever the pitch was to these four people, or some version of it.
+
+**Drew Wilson:** To be honest, I have no idea. It's been so long... I don't remember if I pitched them the banking as a platform, or if it was at that point the branded bank. I think it was the branded bank, and the pitch was just like "High net worth individuals --" Okay, so high net worth individuals are 2% of the population. They make up 98%, 97% of the dollars in the US. And all of these other startup banks are going after 2% or 3% of the money. Whereas I'm saying "I'm going to go after this pot of like 90 plus percent of the money." Maybe it was 70%, 80%, 90%... I can't remember the numbers.
+
+It's been so long. But it's an enormous amount; vastly outnumbering everybody else on the planet.
+
+\[46:06\] And I was like "There's a huge market here. No one's going after this. We're going to go after this and we're going to win, because we're modern, we'll have an app..." Literally even now, other than Letter, that doesn't exist. There's a reason why, is because, like I said, that demographic just doesn't care. But yeah, so that was the pitch, they bought into it, and they're like "Sweet, let's do this thing." So I got into YC.
+
+What's funny is... I was in YC for three months while working at GoDaddy, and had a huge team at GoDaddy, built up the payments team... I think we were at like 16 people was my team at that time... Plus other teams assisting us. And then I'm doing YC at the same time, and then demo day is going to come up, and I've got to like schedule my GoDaddy meetings to have these gaps where I can do demo day, investor meetings... And I think it was like a day or two before demo day my boss was like "Hey, do you have a second?" And I'm like "Yeah, sure." He's like "Oh, how's it going?" He's like "I'm going to screen share." He's like "So I was going through this website", and he has YC's website pulled up. And he's like "And I scrolled down and I saw this person." And it was a picture of me... \[laughter\] A picture of me and Letter, the company... And he's like "So what's going on?" \[laughs\] I'm like "Yeah, I'm probably not going to stay here forever..." So my plan was I was going to get in YC, raise money, finish up the last bits of stuff I had to do at GoDaddy, and say "Hey guys, I'm leaving." But he preempted that. So then at that moment, I had to be like "Yeah, I'm probably not going to stay. I'm sorry."
+
+**Adam Stacoviak:** Well, that was only a few months after the acquisition, though. Didn't you say you had to stay as an employee for a bit? Maybe a year in?
+
+**Drew Wilson:** Yeah, so this was -- so I came up with the idea maybe six months or so after, and then there was a whole bunch of rigmarole about figuring out how to set up the bank, what to do. I started the company, ran the company I think for like a year before we jumped into YC, or something like that. So yeah, there was about a year and a half in between that.
+
+**Adam Stacoviak:** What was the company doing while you were preparing for YC? Like, what was the company, running the company?
+
+**Drew Wilson:** So it was building the product... Like I said, first we thought we were going to be a banking as a service kind of provider, but then ultimately decided we didn't want to do that. There was already a couple of other players in the space. It seemed horrible to do. And it is. And if you'll talk to those companies, it's horrible. As a matter of fact, there was another banking as a service provider in YC, in my group, a guy named Darragh Buckley. He was the second hire at Stripe. Super-cool guy. We almost -- we were talking about merging our companies at one point. But he was doing that. And so - yeah, we decided not to do that.
+
+Yeah. So we were in the middle of building out this idea of a banking as a service provider, and then we switched and we're like "Okay, we're going to be the retail bank." Because what we were going to do is build banking as a service provider, and then have our own pretend sort of active, sort of not retail bank to give people an idea of what you could build on our platform. And then we just decided, "Let's just build that retail thing. That'll be way better." It was not really way better.
+
+**Adam Stacoviak:** What's the hindsight on that? I mean, so Letter didn't succeed long-term. I'm assuming -- I don't know why it didn't succeed. Why did it not succeed?
+
+**Drew Wilson:** \[49:48\] It ran out of money. It ran out of money. That's the practical reason. The real reason is, is we approached the market incorrectly, and we didn't have the cash to correct. By the time we learned how to correct... And anyone who's in that industry, they know this stuff just takes forever. These feedback cycles take forever. So if we had landed our series A, we would have been good.
+
+So we raised a pre-seed, and then I raised a seed round, and then I ended up spending all my money, way too much money on this bridge round, essentially. It wasn't really a bridge round. It was more just an open access to Drew's bank account to pay for the company, to try to get to this A. So what we did is we had this investor - super-cool guy; became a billionaire a couple of years before we started, to the tune of like 4 billion bucks... And had a bunch of money, started buying up all this Western-themed stuff. And by Western, I mean like horses and cowboy stuff. So he bought up Country Magazine, all the country stuff... He bought up all the American rodeos, he bought up all the cable programming for Western-themed stuff, he was building a Western-themed movie studio, because he's a former movie studio exec... And so somehow I got to be on a Zoom with him, and pitched him what we were doing, and he was into it, and wanted to invest... Mostly because he wanted to create a Western-themed bank under the brand name that he's buying all this other stuff under. So we're like "Yes, we'll do that. We'll white label Letter and we'll have this other thing for you, too."
+
+So that was the deal. He was running it all himself. And then he brought in a team to run that brand that he had made, because he didn't want to run it anymore. And the team he brought in was headed up by the former CEO of Hulu, the guy who sold Hulu to Disney. And so then I was interfacing with him and his team, and their idea was to create an Apple TV Plus or a Disney Plus, but for Western-themed content. Well, it turns out they couldn't secure anything; literally anything. They couldn't secure anything. So he's like "Hey Drew, we don't know what we're going to do here at this brand... But we do know that it's not going to involve a bank if we can't figure out how to make money. So sorry, we're no longer going to fund your bank." And so this was me trying to get this deal across the line for like seven or eight months... And it was just like, it was always right there. We had paperwork sent to them, they would revise, send back, and just like... I'm like "It's so close. Just sign the freaking paper." But then it never happened, and that was early 2023 when they pulled out. And that is the exact same -- the same month they pulled out, it's the same month as Silicon Valley failed, and everybody started pulling money out of there. And we felt that at our startup bank too folks weren't super-comfortable with the startup bank anymore, and moved their money around.
+
+So yeah, we got kind of hosed there... But we did get to use that guy's money to lend to people. We created our own in-house lending platform, where you get a loan in a few seconds, and it was 100% programmatic, meaning there was no humans in the loop at all... Which is crazy. And I don't know of any other thing or program that's done that, but we did that, which was fun. So that's why it didn't work.
+
+**Adam Stacoviak:** \[53:43\] What in the world is involved in building this? I mean, I just look at you -- so I don't want to downplay you. I don't want to say "Drew is just..." But just for the sake of just this conversation, you're the person who taught yourself how to design in Photoshop before there was layers. And it's not saying that you can't learn, but you're somebody who came from literally a zero in terms of like knowledge. You didn't have the design skill, you gained the design skill. You didn't have the engineering skill because you didn't really want to; you weren't interested in it. You gained that. And now here you are, creating a freaking bank. Like, come on, Drew.
+
+**Drew Wilson:** People always get hung up on the bank thing, and think it's so impressive... But I don't know, it's just another business.
+
+**Adam Stacoviak:** It's just another business, huh? It's regulated. It's more regulated.
+
+**Drew Wilson:** It's just highly regulated.
+
+**Adam Stacoviak:** Well, it's just unlike me. I would say that's the reason why it's so remarkable, is that it's so unlike -- like, I don't know anybody who's created a bank, besides you. You're the only person I know. And you're the only person I know that -- well, I think it's actually strange now that I'm realizing what my words that I'm putting out there, is that you're the only person I know that created a bank, and you're the person I thought would never create a bank. And in the same person. \[laughter\] I don't understand that.
+
+**Drew Wilson:** Yeah, I don't think I ever had bank on my bingo card either. Really the reason I wanted to do it is I'm like "No more making little apps. Done. I've done that to the nines. I just want to do something big, that could become a big company."
+
+**Adam Stacoviak:** A big swing.
+
+**Drew Wilson:** Yeah. And so that was my big swing. Since then, I've come to realize that, you know, having a big win is such a -- financially, it's so different than having a large cashflow. Having a big win means you can go out and buy a big house and do a bunch of cool stuff and get all this kind of crazy stuff going on, and invest in these properties, and do that. It's like, it's what you would think of as wealth. That doesn't happen when you have a huge amount of cashflow. If you make $2 million a year, you literally can't do what I just said, because you don't make enough money, even though you're making $2 million a year. It's just so different. But if you don't have cashflow, you can't do anything. You can't control your own destiny. You can't work for yourself.
+
+So having like a cash cow, like an app that's like just killing it, on a small scale, that lets you not have to work somewhere, that lets you get to experiment and do whatever you want... That is the successful life idea. That's not the successful finance idea, that's a successful life idea. And the ultimate goal for a lot of people who are very ambitious is to have both of those. You have this windfall, so you can do a bunch of cool stuff, but then you also have this cashflow.
+
+I gave away my cashflow when I sold my company. I was too inexperienced to know all this at the time. But I did get the windfall, which was great. But now I need the cashflow. And so learning all these things that I've learned - there's always so much more to learn. So I've got to get myself another cash cow, and then hopefully I'll consider myself set. But until then, I can't stop the doing of the things. Yeah, I don't know where I'm going with that, but... There. I said words.
+
+**Adam Stacoviak:** You said words. Well, I think -- okay, so from Plasso, you had a windfall, right? You said it was an all-cash deal. You were able to take care of your investors, plus, so they were able to make their money back, plus a percentage or an X, or however you want to frame that. You paid the employees, you paid early folks who had sweat invested into you with their loyalty, and their abilities, and their friendship, and their whatever, however you want to frame that. And then you obviously had your own money, that was from that, you took on, it sounds like, some massive debt just to keep the company going. I assume that cash took care of that debt, no problem, boom, shabam. But then you said in that story - and tell me if this is the treacherous moment, or if you're still okay - is you said that for a bit there to fund Letter it was an open... The bridge loan scenario, that Letter had open access to, in your own words, "Drew's bank account", to make things go. Did you lose a bunch on Letter, or did you kind of come out even?
+
+**Drew Wilson:** \[58:37\] Pure loss. Pure loss. No, no, no, we made money in that there were people paying us money, and money went to our bank account, but we were not profitable. So I lost a lot of money. More than --
+
+**Adam Stacoviak:** Like millions, thousands, hundreds of thousands?
+
+**Drew Wilson:** Yes, millions of dollars.
+
+**Adam Stacoviak:** Millions of dollars. And this is your own cash.
+
+**Drew Wilson:** Yeah. For me, it was like a -- the decisioning was "Well, if I stop Letter now, it's going to take me four years to get to the point with any other company where I'm at right now with Letter, so I may as well just keep funneling my cash into this thing, and try to make it work, since we've already gotten so far." And that was kind of the reasoning behind it. Had Letter worked out, everyone would have said "That would have been the greatest decision you ever made. That was the best investment you ever made, Drew." But if it doesn't work out, they're like "You probably should have cut your losses earlier." \[laughs\] How do you know in the moment? You don't. And so sometimes you win, sometimes you lose. Actually, let me rephrase that... You always lose, and then sometimes you win. \[laughs\] You're usually losing, and then sometimes you win. So yeah, it makes the wins all the more sweeter when they happen.
+
+**Adam Stacoviak:** For sure.
+
+**Drew Wilson:** But at some point, you know, I'm noticing. I'm 41 now, and you just get sick of it. You just get sick of writing code. You just get so sick of having to design another freaking page... It's like "Gosh, dang it. I'm so sick of it." I've designed every kind of page, so many times. I've written every kind of code, so many times. All the languages, all the SDKs, all the from scratches, all the things. I freaking built a CSS interface to your database with FireRift back in 2009. Oh, my gosh. Like, Tailwind is the only thing close to that now, just doing absolutely insane stuff with the class name stuff. Don't get me started there... \[laughs\]
+
+**Adam Stacoviak:** You like that, or -- is that a good thing or a bad thing?
+
+**Drew Wilson:** I don't want to offend anybody, but no, I don't like it. So I don't know... It just gets old. It gets really old. All these people are 20, and I see them like "Oh, just raised money through YC", I'm like "Dang, you look like you're a baby." \[laughs\] And I'm like "Well, it makes sense." They're young and hungry. I still feel hungry, but I'm not young.
+
+**Adam Stacoviak:** Just not young anymore.
+
+**Drew Wilson:** My big attempts overall turned out well, but they were never like big-big. So I still am after my big-big one. You know, I have these ideas, I'm like "Gosh dang it, there's someone out there making supersonic jets. Gosh dang it, there's someone out there making robots. Why can't I freaking do that?" I want to, but I feel in my mind that there's just this massive gap between what my ideas are and what I gravitate towards, versus what some of these guys are doing when they're making robots, and stuff. My mind gravitates towards solving problems that are in front of me, not towards problems that plague Earth 100 years from now.
+
+\[01:02:03.11\] So I've come to the grips with the fact that it's probably how it's always going to be for me. And so my next thing I want to do is I want to build something that has an enormous cash flow, but is like, we never take VC, and it's just me and somebody else. And I think I'll actually be satisfied with that, which will be cool. Because I've always wanted to be the person that builds the big company, but I just never could figure out how to get there.
+
+**Adam Stacoviak:** Yeah. I've always known you as a -- I would kind of classify you as an inventor, in a way... Not like a literal like you made something, but you've definitely -- like, Pictos was not something, and then creating icon libraries became a thing... The way you engineered to deliver that initially... I mean, you invented all these things to make something very simple, that doesn't have to be that simple. I mean, it could just simply -- you could give me some PNGs, or some SVGs now, or whatever. It could just be like "Thank you, Drew." But no, you went and made it API accessible, or interactable, and like all these different things with Pictos over the years. You're an inventor. I would classify you as that.
+
+**Drew Wilson:** Yeah. And I've always been disappointed in myself for not having the ability to -- I could think larger, I could think of things that I want to do that would be big, but I just don't have the ability to do it. And I've always -- practically, it's broken down to the fact that I don't have the network to do that. So part of me doing Letter was like "Ooh..." Especially like in YC, or I'm meeting these higher net worth people, hopefully that will change things for me. And in some ways it has, but I just need to go even deeper there. Like, that piece -- those people that are creating those large companies are usually super-well connected. And that's kind of why that stuff happens. Practically, that's why it happens. That's why movies that get made... You're always seeing the same people making these movies. Well, it's because they have a network. It's very hard to break in unless you have that.
+
+**Adam Stacoviak:** Network is required. You're right. I mean, you can't just have the skills and the ability and the drive. You have to have the other people \[unintelligible 01:04:20.04\] Like the CFO with the license, for example, so you don't have to go and do that yourself.
+
+**Drew Wilson:** Exactly.
+
+**Adam Stacoviak:** Cram for a week, \[unintelligible 01:04:27.02\]
+
+**Drew Wilson:** Andrew Carroll, man. He's a good guy, if you know him. Yeah. One of my big stretches, though, in that I want to be a part of something very large is this thing called Cortical Labs. I don't know if you've seen that.
+
+**Adam Stacoviak:** No.
+
+**Drew Wilson:** I just launched the new site there. So I'm on the team there as well, doing design... So it's corticallabs.com.
+
+**Adam Stacoviak:** How do you spell that?
+
+**Drew Wilson:** corticallabs.com. And so what we've built is the very first commercially-available biological computer. An actual human brain inside of a computer, and that is the CPU. That is the only CPU. It's pretty rad. We could take a prick of your blood, or sample of your skin, regress that down to a stem cell, put it into the proper environment to then grow it into a neuron, put it on a chip, it'll expand and grow into multiple neurons... That chip that we have, especially designed by us, and on the bottom of it are a bunch of electrodes. So essentially, the neuron is growing on a floor that is made of electrodes. And then we have a bunch of connectors to that electrode, and you can program, you can write code, and you can have the neurons do things for you. And you can receive signal back from them.
+
+\[01:06:05.20\] And the thing that we've built is the thing that translates your code into electrical impulses to the neurons. Also the computer, the CL1, has a bunch of hoses in it to feed the brain essentially electrolytes. Because brains make their own electricity through a chemical process where they break down electrolytes and sodium to convert that into electricity. So you can have this computer that virtually takes zero power; just the LCD screen that we have on there is the only power, and the pump, the little teeny pump that pumps the fluid in and out and cleans it. And so you can have a neuron growing in there for up to six months, or a set of neurons. So you can have this mini brain in there, and they have -- in the past, in 2021, they gave it the rules around Pong, and the neurons taught themselves how to play Pong. So you could watch them playing Pong on the screen.
+
+The idea is that instead of artificial intelligence's actual intelligence, and where artificial intelligence is trying to get to - they're trying to get to AGI, so then you can prompt it and it can think like a human... That thing starts at a human, and it's already there, and it already thinks like a human. So it's a different approach. What we're building now is like an AWS server farm, you could think of it like, but instead of a bunch of computers and hard disks, it's a bunch of these CL1s stacked on top of each other. So you can access a brain computer through the cloud, instead of having to have one. So you can either buy one and develop on it locally, or you can develop through the cloud.
+
+**Adam Stacoviak:** Well, now I'm angry with you, because you just dropped this massive conversation here in front of us... It was not part of my plan, Drew...
+
+**Drew Wilson:** \[laughs\]
+
+**Adam Stacoviak:** This was not part of my plan. You cannot ruin an interviewer's plan.
+
+**Drew Wilson:** Brains...!
+
+**Adam Stacoviak:** I'm just kidding with you. This is wild. And so what role do you play for Cortical?
+
+**Drew Wilson:** So I'm the designer there. And I saw a video on Twitter by Robert Scoble a year and a half or more ago, where he interviewed Hon, who is the CEO at Cortical, and the founder... And he was up in Hon's hotel room. Hon was there for like a trade show of sorts... And Robert met up with him, and like filmed what he was working on... I was watching this thing and I'm like "This is freaking crazy." And it's all the stuff that I've been thinking about in my mind about how things are going to go in the future, and where they're going to end up... And I was like "I have to get ahold of this guy." So I asked Robert, I'm like "Dude, can you get me in touch with this guy? I'm a designer. Do they need design help?" And he's like "I think it's too early for that." And so he wouldn't put me in contact... So I was like hitting up Hon on LinkedIn, on Twitter, on all the ways I could find... Eventually, he responds to me, and we meet up, conversation goes well... They're not totally sure they need a designer or anything like that... And I ended up talking to him for like a year, and finally he's like "Alright, cool. We need a designer. Jump in." And so I got to join the team in like December. And then quickly built out this website in January, and launched it in early February... And then had a little bit -- I was able to change some of the colors on the CL1, but it didn't have any influence on the way that that thing looks... It will on the next iterations. And then now going to be working on the cloud interface. So when you log in, you can get yourself set up on coding against one of these brains in the cloud.
+
+**Adam Stacoviak:** \[01:09:50.18\] Brains in the cloud. I love it. Okay. So let me zoom out then... Who was this person again at GoDaddy that pulled you aside and was like "I was on the YC website..." Who is this person here? I kind of felt like that person in this moment here. Who was that person again? Was he your manager, or was it --
+
+**Drew Wilson:** Yeah, he was my manager.
+
+**Adam Stacoviak:** Your manager. So here you are at Clerk - and I know Clerk, because they've sponsored our show before... I met the CEO, great team there... Super-cool, the way they do auth. I mean, very focused on Next, and the frontend kind of things for when it comes to authorization and this whole entire UI kit... So here you are, you're working at Clerk... I don't know what you're doing, honestly. You've got a job, you've got several jobs... How do you be a designer for Cortical, Clerk...? Like, how are you doing all this?
+
+**Drew Wilson:** Yeah. So at clerk, I'm actually on the leadership team there, and I am the head of design. So I've got a team of like - I think there's 12 of us now designers. And then I'm also the head of commerce and billing, which is our new product, which has a very large team right now, as we're prepping for launch in the next week. So yeah, lots of responsibility there. That is like pretty much all that I work on, is Clerk.
+
+The reason I can work on Cortical and Clerk at the same time - Cortical's designs needs are vastly less than what Clerk needs. And there's not a team, essentially, that I have there. It's just me. So it's way easier to manage. Also, it's not like a paid job. I mean, I have equity, but it's not a paid job where I'm making a salary, or anything. And Clerk is also cool with people doing side projects. A lot of people at the company do side projects, and have side businesses. So as long as it's not interfering with work and you're not doing it during times that you need to be at work, then you're good to go. So that's my situation there. But I just really wanted to be a part of Cortical. I'm not kidding, man... It is literally the first commercially-available biological computer. And I freakin' designed some of it... I mean, I wasn't the guy that came up with the idea or anything like that, but I designed the whole site, and helped design a little bit, and it's cool to be a part of that. There's a lot of good that can come out of that technology, personalized medicine being one of them. You can take your skin cells - literally, Adam, you can have your skin cell turned into a neuron, put into a CL1, and instead of testing... Let's say you have some brain disorder, or some chronic disorder that you're trying to get rid of. Epilepsy, cerebral palsy... There's a bunch of stuff. Let's say you want help, but here's these drugs that were developed, and they could be fatal. Like, there could be all these risks. You take that thing, that's your one shot, right? Well, let's say you could just put that thing, drop it into your other Adam, your mini Adam, your brain, and see how it responds. How does it affect that? If it dies off, you're good to go. It didn't affect you at all.
+
+There's just so many cool things that can come about. And then as a drug manufacturer, you're like "How do I test this in people?" Well, you can't; you have to get all these approvals. You have to first do mice, then monkeys, and this, and that, and it takes a freaking decade by the time you get to people. So with this, you can just start testing it right away. There's just so many cool things that can come out of that technology. It's very nascent, so folks don't know exactly how it's going to be used, what it's going to be used for, but there's just so many potential use cases.
+
+And then Clerk, separately, is a freaking awesome company. The thing I like about Clerk is that if I were to run a company, it would be very similar to how Clerk runs. So I feel very at home at that company, which is really cool. And they're like a step or two beyond what I was able to get to with my companies.
+
+\[01:14:10.00\] Clerk is essentially Plasso, my old company, 2.0. It's the exact same thing. Embeddable auth, embeddable billing... That's exactly what Plasso did back in 2014. And Clerk is the same thing, just the React way, right? The newer style. So it's cool in some ways. I have a lot of experience with that stuff, so it ends up working out.
+
+**Adam Stacoviak:** What made you -- this might show some of your cards a little bit, but hearing your story, I suppose, in this conversation, and then being aware of what you've been doing for the last decade, let's just say, to me, I wouldn't think that you would need a job. Or that you would just have enough wealth, and you could just do whatever you want... What makes you take on this job?
+
+**Drew Wilson:** You would be wrong, Adam...
+
+**Adam Stacoviak:** Am I way wrong?
+
+**Drew Wilson:** You would be wrong, Adam.
+
+**Adam Stacoviak:** Oh, my gosh. Sadness.
+
+**Drew Wilson:** No, it's not sadness.
+
+**Adam Stacoviak:** It's not sadness.
+
+**Drew Wilson:** It's great.
+
+**Adam Stacoviak:** Yeah, let me rephrase that. It's sadness that I was wrong. It's not sadness that that's true.
+
+**Drew Wilson:** Yeah, yeah. I invested a lot of money smartly, and I used a lot of that money for Letter. So I made -- like, in the span between 2018 and 2022 or 2023, before the market crashed, I made a lot of money, and I built this huge office, and addition, and movie theater, and all this kind of stuff in my house... So my house is like this premium playground. And then I spent all the other money on the Letter, which wasn't my intention. So a lot of that extra money that I was counting on to have long-term, I made the decision to spend it. And I was like "Well, if it doesn't work out, I'll just get a job. I'll still be able to maintain the lifestyle I have if I can just get a job. And I should be able to get a job fine."
+
+Obviously, ideally, I didn't want to have to get a job. But the play with Letter didn't work out, and so I did have to, and so that's where I'm at. But what's cool is the investors who invested in Letter were all like "We've never seen anybody do this, so this is really cool that you're doing this, taking on this risk." Most of them said "You shouldn't put your own money into your own company. That's what investors are for", but they all appreciated that I was right there along with them. So even though the transaction didn't work out for them financially, I think for me it worked out in that they have more respect for me. And so should I raise money for something in the future, I think their wealths will be faster opened to me than maybe previously.
+
+**Adam Stacoviak:** Yeah, exactly. Interesting.
+
+**Drew Wilson:** You win some, you lose some, man...
+
+**Adam Stacoviak:** You do win some, you lose some, and I've always --
+
+**Drew Wilson:** I'm way better off than when I started though, so it's not like I took any steps backwards. It's definitely -- it was a hundred steps forward and some steps backwards.
+
+**Adam Stacoviak:** Yeah. Well, don't let this seem like I'm judging your -- this is not a report card for drew, okay?
+
+**Drew Wilson:** Yeah.
+
+**Adam Stacoviak:** This is not me examining your English, and your social studies, and your gym - because gym is always a class, too. I'm not examining those things and saying "Okay, great. You've passed the sixth grade. You can go to junior high", kind of thing. That's not what we're doing here.
+
+**Drew Wilson:** Yeah.
+
+**Adam Stacoviak:** \[01:17:43.24\] But I think we're peeling back the layers a little bit to just understand the reasons for the move. So it sounds like you've got some equity in Cortical. The CL1 sounds super-dope. I mean, I'm not even talking about it. Not because I'm not interested, because I just know it'll take over the whole thing. So I'm not even gonna go there yet. And then you have obviously ValioCon. So you brought this back after - what. You said it hasn't happened for seven years; it's been a big deal for a lot of folks... You're bringing back the conference. What can we expect, I suppose, there? What can we expect with Clerk? What kind of other good next moves do you have in you?
+
+**Drew Wilson:** Yeah, I would suggest if you're listening, go check out valiocon.com, sign up, come out, hang out... It's going to be a good time. I've got a lot of cool speakers that're going to be there. A couple that I'm just waiting on the final -- not actual signature, but final go ahead, and then they'll be up on the site, too. But it's going to be a good time. It's like a vacation. So if you want to like get away, bring your family; if you don't have family, just bring yourself. Come hang out, hang out with old friends, meet new people... It's right on the beach. We do s'mores on the sand, we'll all get lunches together... It's a good time. There's one track, so everyone's in the same room at the same time... And you'll end up making a lot of friends. Folks really love their experience at ValioCon. I've done it for, like I said, six years, and then there was a break. So lots of good things that people have said about it. So come out. ValioCon will be a good time. If you're in design, in engineering, in tech at all, there's going to be a ton of stuff that you can learn from it.
+
+One of the other focuses of ValioCon is giving people who don't normally speak at conferences a chance to speak at a conference. And if they're doing something really cool, I want to feature them. I want them to have a platform to talk about what they're doing. So that's also another focus on ValioCon, is getting folks who've never spoken to come talk about what they're doing. So yeah, definitely come out for that.
+
+What's new in Clerk is we are right on the cusp... I mean, maybe when this thing -- I'm not sure when this episode will go out, but maybe when this episode will go out, Clerk Billing will have launched.
+
+**Adam Stacoviak:** Next week. It ships next week.
+
+**Drew Wilson:** Yeah. Clerk.com/billing is where you'll go for that page... And what's really cool about Clerk is that it makes authentication super-easy. Kind of like Stripe did 15 years ago, or whenever they started. They made this pitch that "We want to own the credit card data at your company, and you're no longer going to have ownership ownership of it. We're going to own it, but we're going to build some really cool software around it to make it really easy to use, and build in a bunch of fraud and safety and monitoring protections for you." And they actually won that argument. People stopped using whatever they were using to reconcile credit cards on their own, and they gave all that data to Stripe. And Stripe built some really cool APIs and stuff around it to make it really easy, and actually easier to access than if you'd done it on your own.
+
+And so Clerk's doing the same thing with the user table... Which is for a lot of engineers a harder thing, but as we scale engineering into, and as it goes downstream into less and less technical folks, with AI now happening, they don't really care. They are happy to give up their user table to Clerk, to say "You handle the user record, because you're going to build this amazing software around it, these cool APIs, these cool UIs... You're just going to make it where I don't even have to think about this stuff."
+
+So Clerk's done a really good job to date on that argument, and winning folks over, and showing how much better it can be to build your app on Clerk that not build your app on Clerk. The next major step to that user, from a company perspective, is like "I want to make money off this user." And so that's where this new product billing comes in. So it's really the second product that Clerk will have ever offered off, and now they have billing.
+
+\[01:21:59.11\] So everything that we're building at Clerk in terms of billing is all native to Clerk. So we're not actually using Stripe's billing product. We're just using Stripe for payments, and eventually PayPal for payments. And you'll be able to pick between them. For right now though, all of the recurring billing logic, all of the invoices, the payment attempts, the emails, all that stuff is native and built into Clerk. And what that means is we can tailor these plans that you create, these subscription plans you're going to create to SaaS, and we don't have to account for every other type of business like Stripe has to. So we can get very narrow-focused to make the UI and the UX way better.
+
+Then separately, there's this idea of entitlements. Okay, someone signed up for a bronze plan, and this is a person who signed up for a gold plan. Well, what do they have access to? Well, usually, if you're using Stripe or some other system, you've got to then build out these entitlements in your app internally, and you have to somehow reconcile them and keep them in sync with the plans that they're signing up with. And you have all these if statements all over your code saying "If bronze plan, and it has this feature, and if that entitlement, do this, that, and the other thing. But if this..." etc. So it gets crazy. And so at Clerk, what we're doing is we're merging those all into one thing. So when someone signs up, or when you're creating your plan, you can assign features to it and entitlements, and you can set the parameters on those, and you can have them be the same on gold, and on platinum, by just adding the features to both of them. Or you can slightly tweak that feature to have this number of credits, versus that one has this number of credits... You could do all that right from the UI inside of Clerk. And then in your codebase, you just use our has function \[unintelligible 01:23:45.20\] the name of the feature. And then it will automatically check to see who are they logged in as, and if you're in an organization situation - because clerk also supports orgs - where you have roles, this has function will automatically look at the role this user is logged in as. Are they logged in as an admin, or a member? And maybe admins have the thousand credits and members only have the 50. And it will just automatically figure that out for you.
+
+So then the entitlement scenario on actually building out your app becomes way easier than it ever would anywhere else. And I mean that sincerely, because that is the truth. So it ends up becoming just a way better experience as an engineer, and you can get to making money so much faster.
+
+**Adam Stacoviak:** Which is the goal, right? I mean, when I recall some of the details behind Clerk, it was really great -- I think they were struggling with migration from other auths, but like really great for green. Like, if you build here, it's amazing, because you don't have to backport all this cruft that you are used to building. You could essentially just start green, and use the has function and other things that are available to Clerk to make it more of a clean flow into what you're building, versus migrating from something else like Auth0, or whatever that might be out there.
+
+**Drew Wilson:** Yup. Yup. I think that's accurate. We are working right now on migration tooling and improving that stuff. That's like a separate team working on that stuff right now. So it'll all get better in the future... And what's cool is the reason Clerk is called Clerk is because we want to take on the boring administrative duties that you don't want to deal with. And to start with, to actually do that -- you know, all software, if you think about it, all software is just decoration on top of the user record. Every single piece of software is that. And so you've got to start at the user record, and then you can build out these other products, like auth, and like a number of other things that we're going to launch in addition to auth. And we're going to move into other business verticals, which is going to be great...
+
+\[01:26:02.07\] So what's crazy is if you look at auth, it's like a 10, 10+ billion dollar industry; tens of billions of dollar industry. But if you look at payments, and you look at commerce, that's a multi-trillion dollar industry. It's like a whole other scale. And so the way I see it is that our billing products will become the breadwinner; not overnight, but decently quickly, within a couple of years, at Clerk. And we'll outpace in terms of revenue all the businesses that we launch by a factor of many, which is great. We help people make money, so it'll be good.
+
+**Adam Stacoviak:** It'll be good. What else? What is left? We've got a few more minutes to chat, about three-ish from my examination of the clock here... What else is left unsaid? What's the next big idea? Can you tease your next big thing?
+
+**Drew Wilson:** Yeah. So I'm working on something - and when I say that, I mean that very loosely... Every week, I get like an hour. It'll slow down soon for me, after launch and everything... But I'm barely working on a thing called Opacity. And this comes from my experience working with teams. And I've teased Opacity on my Twitter account, but I don't really have a website or anything for it yet, though I do have an entire product designed in Figma. And we're working on the engineering side of it now.
+
+So this is something I'm working on with Michael Jackson. So in our spare time he created React Router and Remix, and he and I -- we live very close to each other, our daughters are friends, and we see each other all the time, so we've always wanted to build something together. And so we're working on this now together.
+
+And so Opacity is a way for designers - and engineers too, but mostly designers - to build and maintain components. So the idea of building a web interface... Not a marketing site, but like a web app - components are the way forward, and people have known that for a while. You don't really build things willy-nilly anymore, you build via components. Whether you're choosing React, or Web Components or whatever the crap. You have this idea of this isolated set of functionality based on UI. And since that's what everyone's doing, and that's the way forward, there hasn't really been a tool yet to make that happen in an easy way. If you look at Figma, you can build a component, you can drag things out and create instances... But then, when you actually want to turn that into code, they have this thing called DevMode. And the idea of DevMode is a mission that says "We will never be able to make these things, the idea of a design and the coded version, be the same." The fact that there is a DevMode toggle means we have resigned to the fact that they will never be the same. If they could be the same, there would be no need for a DevMode. They would just be the same.
+
+Figma's canvas is not a DOM. It is a giant WebGL thing. So they can never match the same. As a matter of fact, if you look at the way they do blurs, if you look at the way they do drop shadows, they're never one-to-one with what you see in code. They're bordering; all that stuff works totally differently, because it's not the way the DOM works. Not that one's better than the other, it's just, it's not how the DOM works. So you can never visually get one-to-one.
+
+And they always have to build all these features, like auto layout, and this, that and the other thing, just to try to like match the DOM... \[laughs\] And so what Opacity is is a canvas that is the DOM. And you drag and drop, just like Figma, you build boxes... And the boxes that you build and the inputs that you put in there, everything is all actually just DOM.
+
+\[01:30:09.28\] And so you can style and design it, and you can create a component, and then you can go to the pages tab and like build out a whole page where you're creating instances of those components, dragging them, dropping them... But the thing is, everything that you're doing is actually code. And so the export of Opacity is an npm package. It's a package that you can include into any of your projects, that is essentially all of your components coded out. You don't have to care what we use to style it in CSS. You don't have to care about any of that stuff. All you have to do is like "Okay, export this thing..." And by the way, it looks like GitHub. It's like full version controlled and everything. Because that's the other major piece I want, is like, there has to be -- like, in Figma, if you change something, you don't know who changed it, when did it break, how did it break. You just destroy everything just by one click. And there's no rollbacks, nothing. Opacity has all that. Opacity is like GitHub and Figma combined. But the export is an npm package. And then you have all these props that you set up in Opacity. So you could say, "Okay, the title color can be this, the text size can be that." You can set up any props you want. And then you can expose those... Let me rephrase - you can set up any variables you want, and then you expose those variables as props. You can choose which variables you want to expose and which you don't. And then those props are now available in code. So when you're including my cool new checkout component, all these props that you said you have, you can include them in code.
+
+And as you know, with any component-based system you can actually extend props. So if you have your business logic that you want to add to these components, then you can. You can create an - an actual checkout is the name of the component, which is really a wrapper around the checkout component, and it has all your business logic in it. And maybe you really just use the actual checkout component throughout your codebase. So you can attach these other props to it. But that's up to you. That's however you want to do it. We just care about the look and the feel, and also some of the eventing. There's events built in, so you can have click events, have hover events, all that stuff, that is just not going to change. Maybe you want a hook for the click event, and we give you a way to do that.
+
+So it's the idea of like if you were to build actual products visually, this would be the ideal way to do it, and that's what Opacity is.
+
+**Adam Stacoviak:** Hmm... Eventually a Figma killer, or a Figma acquisition? Could this be your whale?
+
+**Drew Wilson:** Maybe. We don't want to take any investment. We just want to build it ourselves, and build a company.
+
+**Adam Stacoviak:** Cashflow.
+
+**Drew Wilson:** Yeah. A cashflow company. I'm not saying we would never sell. I'm not saying that whatsoever. But I don't think it would be a Figma killer and that people would use this in place of Figma, because if you're designing a website that's not an app, you're going to probably want to use Figma. It's built for that sort of thing. But if you're going to build a UI, I would think that hopefully most people will be moving off of Figma, and onto Opacity for that. That's kind of the goal.
+
+As a matter of fact, one of the things you get - whenever you create a new project in Opacity, you ultimately create a set of components in that project, and you can share that and you can spread that out on the web to anybody you want. And when they click to go to view that page, they don't have to have an Opacity account login, or anything. They're viewing essentially a storybook. Are you familiar with React Storybook?
+
+**Adam Stacoviak:** Mm-hm.
+
+**Drew Wilson:** \[01:34:02.11\] So you see your components visually, and you can see the props, and change them and edit. We're not using Storybook, we're using our own thing. But it's essentially Storybook, and you can see all the components and mess with them. So when you think about it that way, that's what we're building. And that's what you would want to see. If you're like "Here's our design system. Oh, just use this link. It's already built for you. This is exactly how everything works, and here it is all visually." What's crazy is this product doesn't exist anywhere. Yet it's the thing that everyone would want. So it's cool.
+
+**Adam Stacoviak:** So all you've got to do is get more random hours over the next however long to do that.
+
+**Drew Wilson:** Yeah.
+
+**Adam Stacoviak:** Well, I'm sure this next big idea will come to fruition, because you're Drew Wilson. That's how it works. At least in my experience with you. Maybe your own version of your own path is a little different, but from the outside looking into your world, that's how it seems. It'll happen.
+
+**Drew Wilson:** You've got to add an extra F in there. It comes to fruition and failure. That's usually what happens. \[laughs\]
+
+**Adam Stacoviak:** Ah, yes, yes. Well, I mean, there isn't a sweet unless you've got some bitter, right? The bitter is never as bitter unless you've tasted the sweet.
+
+**Drew Wilson:** Yup. Yup.
+
+**Adam Stacoviak:** Yeah. Cool. Well, I'm sure we could talk for longer... It's been awesome having you back on the pod. It's been way too long. If this didn't cover everything you wanted to, I'm happy to have you back on and dig even deeper. I mean, we barely scratched the surface of things I think we could talk about, but it has been...
+
+**Drew Wilson:** Dude, I'm totally down. It's fun chatting with you. I've missed this, it's been great. I mean, if folks don't know, we used to run a podcast together called The Industry back in 2012...
+
+**Adam Stacoviak:** Oh my gosh, man...
+
+**Drew Wilson:** Yeah. So it's been a while.
+
+**Adam Stacoviak:** It has been a while, man.
+
+**Drew Wilson:** And it's always fun. It's always fun.
+
+**Adam Stacoviak:** Jared Erondu, me and you, The Industry...
+
+**Drew Wilson:** Yup.
+
+**Adam Stacoviak:** The Random Show...
+
+**Drew Wilson:** Yeah.
+
+**Adam Stacoviak:** The days, man. The days. And yet, we've never met face to face, I've never been to ValioCon, we've never shaken hands...
+
+**Drew Wilson:** Yup. Dude, the internet's crazy. You could do stuff like that.
+
+**Adam Stacoviak:** Yeah.
+
+**Drew Wilson:** You can be a founder/co-founder situation with a buddy and never really meet them. It's wild.
+
+**Adam Stacoviak:** And I realized in this conversation, in the first few minutes of it, was that you were the person I think I had learned or heard the phrase more frequently than I ever have before, called "What the crap." That's one of your phrases you like to say, because I'm sure you like to cuss a lot, so I'm sure you say "What the crap" instead of something else.
+
+**Drew Wilson:** Yeah.
+
+**Adam Stacoviak:** And I think I began to say "what the crap", because I'm also like that. I like to just cuss less. I'm just not much into profanity. Although I do do it... I just don't do it publicly a lot. You know, I stub my toe, I'm gonna say a bad word. I'm not gonna say "What the crap."
+
+**Drew Wilson:** \[laughs\]
+
+**Adam Stacoviak:** But on a podcast, it's "what the crap", kind of thing.
+
+**Drew Wilson:** Yup. Yup. Dude, I hear you. I hear you. Yeah, it's good. My kids now are old and grown... Almost all of them are teenagers now, and you've gotta watch yourself. They like to do what you do.
+
+**Adam Stacoviak:** They do. And they emulate you. Even if you don't want them to. If you're like "Don't do what I do", they're like "Listen, my default gear is dad or mom gear. Like, wherever they're going, I'm going that direction."
+
+**Drew Wilson:** Yeah, yeah.
+
+**Adam Stacoviak:** It has been fun, Drew. Thank you so much for sharing your journey again, and getting into some details, and we'll see about getting you back on, man.
+
+**Drew Wilson:** Yeah, dude. It's been great. Thanks again for having me, and we'll chat later.
+
+**Adam Stacoviak:** Alright, man.

Creating communal computers (Interview)_transcript.txt

@@ -0,0 +1,507 @@
+**Jerod Santo:** Today, I'm joined by Spencer Chang, who made my day earlier this week. Spencer, you made my day with your Alive Internet Theory. I emailed you, even before, I think, I put it in Changelog News. I'm "I've got to talk to this guy. This guy is making me happy." So welcome to the show.
+
+**Spencer Chang:** Thanks for having me, Jerod.
+
+**Jerod Santo:** So I have been fretting the Dead Internet Theory for probably the last 18 months, off and on, whether it's in Changelog News, or just in conversation on the show... Because I feel it. I feel the encroaching of the robotic content increasing the upward swing. And so the theory to me feels a very good one. \[laughter\] And the recent news, or the recent study that showed -- they've found about a 50/50 break in AI-generated content versus human-generated content, so I'm thinking "Well, if half the internet, at least new stuff being created is taking over, then this theory has some legs to it." You came out with a competing theory, on a website - I'm not sure how theoretical your theory is, or if you're just trying to make us all happy - about the alive internet. And kind of saying "Don't freak out, everyone. The humans are here. We are here and we're staying." I'm putting words in your mouth, but that was kind of the gist, right? Like, let's celebrate the humanity out there. Do you want to tell the story of this Alive Internet Theory?
+
+**Spencer Chang:** Yeah. I mean, I didn't come up with this name. It's interesting, I think it just sort of emerged naturally online. I found a lot of random Tumblr posts referencing this, TikTok posts, some people made YouTube videos about it... And I think it just -- it naturally emerged in response, and it's basically like... The essence is yeah, humanity will always be here. We've created the internet, so much of it so far, and we're not going away.
+
+My take on it - I felt very inspired coming across this stuff, and I've also been thinking about the dead internet theory... And I created this -- the prompt of it was I was commissioned to do a piece for the Internet Archive, and it felt very fitting to sort of bring these two things together, because they've archived so much content that has been created by humans for more than the past -- yeah, three decades or more.
+
+**Jerod Santo:** Three decades, pretty much, yeah.
+
+**Spencer Chang:** They have tens of millions of content, which is just a very crazy scale to think about each of these things was uploaded by a human in some point in the past. And I was just going through the archive's contents from different years, and it's very cool to see... I don't know, it's just very cool to see this content from the past, because there's this human touch to it. And I don't think that's lost in today's world. I still feel a lot of that -- even in social media spaces, where a lot of bots flourish, I think every day they're just... I feel people are finding just crazier and crazier ways to show that they're human, really pushing the human ingenuity... And I'll just see videos that's someone animating a piece of bread, or it's just them throwing a piece of bread on the ground... It's like, AI would never --
+
+**Jerod Santo:** It'd never do that.
+
+**Spencer Chang:** It's too weird.
+
+**Jerod Santo:** Yes. There is a fluidity and kind of a roundness to AI-generated content that humans don't exhibit. We are jagged little pills, as they say... And there are things that we can do, that the current crop of technologies are never going to be able to do. Now, maybe the next crop will pin us down a little bit better... But yes, the essential human aspect of creation is not to be matched by what's out there... Which is why we still can feel -- we can feel the encroachment, because it's still obvious, "Yeah, that's not funny", or "That's not interesting."
+
+Now, humans wielding AI can do new things. You start to get some pretty crazy creations. You're "You know what, that is actually pretty smart." But it was the idea behind it that made it what it was, versus the actual execution... Which is totally fine.
+
+So tell me about this website... I know how it works, I've described it on the show... You load up a time period, and then you hit Go, and now you're just throwing a collage of just insanity at our faces. That's pretty much the idea, right?
+
+**Spencer Chang:** Yeah.
+
+**Jerod Santo:** How do you do that?
+
+**Spencer Chang:** \[07:57\] I basically, just pull random content from the archive for a specific time period, and it's just filtered by media type. So I'm pulling images, I'm pulling audio, and I'm putting video files... And there's no -- yeah, almost no curation whatsoever, because I wanted it to be a very... I wanted it to be an authentic look of the spectrum of things that are on there.
+
+**Jerod Santo:** Gotcha.
+
+**Spencer Chang:** So you might run into NSFW content as well, because that's just part of what's on the internet...
+
+**Jerod Santo:** Because you're not curating it. It's just what it returns, you know?
+
+**Spencer Chang:** Yeah, because if I curated it, it would be my version of the internet.
+
+**Jerod Santo:** Yeah.
+
+**Spencer Chang:** And yeah, part of me making it this overwhelming experience was I thought there was this paradox for the Internet Archive that has archived all this human content, and we also know that it's one of the big sources that a lot of AI companies have trained on... And I sort of imagined an LLM training on the archive... Visually, if I picture what that would look like, it would probably feel a little bit like this.
+
+**Jerod Santo:** Right. It's like the LLMs fire hose of what it had to consume.
+
+**Spencer Chang:** Yeah, exactly. So I don't know, there's interesting empathy with the machine; a bit of that experience as well that I was trying to get at.
+
+**Jerod Santo:** Right. That is interesting. So I'm not sure how long you've been around... I've been around for the majority of it. Do you remember Flickr? Were you a Flickr user?
+
+**Spencer Chang:** I do remember Flickr. Barely...
+
+**Jerod Santo:** Barely. Alright. So you're quite a bit younger than me. That's fine. A lot of people are. Flickr had this really cool thing... And of course, for those who don't know Flickr, it was kind of the original web 2.0 image site where you're sharing your pictures. And it was very popular, very good, built by the same fella who built Slack... Is that right? I think that's right. Check my work there, but - going off memory. And the reason I bring it up is because Flickr had some of the most interesting pictures in the world on it for a very long time. And they had an interesting sort, where it was like "Give me pictures", and you could search an area, or a word, or a tag, and stuff. And then you could say "Sort by interest in this." And I always love that concept of like "Well, give me everything, but let's go ahead and somehow-" I'm sure it was some fuzzy math they were doing inside of Flickr, or some hardcoded thing, who knows, based on stars... Who knows what it was...? Views, probably... Just start with the most interesting and work your way down from there. And I loved that inside of Flickr. And I wonder if the Internet Archive at least has one of those buttons, where like - yeah, you're not curating it, but you could be like "Please send the most interesting things... Because I don't want to bore people as they enjoy the Alive Internet." Or you're just fire hosing it? Whatever they've got, I'll take it?
+
+**Spencer Chang:** Yeah, they have the basic sorts of like - you can sort by views, and stuff like that... But it's almost like because this is historical content... Like, I'm wondering what you saw on a given day on Flickr, when you would turn on that filter... Because this content isn't something that people are actively going to, or returning to. I don't even know how someone would stumble upon it. Maybe only if you're searching for that specific kind of content. So for this case, it's almost like it's better at that -- like, the more interesting stuff is the stuff that's buried, in some way.
+
+**Jerod Santo:** It's funny that you say stumble upon. Now, do you remember Stumble Upon?
+
+**Spencer Chang:** I don't know if I was -- I don't know if I knew about it when I was using the web, but I do know about it. Yeah.
+
+**Jerod Santo:** So Stumble Upon used to be -- it was a Firefox extension. I'm sure it's still out there, in some form or the other; probably bought by three different private equity companies at this point. It was just for fun and discovery. And it was a Firefox extension, and probably other browsers too, that you would throw in your tab bar when you're bored, and you would just click a button and it would send you to a quasi random place on the internet.
+
+\[12:03\] And they had curation, I think, and they were trying to send you interesting stuff, which is kind of why you appreciated it... Because you're going to stumble upon some pretty cool stuff. It wasn't terrible. You're almost guaranteed to enjoy what you're going to land on. And your website reminded me of that, because it was almost like you hit Stumble Upon and you just like clicked on the button like 75 times, and you're just loading them all into the DOM. That's what it felt like.
+
+**Spencer Chang:** Yeah.
+
+**Jerod Santo:** So they hired you to build this.
+
+**Spencer Chang:** \[laughs\] That would've been nice.
+
+**Jerod Santo:** They asked you to build it?
+
+**Spencer Chang:** Oh, you mean --
+
+**Jerod Santo:** Not Stumble Upon. The Internet Archive.
+
+**Spencer Chang:** \[laughs\]
+
+**Jerod Santo:** Oh, you mean Stumble Upon? Yeah, I don't know who they are or where they are. I'm saying you -- this wasn't... You caught a random hair. This was like -- or a wild hair.
+
+**Spencer Chang:** Yeah, well, they asked me to make an art piece to celebrate... They recently hit a trillion web pages, and I was trying to figure out what is the best -- what could I make though.
+
+**Jerod Santo:** Did you have to pitch it back to them, or did they just let you go wild?
+
+**Spencer Chang:** It was pretty -- just a lot of flexibility for the artistic vision.
+
+**Jerod Santo:** Cool.
+
+**Spencer Chang:** Yeah. As long as I used the archive content, or it was inspired by it...
+
+**Jerod Santo:** Yeah, like "Use our stuff somehow to build something cool."
+
+**Spencer Chang:** Yeah.
+
+**Jerod Santo:** And you certainly did.
+
+**Break**: \[13:20\]
+
+**Jerod Santo:** Is this a common thing for you? Or do you pick up grants, or - what do you call them? ...people give you money to create... Or is this an uncommon thing?
+
+**Spencer Chang:** Yeah, I wouldn't say it's common, but I guess I now wouldn't say it's uncommon... Yeah, I'm sort of still even now figuring out what it means to be this independent artist, designer, technologist... It's been almost three years now of doing this full time. I started in January of 2023.
+
+**Jerod Santo:** Really?
+
+**Spencer Chang:** Yeah. And I used to work as an engineer in tech, and I found my -- I mean, every now and then I'll sort of get these commissions from very different organizations; sometimes they'll be more arts-based, sometimes they'll be more technology-based... Yeah.
+
+**Jerod Santo:** And you've found this more interesting than your engineering job.
+
+**Spencer Chang:** Yeah, I sort of stumbled into it. And I wouldn't call it a break from my engineering career.
+
+**Jerod Santo:** Sure.
+
+**Spencer Chang:** \[16:13\] I think it's actually just like -- it felt like the most natural progression of it. So I worked at Coda for three and a half years, which - for those who aren't familiar, is in the document database editing space. And I got into that because I was very inspired by a lot of the research and HCI stuff behind it, like Dynamicland, and Bret Victor, and creating software a lot more accessible to people. Now it's interesting, because a lot of that is coming true, at least the democracy of creating software... But yeah, I felt like what we were missing was -- I thought there needed to be an equal cultural change, as well as technology change, to get people interested and invested in taking agency over technology. If you look four or five years back, there wasn't a broad societal sort of interest, or even awareness of the political power of software. I think it's changing now, with all of the effects of AI and stuff... But yeah, I felt like I wanted to bring those two together. So sort of where I am, I thought that allowed me to both push for a technological change, but also a cultural change.
+
+**Jerod Santo:** And you're pretty good at it, because it's allowing you to live, which - the stereotype of the starving artist is not just a stereotype. There's a lot of artists that don't make any money off of their art, but it's nice that you're able to create... You kind of live at this intersection of art and technology, and there's lots of money on one half of that intersection, on the technology side. So it's kind of a nice place to live, right?
+
+**Spencer Chang:** Yeah, I feel very grateful that I am able to make a living doing this. I'm honestly not sure... It still feels weird to be able to say that. I think it's very -- not as steady of a path, compared to working a standard engineer job. But yeah, I've found ways to think about making a business creative as well, and I think that was one big \[unintelligible 00:18:42.11\] I had to get over, of treating my art and my creations also as a business, from that lens. Because it's important that it sustains. If I want to keep doing this work, if I care about the ultimate impact or possible change that I can create from it, I also need -- it's my responsibility to make sure I can keep doing it.
+
+**Jerod Santo:** Yeah. A couple of weeks ago I was talking to two guys in the Ruby community, who are longtime open source maintainers. And we were talking about that push and pull between sustainability, finance, your hobbies, your passions, and the unfortunate, sometimes, result of parlaying your passion into a job. Because it allows you to sustain it, but it also turns into your job.
+
+**Spencer Chang:** Yeah.
+
+**Jerod Santo:** Have you felt any of that, where it's like "I love making art, but I don't want to make art today, because I have to, because somebody hired me to"?
+
+**Spencer Chang:** Yeah... I have felt slivers of that tension. Only in the space where I've had to be the most business about it, which is - I have a shop for these Internet Sculptures I make, where I put digital experiences into ceramic and other material objects... Just because that involves a lot of manual labor. I'm making every one of them by hand, so there's a lot of repeated processes. And that definitely sometimes could be like --
+
+**Jerod Santo:** \[20:17\] So everytime you get an order, you're like "Dang it." \[laughter\]
+
+**Spencer Chang:** Yeah. Sometimes it's just like "Okay, I'm spending several hours making all these by hand. Dang."
+
+**Jerod Santo:** Yeah. You start doing the math, like "How much am I making per hour on these pieces of art...?" Well, let's go there, because that stuff is so cool.
+
+**Spencer Chang:** Thank you.
+
+**Jerod Santo:** You have this concept that you put on your website of creating communal computers, that people gather around and play, connect, and create. And it seems like that particular thing that you're making is very communal, and is melding the real world with the digital world. Can you tell us more about them, and where this idea came from?
+
+**Spencer Chang:** Yeah. So they're effectively like design objects or art objects, that you would picture putting up for display in your house. I have one that's a ceramic fortune cookie. But what makes them special is I have put RFID tags in all of them, and program a specific digital experience that fits the sculptural form. So the fortune cookie gives you a new fortune every time you tap your phone to it. And yeah, they sort of connect the physical to the digital in this interesting way, and it's cool that other people can experience it as well.
+
+I sort of just stumbled upon it. I was taking a ceramics class with my partner, and just making stuff for fun, and I was feeling this issue of telling people my website whenever I met them... And I literally was just like "Well, it'd be really nice if I just had something I could give people my website."
+
+**Jerod Santo:** Right. "Here's my website" and you hand them a token, or something.
+
+**Spencer Chang:** Yeah, exactly. And I have this stamp on my website that's sort of the website logo, that's made out of my Chinese name, basically. So it's sort of like a sigil. So that was the first piece I made. I wanted something that represented my website, so I put the stamp in this small token, and then put my website on it... And not really expecting too much of it, honestly; just for my personal use. But people reacted very strongly to it whenever I pulled it out. And just being able to feel -- I feel like ceramic especially has a very earthy feel, that's very different from how you perceive software...
+
+**Jerod Santo:** For sure.
+
+**Spencer Chang:** ...and I could see how the experience of my website connected to that was so much more meaningful than just them experiencing it in isolation. And so yeah, that's how I just ended up like "Okay, there's something interesting here", and how I started experimenting more.
+
+**Jerod Santo:** \[23:08\] Yeah, I love that concept. I have heard it one time previously, and I can't quite put my finger on it, but it was at a conference and there was a person who was like... I think they said "Do you want my website?", or something like that, and they handed us a thing. And I'm like "What is this?" And it's an experience. So like "Put your phone up to it." It's a conversation starter, it's different... The technology is pretty straightforward, you know. It's like the same thing as a QR code, basically. You're just encoding some sort of address that gets opened, or whatever. But that doesn't make it not cool. I mean, it kind of makes it even cooler. It's so simple, and yet the melding of the two worlds in a unique way is what is memorable. And unfortunately for this person, it's not memorable enough that I can give them a shout-out, but...
+
+But I've also seen this recent trend at conferences of like coinage being handed out as gifts, or as rewards, or even just a way to stay in touch... And I don't think I've seen any of those. I know Amazon does one, I have one from -- we were at Oxide's office a couple weeks ago and they had little Oxide coins, that are just cool things that you're going to go and... I think I put it over on the shelf over there. It'd be sweet if it had some sort of thing in there that -- and maybe it does, and I should go try. But no one said that. So what other stuff? I think the fortune cookie one's a cool example, because it's not just a URL; it's a fortune, right? So what other stuff are you putting into these objects?
+
+**Spencer Chang:** Yeah. So I did a bunch of experiments... One of the first ones was this pillow that you put your phone on, and it turns on Do Not Disturb.
+
+**Jerod Santo:** Oh, nice.
+
+**Spencer Chang:** And I'm finally actually releasing that tomorrow.
+
+**Jerod Santo:** That's cool. I like that.
+
+**Spencer Chang:** Yeah.
+
+**Jerod Santo:** So can you sow?
+
+**Spencer Chang:** No, it's made out of concrete, actually.
+
+**Jerod Santo:** Oh, it's a concrete pillow.
+
+**Spencer Chang:** It's a concrete pillow, but it does have a fabric tag that you would expect from a pillow thing incorporated into it...
+
+**Jerod Santo:** Yeah, yeah, yeah.
+
+**Spencer Chang:** But I don't know, I just like contrasts, so I wanted it to be concrete, because it looks a pillow, but it's very hard.
+
+**Jerod Santo:** I like that. I just expected it to be an actual pillow.
+
+**Spencer Chang:** Yeah, exactly. I made this bathtub, where you put your phone in the bathtub while you shower, and it plays your shower playlist...
+
+**Jerod Santo:** Is that made out of concrete as well?
+
+**Spencer Chang:** No, that was made of ceramic.
+
+**Jerod Santo:** Okay.
+
+**Spencer Chang:** Most of my experiments are ceramic. Yeah. I made a WiFi router that gives you the WiFi, your home WiFi, for guests when they come over. I made these tiny -- it was like a tiny photo frame with interchangeable pictures, basically. These are all ceramic. And the pictures would link to different photo albums, and they would give you a random photo from the album.
+
+**Jerod Santo:** Okay. All the places this idea can go... Are you tapped out? Do you have other ones in the works? Are you thinking --
+
+**Spencer Chang:** Yeah, I've still got a bunch. I actually just started a content series on Instagram called "Will it internet?"
+
+**Jerod Santo:** "Will it internet?" Nice.
+
+**Spencer Chang:** Yeah. And I just try to turn everyday objects into it, instead of creating a special form for it. The first one I did was just on my -- this is my Casio watch I wear every day, and I just put a chip on it, and it gives me a photo from this minute of this day in the past.
+
+**Jerod Santo:** Okay...
+
+**Spencer Chang:** So right now it's 12:32, November 13th. It would give me a photo that's closest to that in previous years, November 13th.
+
+**Jerod Santo:** So you hold your phone up to it, and it would go through your phone's photo album and find the closest picture to that date in some past year.
+
+**Spencer Chang:** Yeah, exactly. Yeah.
+
+**Jerod Santo:** That's sick. I love that.
+
+**Spencer Chang:** And so it's tied to the minute, which I really like. If I'm seeing a sunset, I might see a sunset from a previous year. It's it's tied to the year and the -- or the minute and the day.
+
+**Jerod Santo:** How does it pick a previous year? It just finds the closest one, it doesn't care what year it is?
+
+**Spencer Chang:** It just picks a random one. It'll find all of them and it picks a random one.
+
+**Jerod Santo:** Right. I love that. How many different aspects -- I guess some of it is informed by, or perhaps limited by what the phone gives you to actually do from, and see...
+
+**Spencer Chang:** Yeah, totally.
+
+**Jerod Santo:** Do you have -- what all can you access, or what all can you do? Obviously, you can launch a photo album to a spot.
+
+**Spencer Chang:** Yeah, yeah. Yeah, it's very --
+
+**Jerod Santo:** \[unintelligible 00:27:40.10\] Do Not Disturb. What else can you do?
+
+**Spencer Chang:** It's very OS-specific. So this is all pretty limited to iPhone stuff. For Android I think you have to get some third party software. I mean, you can trigger a lot of stuff. I don't know. It uses Shortcuts, basically, on iOS, where there's a lot of different stuff you can program.
+
+**Jerod Santo:** Gotcha.
+
+**Spencer Chang:** \[28:04\] Most of the experiences I do rely on just websites, and doing interesting stuff there... But I do wish -- iPhone in general is very protective about NFC access, which definitely limits what you can do...
+
+**Jerod Santo:** \[unintelligible 00:28:20.15\] just walk by somebody and get close enough that they have an NFC chip, and they can -- whatever they want to do. So I understand why they keep it close to the vest... But it certainly limits the art, doesn't it?
+
+**Spencer Chang:** Yeah, yeah.
+
+**Jerod Santo:** That is really cool. So when it comes to the technical underpinnings of this, how are you programming the chips, etc?
+
+**Spencer Chang:** Yeah, so I mentioned on iOS stuff I'm using shortcut programs to run programs on the iPhone.
+
+**Jerod Santo:** Well, you're triggering it from the device, to run on the iPhone.
+
+**Spencer Chang:** Yeah, exactly.
+
+**Jerod Santo:** What do you actually -- is there an API that Apple gives you? I don't understand how NFC stuff works.
+
+**Spencer Chang:** Yeah, it's just Shortcuts, which is this native iOS application...
+
+**Jerod Santo:** It's literally just Shortcuts.
+
+**Spencer Chang:** Yeah. And you can program it to run a shortcut when you --
+
+**Jerod Santo:** So do you write that in Swift, or...?
+
+**Spencer Chang:** No, they have this no-code-esque builder thing for Shortcuts.
+
+**Jerod Santo:** Really?
+
+**Spencer Chang:** Yeah. I mean, now it's really annoying... \[laughs\]
+
+**Jerod Santo:** Wow. So then how do you get it onto the chip?
+
+**Spencer Chang:** It's just like an address thing. It's like a URL style.
+
+**Jerod Santo:** It's like an app-specific URL.
+
+**Spencer Chang:** I'm not actually sure exactly how it works under the hood, but yes, I imagine it's something like that.
+
+**Jerod Santo:** Okay. And where do you acquire these chips from, that you put in the ceramic? It's like you buy them off Amazon kind of thing?
+
+**Spencer Chang:** No, I bought them just from some suppliers I found on Google... But yeah, I've started to source them directly from China now, just because I need very specialized --
+
+**Jerod Santo:** And you might need a lot of them, as you start to sell more of these tokens. What about talking to the chip? Is it just like a USB cable that just connects into them, or a firmware thing, or...?
+
+**Spencer Chang:** Yeah, it's just a phone thing. So you can -- there's a bunch of apps available.
+
+**Jerod Santo:** Oh, so from your phone you connect to it.
+
+**Spencer Chang:** Yeah, exactly. Yeah.
+
+**Jerod Santo:** Okay. You can tell I've never done this, before I'm completely clueless.
+
+**Spencer Chang:** Yeah. It's really magical, honestly.
+
+**Jerod Santo:** Yeah.
+
+**Spencer Chang:** Yeah.
+
+**Jerod Santo:** It's super-cool. So you call that Internet Sculptures? Is that the name of this -- or computing-infused objects? I'm reading some of the words off your website.
+
+**Spencer Chang:** Internet sculptures. Internetsculptures.com is the website for the shop.
+
+**Jerod Santo:** Yeah. But Internet Sculptures, it sounds, is just a step along the road, perhaps? ...a move down the path, but you're going further now.
+
+**Spencer Chang:** Yeah.
+
+**Jerod Santo:** To computing shrines, which is like - okay, similar, but different. Tell us about computing shrines.
+
+**Spencer Chang:** So Internet Sculptures is sort of the business or the shop container for all this stuff.
+
+**Jerod Santo:** The thing you're selling.
+
+**Spencer Chang:** Yeah. I do these experiments, some of them feel cohesive enough, repeatable enough to be products, they go in the Internet Sculptures shop. It also serves as a vehicle to do collaborations with companies or organizations if they want custom ones... And yeah, the manifestation of all those things is in this latest project, Computing Shrines, which is a sculptural version of these same concepts, where you take larger sculptures and create a more immersive digital experience that's only accessible at the sculpture. So it creates this very direct physical to digital bond. And all of them are basically collective experiences. So they have some interaction with visitors who have been there before.
+
+\[32:09\] So one of them is like a -- it's like a phone booth made out of acrylic. You place your phone -- it's like a phone booth for your phone. So you place your phone inside the phone booth, and the website or the web experience is a voice sharing thing, where you hear just the last person's voice, and then once you listen to it, you can record a message for the next person...
+
+**Jerod Santo:** Oh...
+
+**Spencer Chang:** ...and that replaces the message. So you know the only one person, sort of -- your message isn't there forever. It's just --
+
+**Jerod Santo:** Right, it's only the next person in the chain. That's kind of cool... That reminds me of geocaching. Have you ever geocached?
+
+**Spencer Chang:** Yeah, exactly.
+
+**Jerod Santo:** Because they have that concept of like leave a thing -- or take a thing and leave a thing, right?
+
+**Spencer Chang:** Yeah.
+
+**Jerod Santo:** This is cool, because it's that same concept, only there's only one thing. With the geocache, you might open it up and there might be a few trinkets in there, or whatever it is, or notes, or memories, pictures... And you take one with you and you leave your own.
+
+**Spencer Chang:** Yeah.
+
+**Jerod Santo:** And then there's still be a few others in there. Or there might not be. But with this, there's just one message, you're going to listen to that message, you're going to leave one message... And then only one other human in history is going to listen to that one.
+
+**Spencer Chang:** Yeah.
+
+**Jerod Santo:** I love it.
+
+**Spencer Chang:** Yeah. I'm really inspired by geocaching, and all of the -- just very, I would say folk practices of leaving traces of oneself in public places... Like, I even think about people who will write in chalk, or scribble something in the sand... Like, write in chalk on stone, or whatever. There's all these small -- or stacking stones, if you've ever seen that...
+
+**Jerod Santo:** Yeah, I have seen that. When you come across one, you're like "Wow. Somebody spent three hours", or however long it was.
+
+**Spencer Chang:** Right, right. And sometimes you can tell that multiple people have added to it, just by the arrangement of it. I like that there's this sense of how many people have been here...
+
+**Jerod Santo:** Oh, that's interesting.
+
+**Spencer Chang:** Yeah. So all of these are sort of inspired by those practices, but using a digital form, so that it can -- yeah, I'm interested in what other interactive experiences you can do, that amplify those practices. And there's environmental concerns of some of these practices, tagging rocks especially... So this is sort of a way to keep doing those practices, but without -- still tying it to the physical place, but without having to deface anything.
+
+**Jerod Santo:** Yeah. Well, as we began talking about this, my mind immediately went to graffiti, which - I have two minds about graffiti. Sometimes it's beautiful art, and other times you're like "This is just desecration, this is vandalism." And it's so permanent. Of course, you can paint over it, but not without major effort.
+
+So I like this idea of ripples in water... They have this ripple effect, but eventually it goes back to calm. And that's kind of what you're doing with that particular photo booth experience. I think that's a super-cool idea. What are some of your other sculptures you've built, or are thinking about building?
+
+**Spencer Chang:** Yeah. So there's one that's -- so the phone booth is shrine to voice, and there's another one that's shrine to earth, and it's inspired by the rock stacking practice specifically. So I transformed this large boulder, and embedded it with some chips, and the experience is effectively like a rock stacking experience. You load up the website, all of these rocks come falling down, and they're all rocks that have been uploaded by someone who went there... And so you're able to offer a rock to the shrine. So it accesses your photos, it prompts you to search for a rock, it'll verify it's a rock...
+
+**Jerod Santo:** \[35:55\] \[laughs\] I can tell you've put some work into that one... That was very important to you.
+
+**Spencer Chang:** \[laughs\] \[unintelligible 00:35:59.02\] I wanted it to be like -- if I could tweak the model, I wanted it to be like if someone uploaded Dwayne Johnson, the Rock, I wanted it to accept that.
+
+**Jerod Santo:** One exception.
+
+**Spencer Chang:** Yeah. But I'm not that \[unintelligible 00:36:18.11\]
+
+**Jerod Santo:** Yeah, but now, is it set in stone, or can you actually -- can you still access the software side?
+
+**Spencer Chang:** I can still access the software.
+
+**Jerod Santo:** Can you upgrade that sucker? Somebody is going to -- you made my day, but that's going to make somebody's week. If it rejects other humans, but it accepts Dwayne Johnson... I mean, that's the kind of stuff that goes viral.
+
+**Spencer Chang:** I know...
+
+**Jerod Santo:** \[laughs\] I'm just here to encourage you to go ahead and work on that. That's all I'm gonna say.
+
+**Spencer Chang:** On the next version.
+
+**Jerod Santo:** Yeah.
+
+**Spencer Chang:** Yeah, so it verifies it's a rock, it removes the background, and then you're able to add your rock to the pile.
+
+**Jerod Santo:** So are these sculptures then -- is the selling of the objects funding this stuff that you want to do? Or is this also commissioned for other -- because these are living in public spaces, right?
+
+**Spencer Chang:** Yeah. I mean, I'm working on a public art version of it, to be installed in a park.
+
+**Jerod Santo:** Oh, that'd be awesome.
+
+**Spencer Chang:** Yeah. Which would be really awesome. Yeah, I don't know exactly how the funding is going to work, to be honest. I received a grant from San Francisco to try to do this in a park... We'll see, based on permitting and stuff, how that turns out. But yeah, it's tricky for this stuff, especially public art... Public art broadly is just -- as I've learned more about it, it's usually just funded from private patrons or donors, because it's not like a museum installation. It's not something that brings in a lot of money. It's literally a free thing in public.
+
+**Jerod Santo:** Yeah. It's not exactly a line item on the budget when there's so many other things that they're already paying for.
+
+**Spencer Chang:** Yeah. Which is interesting. It's like, that's the kind of thing I want to do, because I've worked on open source on a lot of my projects as well too, and I think a lot of my philosophy is around technology as a sort of public good... And so this is an interesting parallel to that in a very physical universe... So yeah, I'm going to figure out how the funding works. But at least for now it hasn't gone large scale enough that it's prohibitively expensive. I guess I'll cross that path when it comes to it.
+
+**Break**: \[38:47\]
+
+**Jerod Santo:** One of the things you say on your website is that you try to create perpetual energy. Now, I grew up in a world where there was no such thing, so I assume this is aspirational... But I'm curious your thoughts behind that statement.
+
+**Spencer Chang:** Yeah. It's funny, because we know this thing is not real... But I feel like the aspiring towards it allows us to feel that it's present in various points, or various situations. I think I feel that perpetual energy in some of these folk practices that we've talked about, like the rock stacking, or seeing these traces of people... There's nothing physical that is saying that this energy is circulating. But I can just make that gesture, and so many people will see it and might interpret it, and might feel this abstract sense of my presence that feels like there's some cycle happening.
+
+I think it's the same if we look at the software side, things like open source, or this public good style technology... And in that case, there's more of an infinite possibility, because software is free, and can be infinitely replicated... But I think you just see all these projects building off of one another.
+
+I think the thing that I strive for most in the things I make is that people can surprise me with what they do with it. And that also feels like this sort of infinite cycle, or perpetual cycle that gets produced. You just come up with one idea, or a project, and it inspires or enables all these other versions of it.
+
+**Jerod Santo:** Right. Yeah, it's perpetual not through exerting of energy, but through inspiring to exert energy. What's that old saying? "Creation is one percent inspiration and 99 percent perspiration." And when it comes to the perspiring, there's no perpetual for that. That's just good old fashioned motivation, plus discipline, plus whatever it is; stubbornness... But the inspiration part perhaps we can get a little bit of perpetuality moving in that direction, where your idea ripples and inspires another idea, which ripples and inspires another idea. And we've still got to put the work in, we've still got to go build... You've still got to make that new ceramic thing tomorrow, or whatever. But there is a perpetuating of the energy of the idea and the human behind it, which I do think is still probably aspirational, but sounds more achievable than just raw energy perpetuated forever.
+
+**Spencer Chang:** Yeah.
+
+**Jerod Santo:** So that's cool. If you do stumble across that one, let me know, because I'm sure we could put it to good use.
+
+**Spencer Chang:** Yeah. \[laughs\]
+
+**Jerod Santo:** \[laughs\] You could probably fund all your art installations.
+
+**Spencer Chang:** Yeah, that might be my business.
+
+**Jerod Santo:** Yeah. "This is my sustainability model, is a perpetual energy machine."
+
+**Spencer Chang:** I'll pitch that to some investors.
+
+**Jerod Santo:** So much cool stuff, Spencer... What else are you working on? What else are you thinking about? What else haven't we talked about? Because you have a whole -- I just encourage people... Of course, check out Alive Internet Theory, but spencer.place - he's got a bunch of stuff on his website that I'm sure you can just dive into the depths of each of these projects you've had. But what else are you currently working on or mulling on?
+
+**Spencer Chang:** Yeah. One of the projects that I'm really excited about is more on the software side. It's called playhtml. So it's this open source library I made to make making collaborative web elements really easy; as easy as creating a single player web element. So how it works is you can just add a single attribute or program interaction as you normally would, and it syncs that state to everyone. So instead of -- you know, I have a lamp on my website, for example, and you can turn it on and off, and that is the same state for everyone.
+
+**Jerod Santo:** Oh, I did that for everybody that was on your website?
+
+**Spencer Chang:** Yeah. And it persists. So it's like mapping that to a more permanent thing.
+
+**Jerod Santo:** I see.
+
+**Spencer Chang:** And so yeah, I just felt like this paradox of so many millions of people are using modern social media platforms at any given moment, but the actual experience of using the site is very solitary. You have a personal feed, no one else is seeing the same thing that you are in that given moment. There's sort of like a gap between that. And I think having this primitive of like what you do has effect for everyone else - that forms an interesting foundation for new, I don't know, digital social interactions. So that was sort of the premise of the project, and trying to make more interesting digital social spaces using it.
+
+**Jerod Santo:** \[48:27\] There's certainly been a lot of interesting gaming, and maybe even just toy experiences online the last 5 or 10 years that have been driven by a similar component, where it's like we're all experiencing the same thing at the same time. I can't remember who it is, whether it's Neil.fun, or there's another website where they create a lot of cool, interactive games...
+
+**Spencer Chang:** Yeah... Nolan?
+
+**Jerod Santo:** Yes.
+
+**Spencer Chang:** Yeah, yeah, yeah.
+
+**Jerod Santo:** Where it's only fun because there's 7,500 people there, and they're all screwing around on it... And if you were here alone, you'd be like "30 seconds and I'm out." But the fact that there's so many people, and everything I do affects everybody else does bring a new... I don't know what it is, a new excitement to what is otherwise a solitary experience.
+
+**Spencer Chang:** Totally.
+
+**Jerod Santo:** So this is an open source project for other people to build stuff like that. It's not just driving your works, but you're trying to get this to be a thing.
+
+**Spencer Chang:** Yeah, exactly. And it's also I guess an implementation of this philosophy of public good technology. It's an open source library, and I also run the the infrastructure that powers it, basically as a service... For a free service, basically, for anybody who uses it.
+
+**Jerod Santo:** So is there a key-value store in the background, or...?
+
+**Spencer Chang:** It's basically -- yeah, there's a standard sort of database store behind it, and I run the real time streaming infrastructure.
+
+**Jerod Santo:** So you have a server behind this, as well as a data store that you're basically just donating -- so if I'm using playhtml, the open source project, I can basically just piggyback on top of your infrastructure, or I've got to run my own?
+
+**Spencer Chang:** Yeah, you can basically piggyback off of mine. \[laughter\]
+
+**Jerod Santo:** Shh, don't tell anybody...
+
+**Spencer Chang:** \[laughs\] I have audits and I have some checks in place for abuse, but... And it's set up for -- if you're doing tons of data or something, you can run your own infrastructure.
+
+**Jerod Santo:** Yeah. Which would be the kind thing to do, versus...
+
+**Spencer Chang:** That would be the kind thing to do.
+
+**Jerod Santo:** ...effectively hotlinking... It's effectively hotlinking, which we've all learned is not something that you do on the internet without somebody making you pay.
+
+**Spencer Chang:** Right, right. But I think that's such a big barrier for so many people who create interesting experiences on the web. When it gets to the backend infrastructure part, that's the biggest barrier. And usually, these experiences don't need that much; don't need that much bandwidth, and don't need that much storage. And so it just feels like -- it's a very high leverage thing for me to do to provide this, because I've seen so many people who are not from technical backgrounds, but they have a personality to their websites that they want to make... Enabling them to create these experiences has been really rewarding.
+
+**Jerod Santo:** Yeah. It's like, if you're trying to make a massively multiplayer version of Solitaire, that wouldn't make any sense. Or a version of Minesweeper... I mean, you don't want to go take a class in data structures to get your Minesweeper game out there.
+
+**Spencer Chang:** Yeah, exactly.
+
+**Jerod Santo:** It's just not -- this is what a yak shave is, like "Why am I over here shaving this yak? I'm trying to make a game." And so that's cool, that you're providing that for everybody, and obviously as an open source project, as well as anybody -- have you got anybody to use this thing, or you've already got people saying "Thanks, man. Check out what I built with playhtml"?
+
+**Spencer Chang:** Yeah, yeah. And I've run a few workshops and get-togethers, so there's been a lot of --
+
+**Jerod Santo:** Oh, cool.
+
+**Spencer Chang:** Yeah, it's just cool to see what ideas are possible with it. For example, someone made a horse racing game, where there's four horses, and then you just pick a horse, and then you just click on the horse really fast, and it starts moving...
+
+**Jerod Santo:** \[52:17\] That's like straight out of the carnival game. You just point the water at it, you know? \[laughs\]
+
+**Spencer Chang:** Yeah, exactly. Someone made this tug of war game, where you click one side or the other, and everyone's clicks are powering it, and \[unintelligible 00:52:32.10\] going between the two poles...
+
+**Jerod Santo:** Love it.
+
+**Spencer Chang:** Yeah. \[unintelligible 00:52:37.22\] Someone has on their website a knock, they have like a door, and you can knock on it, and you see everyone who's knocking on it at the same time... And then probably the biggest one, it was one I made, but it's this fridge poetry game... Basically, it's playchannel.fun/fridge, and it's just like a huge fridge wall for the internet. You can add words, and you can move them around, and you can delete words, and just see these poems emerge on the website... That's had probably multiple tens of thousands of players go through that.
+
+**Jerod Santo:** "Sometimes I dream of a world full of love, where lightning surrenders with vanishing tenderness." Just a small sampling from your current fridge wall.
+
+**Spencer Chang:** \[laughs\] Yeah.
+
+**Jerod Santo:** That's pretty nice. I assume you didn't write that.
+
+**Spencer Chang:** \[unintelligible 00:53:37.23\] for a while, yeah. And you can see the color under each word... That's a different person. Every person has a different cursor color.
+
+**Jerod Santo:** Gotcha.
+
+**Spencer Chang:** So whatever color...
+
+**Jerod Santo:** And then you can delete words off the fridge. So if you're a jerk, you could just delete this amazing piece of poetry that somebody wrote.
+
+**Spencer Chang:** Yeah. It does stop you if you're --
+
+**Jerod Santo:** It stops you eventually?
+
+**Spencer Chang:** Yeah, yeah. It's actually a similar approach that Nolan did for I think 1 million check boxes, where if you're \[unintelligible 00:54:06.00\] and you're doing it a lot, it tells you to chill.
+
+**Jerod Santo:** It says "Knock it off, jerk." Very cool stuff, Spencer. I'm kind of -- I'm inspired, I'm also jealous of your creativity, and I'm happy that the internet is alive and well with people like yourself creating cool things... Playhtml.fun, spencer.place, aliveinternet...
+
+**Spencer Chang:** Alivetheory.net, yeah. Lots of URLs.
+
+**Jerod Santo:** Many URLs. Internetsculptures.com...
+
+**Spencer Chang:** Internetsculptures.com, yeah. There's shrine.computer...
+
+**Jerod Santo:** Okay...
+
+**Spencer Chang:** I have a lot of domains.
+
+**Jerod Santo:** You're going to have to figure out a funny model for your domain name renewals. That's your biggest expense, probably. \[laughter\] Let me teach you about a thing called subdomains, you know? They work, and they're free. Cool. Love your work, love your perspective and your attitude on the internet and life... I appreciate you joining me, especially on short notice, to chat about the Alive Internet Theory. Anything else you want to say to me or to the audience before I let you go?
+
+**Spencer Chang:** Yeah. I mean, first, just thank you so much for having me on.
+
+**Jerod Santo:** Of course.
+
+**Spencer Chang:** I really enjoyed the conversation, and I feel like we should be talking more about this stuff. Also, it's up to all of us to keep the alive internet a thing. I hope my work is just a seed for all the amazing things that the person listening to this right now could do. I think it takes all of us coming together to preserve the internet that we know and love.
+
+**Jerod Santo:** Beautifully said. Did you hear that, to you out there listening to this? Take this seed, plant it in the ground, and hopefully something beautiful will grow. Awesome. Thanks, Spencer, thank you all for listening, and we'll talk to you on the next one.

DO repeat yourself! (Interview)_transcript.txt

@@ -0,0 +1,339 @@
+**Jerod Santo:** Today we're joined by Sean Goedecke, who is over there in Melbourne, Australia, and has been blogging like a fiend, like a prolific fiend...
+
+**Adam Stacoviak:** Non-stop.
+
+**Jerod Santo:** ...and I've just been like slurping up his blog posts and putting them in Changelog ews. Sean, welcome.
+
+**Sean Goedecke:** Hey. Great to be here.
+
+**Jerod Santo:** Well, because you are where you are, and we are where we are, we'll definitely be dealing with a little bit of lag and latency, so hang with us, everyone... I'm sure it'll get smoothed out in post, but if Adam is telling stories over the top of Sean's answers, that's why.
+
+**Adam Stacoviak:** Blame me. Blame me.
+
+**Jerod Santo:** No, I'm blaming the ocean and the speed of light. I'm not blaming you.
+
+**Adam Stacoviak:** Well, let's get a key and open that room, okay?
+
+**Jerod Santo:** There we go. I actually went back, Sean, and checked... In the last 50 issues or so I think I've included posts by you in Changelog News, one, two, three, four, five different posts of yours, in like the last... That's probably a year. At this point I should just put a redirect into your blog, you know? "Welcome to Changelog News. Here's Sean." You've been killing it, man. Where do you get all these blog posts from? Where are you pulling them out of?
+
+**Sean Goedecke:** Yeah, that's a good question. I'm not quite sure myself. So I didn't write for the longest time. I didn't write for the first eight years of my career. And then all of a sudden, in the last two years, I find myself having a lot of things to say. I think it's kind of self-reinforcing, in a way. It's easy to feel like you don't have a unique perspective to add. But if you write one thing and people like it, it makes you more likely to write other things, and it kind of just keeps going like that.
+
+It all kicked off for me with a post I wrote about how to ship projects, maybe a year ago. And I wrote that out of frustration with people I'd worked with in the past, and people I'd seen work, engineers who were treating the shipping part as like an afterthought, and they sort of felt if they wrote all the code and they completed all the JIRA issues, then the project would be done, and that's how projects work. And I felt and feel now pretty strongly that that's not how projects work. So I wrote a blog post about that, and it hit the top of Hacker News, and got really popular, and it kicked off a whole bunch of stuff... And I thought "Well, okay, if people enjoyed reading about that, I have a lot of other things to say. I'm a pretty opinionated guy." So I kind of started writing more from there, and it turned out to be, I think, a pretty good fit with the kind of people who want to read blog posts about tech... Which is mostly just luck.
+
+**Adam Stacoviak:** I do find that having something to say is obviously the barrier to entry, right? I mean, you have to have something to say, to say something, right? That's pretty clear for everybody. But then there's also the exercise and the discipline to just show up and do. You mentioned in the pre-call that you've got a lot of drafts... We mentioned that we're going to talk about your blog primarily, and obviously the wisdom you've shared in that blog... But break down kind of how you approach, the pragmatic approach to showing up and actually executing.
+
+**Sean Goedecke:** Yeah, sure. So I don't see myself as a particularly disciplined guy. Like, I don't have the kind of three blog posts a week, rain or shine, got to hit my deadlines, got to do the thing every day. I don't have that at all. What I do have, I think, is a pretty good sense of how I like to work, and I'm pretty opportunistic about how I kind of exploit my own peculiarities about work... So I have -- instead of sitting down and working on a blog post and just grinding it out, I'll have like 13 drafts on the go, 13 kind of rough ideas... And often, most of my blog posts come not from a draft that I've been grinding on, but as I'm working on a draft, as I'm writing things down, I write this sentence that's like an unrelated point, and I think "Oh, that's actually kind of interesting." And then I just sit down, write top to bottom a blog post about that sentence, post the blog post, and that's it.
+
+Some of my drafts, I think, will never see the light of day. They're just kind of like breeding grounds for kind of other interesting ideas. And often, it's like -- you say that having ideas is the barrier... Well, I think people have more ideas than they imagine. Every time you get annoyed at like a coworker or annoyed at something you read on the internet, that's an idea. That's something you have that's maybe a little bit controversial, or certainly an idea that not everybody shares. And a lot of those, I think, are kind of worth writing up. That's how many of my blog posts start, where I read something and I think "Well, that's silly", or "That can't possibly be right." And then I go and write about it to kind of convince myself whether I'm right or wrong. And that ends up being a blog post.
+
+The reason when I wrote about 95% of projects failing I think is a good example of that... Where I read a bunch of tweets saying "Oh, 95% of AI projects fail at large enterprise companies. This proves that AI is in a bubble." Well, no, it doesn't. You have to go look at the base rate of how many complicated enterprise IT projects fail. And the base rate is pretty close to 95. So that's what I suspected, and then I went and did a bunch of reading about it and wrote it up. That's what I think is the case.
+
+**Adam Stacoviak:** \[08:22\] Yeah... Well, let me clarify, just in case that was a little off in what I said. I can't rewind and determine exactly what I said, but... Having something to say is different than having ideas. I think that you can have ideas, but I think having something to say is like being down the road a bit, and having some version of experience or wisdom. And I kind of pull this, oddly enough, from a movie... We mentioned in the precall - listeners, you'll get a little taste of it in the after show, potentially... But I pull this from this movie called A Star is Born. It's the most recent iteration of it. It's been a couple versions of it before... Barbara Streisand, I believe, was the most recent one... Well, I guess the most recent one compared to the current one. But in there, the kind of main character - I believe it was Bradley Cooper - telling Lady Gaga, who's his co-star, he says "You've got to have something to say. It's about being an artist, and in particular being a singing artist, a performing artist." And I took that to heart, because I was like "You're right. You can have ideas out the wazoo." Right? Sure. We've all got ideas. But having something to say is different than just having ideas. And I think you have something to say.
+
+**Sean Goedecke:** I appreciate that, firstly. Thank you for saying that. And yeah, I do agree with you, I think, and it ties into how I didn't write much of anything at all for the first seven, eight years of my career. And I think partly that's because - yeah, maybe I didn't have anything to say. And I'm still 10 years in, it's still pretty early to a career. I'm sure my opinions will evolve. But yeah, there's a sense in which like everything I write is the same thing. I'm always making the same points. And one thing I had to get over, I think, in my early few posts was I felt "Well, I can't write this. That's just saying what I've said before. I can't say this, I've already made this point." And I kind of had to get over that in order to keep writing, and I had to kind of convince myself that continuing to articulate my perspective is worthwhile and interesting, even if it is just continuing to articulate my perspective. Even if these points aren't kind of new to me.
+
+I mean, I have a hundred times where I've written "Engineers should be more mindful of how the organization functions, and they should see their role as more than technical." I've made that point in a hundred different ways, and I'll continue to make that point. I guess that's part of what I've got to say. Maybe if you have something to say, you'll just keep saying it forever, until people get sick of it.
+
+**Jerod Santo:** And it's well said, because we do - I'm saying 'we', but I do sometimes discount, or do not want to repeat the same thing over and over again, because you think "Well, I've already said that. It's not interesting anymore." You want something fresh... It's not a fresh take, it's the same old take. But now that we've been podcasting for so many years, Adam, I'm sure we say the same things over and over and over again.
+
+**Adam Stacoviak:** For sure.
+
+**Jerod Santo:** And that's really how you get a message out. I mean, there's a reason why I call it beating the drum, is you're just -- it's the same beat every single time. It's just like, people aren't listening. Different people, different mind spaces... They're not ready to hear it...
+
+**Adam Stacoviak:** Time, even.
+
+**Jerod Santo:** ...or they didn't notice...
+
+**Adam Stacoviak:** Time changes.
+
+**Jerod Santo:** Time changes... And eventually, the message gets out. You can't actually say something just once, in the world that we live in, unless you already have a gigantic microphone, or it's very bombastic, and have it resonate. Like, you do have to actually say the same things in different ways, to reach different people with those messages. And so if you do want your opinion out there, don't feel bad about saying it more than once... He says to himself.
+
+What are some of the other things you've been saying over and over again? So I liked that one: "An engineer must be aware of the politics or the..."
+
+**Sean Goedecke:** \[12:10\] Yeah, it's broader than the technical engineers ought to be aware of the politics of the organization. And broader than that, one thing I've said over and over that I think is a bit more controversial is that senior engineers specifically ought to be heavily involved in the politics of their organization. It's a dereliction of duty if you're a senior engineer and you're like opting out of these political decisions about how the team should function, and what the team should do. It's your responsibility to kind of bring your technical expertise to those discussions.
+
+And part of that, which again is a point I've made over and over, that a lot of people don't like hearing, is that you want to be confident in the workplace as a senior engineer. You want to be more confident than you feel internally. A huge part of the job is being more confident than you feel internally. If you're in a conversation and you're best positioned to know the answer to something, even if you don't feel confident about it, you ought to represent that answer to the organization, because nobody else is in a better position to do so.
+
+Anybody who has to make a difficult decision is going to be kind of dealing with their own uncertainties about it... And if you're in the best place to make that decision, it's your responsibility to deal with your own uncertainties, and not simply to be a technical kind of "Well, it could go either way. There's lots of risks. I wouldn't like to say for sure." No, no. If you're a staff engineer, you ought to say "This is the direction we should go. This is what I believe." Otherwise you're not doing your job.
+
+**Jerod Santo:** So to those who might say "I get it, Sean, I'm supposed to be more confident in my position... But I'm not." How did you and when did you acquire and accrue the confidence necessary to be confident?
+
+**Sean Goedecke:** I don't know if I have good advice on that front. I think I'm naturally very confident in my own ideas... So I wouldn't like to tell people steps that they could take to do that when I didn't follow any steps myself. That's just kind of a fortunate a fortunate way in which my personality fits, I think, the tech industry.
+
+**Jerod Santo:** Gotcha.
+
+**Sean Goedecke:** But in general, I haven't always been confident about tech, specifically; a lot of it was just time, and years working in the same industry, solving similar problems, seeing the same things crop up... Particularly if you're fairly new to the industry and you're not confident, I wouldn't worry about it. I would assume that in 5 years, 10 years, you might get there.
+
+**Adam Stacoviak:** Yeah. Confidence as well, because you kind of have to get on the road a bit, and you kind of have to make mistakes... And when you're new, you're like "Oh, the last thing I want to do is make a mistake, because I might get fired, or I might be seen as less valuable", or "I really am not an engineer." And we continue to perpetuate this imposter syndrome, which even 20 years in I still feel on the daily. I'm a little bit more confident about my awareness of it, but I still feel this level of imposter syndrome... Maybe even in this moment, I don't even know. But to somebody new coming in, you have almost no perspective, so you can't really blame yourself for not having strong opinions or confidence, because you have no reason to have confidence yet.
+
+**Sean Goedecke:** Yeah, that's absolutely right. I think there's a sense in which in tech it's almost easy to gain confidence quickly, if you're working as a software engineer... And that's because codebases are so complicated, and are so different, and contain so much accidental detail that if you're even a fairly junior engineer working on a codebase, as you become familiar with that codebase and able to ship changes, you will rapidly have more technical familiarity with that codebase than even very senior engineers in the organization. In your narrow domain you'll have much more ability to kind of answer questions confidently, and say "No, this is how it currently works. Here's where you'd want to make that change", those kinds of assessments. So I think like there's a sense in which kind of working as a software engineer gives you narrow domain confidence fairly early on, assuming you do stay in the same codebase for long enough to get it.
+
+**Jerod Santo:** \[16:21\] So let's zoom in on that post about company politics, since that's kind of the topic that we're on right here. You have the confidence, you've gotten to become senior enough that you know that you have good things to say, and good direction with regard to where the company should go, how the company should operate... Maybe it's which features are worth developing or not, or how to go about building things... You wrote "How I influence tech company politics as a staff software engineer". I guess, side note, Sean is a staff software engineer at GitHub, for those who don't know... And this was back in October. Do you want to walk us through some of -- maybe the bullet points, or the high points of actually how you go about effecting change inside the org that you work in?
+
+**Sean Goedecke:** Yeah, sure. And I want to preface this by saying that my experience has been with large orgs, whether it's 10, or 20, or 30 engineering teams... So the advice I give is probably not as relevant to smaller companies or startups. But certainly in large orgs, kind of paradoxically, the process of influencing company politics begins with accepting that you're never going to set the direction of the company. It's just too big, the decisions are being made at too high of a level, with information that you straight up don't have.
+
+I'm a staff engineer at GitHub, I'm not going to be able to point GitHub the company in a different direction, and certainly not going to be able to point Microsoft the company in a different direction. Very, very few people are in a position to do that. But what you can do is you can influence your local politics. And I don't mean kind of the company like strategic decisions here so much as I mean the technical decisions about what technologies get chosen, what projects get funded, that kind of thing. And the way to do it - well, there are two ways. The first way is like the hard way, the kind of traditional way, where you have an idea that you want to do, some project you want to build, some feature you want to ship, and you just kind of campaign for it like normal. You write a proposal, you talk about it at talks, you try and convince people, you talk to your manager, you talk to your skip, you're drumming up support for this idea... And that's very hard. It can work. I've seen it work. But it requires a lot of persistence, and it's usually less sticky.
+
+The easy way to influence company politics is to - instead of having one idea that you're pushing, have a kind of stable of ideas that you have ready to go at any given moment. And then when it becomes time for that idea, you'll have something you want to do in the bag already. And when I say time for that idea, companies kind of go through what I like to think of as these waves of interest in different areas. So suppose you have a couple of ideas about how to make the service you're working on more reliable. Maybe you want to do cross-region stuff, or multi-cloud stuff, something like that. It's going to be pretty hard when the company is interested in shipping features to push this big reliability project idea. But in the aftermath of, say, Amazon's recent incident, or perhaps an incident affects your company, there's usually a month or two where people really care about reliability. And it's even coming down from on high, where your VPs, your directors, your managers are looking for reliability projects to fund. And if you're already there with a couple of reliability projects in your back pocket, you're in a position to very easily make those happen, because they're no longer your projects that you're championing; they become your directors or your VPs projects, that they're championing. And so you can get what you want done very effectively, as long as you're willing to wait for the right time. That's the kind of process that I've used, and it's worked very well for me... But it does require you at any given moment to have brief write-ups on 10 to 15 things that you want to see happen.
+
+**Jerod Santo:** \[20:18\] Well, I don't have anything to add on that one, Adam, because I've never been in a large organization. So I can just take Sean's words as gospel truth, I guess, that it just works. I can't even push back, or anything. How about you?
+
+**Adam Stacoviak:** I can definitely resonate with the having many ideas in the bag sentiment, because I think I kind of -- you know me, Jerod, I kind of do that. I kind of have this smorgasbord of things that have just been interesting to me, and I think it really is just timing. So I can resonate with that, at least, but I can't really resonate with the large org lack of control, really fool's errand of trying to influence change at the biggest scale without being CEO. As an IC, that kind of makes me sad that that's the case... But I like the -- you kind of have to not go into those scenarios or those kind of employment opportunities with a control freak mentality... Which I don't think I have, but I do like controlling things. And so it would be challenging for me to be in a large org, I suppose, given that sentiment. I hadn't considered my inability to change the future of things, because I just -- I'm a future changer. It's just -- it's my DNA, it's how I work... And that's kind of where I focus at. So if I can't do that, it might be disappointing for me. Ideas, for sure, though.
+
+**Sean Goedecke:** That's a really interesting point. One theme I get a lot in emails and messages from readers is that I've helped them feel less sad about working for large enterprise companies. I've helped them reconcile themselves to their roles as cogs in the machine... Which - I don't know if I would have picked that as my mission in writing, but it certainly seems to be something that my blog posts are doing. And that definitely is another theme of what I write, which is that... I don't know, I didn't come to the industry from tech. I don't have a computer science background. I have a background in academic philosophy. I did a master's in moral epistemology, and then transitioned to an internship in tech. So I think I lack some of the high mindedness about tech that many of my colleagues have, and I do kind of fundamentally see it as labor... And it doesn't rock me to perceive of my job as being subordinate to capital, because I kind of always imagined I was subordinate to capital, and always will be subordinate to capital in the current system.
+
+So there's a sense in which how I think of making technical decisions in large organizations is kind of flowing on from this kind of broader political commitment to -- you know, there's just not that much you can do about the broad structure of capitalism. You just kind of have to exist in it as best you can.
+
+**Break**: \[23:18\]
+
+**Adam Stacoviak:** Do you know the term emotional intelligence? And if you do, do you think that you are a relatively strong emotional intelligence person?
+
+**Sean Goedecke:** I do know the term emotional intelligence. If you'll forgive a brief digression - I have two dogs, I have two greyhounds. One of them is a complete maniac, and the other is a giant baby, a big boy. Not very smart, but my mom, who's head over heels in love with the dogs, likes to say that he has great emotional intelligence. That's what she always says when he does something stupid. He's like "Well, he's very emotionally smart." That's kind of where he's at. \[laughs\] I had to say that, I'm sorry.
+
+About myself - no, I don't think I have particularly great emotional intelligence. I don't think that's one of my strengths. But what I do have in that area, I think I've come to via analysis, by kind of thinking through it carefully, which I don't think is how you're supposed to do it, but... I don't think it comes particularly naturally.
+
+**Adam Stacoviak:** You're augmenting the EQ for the IQ.
+
+**Sean Goedecke:** Yeah, although I don't know if I have great IQ either. I think I just have a lot of persistence when it comes to thinking about the same problem for years.
+
+**Adam Stacoviak:** The reason why I bring that up is because I'd actually talked to a fellow just last night, as a matter of fact... My son does ninja training; not like ninja fighting, but like ninja, American ninja warrior. I think it's also become popular in other countries, that's where I kind of became aware of it... But like ninja skills, parkour, this athleticism that's just uniquely different than any other sport. And one thing that the guy that runs the place shared with me last night in regards to being an athlete - I think this is a sentiment you share in your blog posts, which is why I bring it up, is this idea of like working on your mental fitness more than just simply your athleticism. Because it seems, like you say, sure, you can have technical skills, and you can understand a programming language, and you can understand TDD, or whatever it might be that makes you a good software engineer. But then you also have to have this mental toughness, this mental clarity, maybe even emotional intelligence, this working on your mind fitness... And that's something I've seen in your blog posts, where you seem to suggest being mentally fit - and I don't mean not crazy, or being off the norm path, but mostly just the mental toughness and the mental game more so than just simply the skill game.
+
+**Sean Goedecke:** That's really well put, and that's absolutely correct. Before I even blogged like I do now, maybe eight years ago, one of the really, really early blog posts I wrote, called "Avoiding Worry-Driven Development", and absolutely most software engineers I've worked with have not been primarily held back by a lack of technical skill... To the extent that they've been held back, it's been by lack of emotional regulation. And there's all kinds of ways this can happen. It can happen even really subtly, where you become afraid to work on things that you might get wrong. You become afraid to, for instance, deploy, or to make changes that could be risky.
+
+So you might be technically very capable, but there are certain kinds of technical change that you're just not really willing to do, that you're like letting other people do who maybe don't have your technical expertise, which - that can really hurt a team. But yeah, there's more ways than I could articulate where kind of poor emotional regulation makes you a worse engineer. I think I'm going to continue writing about that for a long time.
+
+\[30:04\]
+
+***Jerod Santo:** Jerod here in post... Just wanted to let you know that Adam had to step out unexpectedly. So when you don't hear from him for the rest of the show, that's why.*
+
+**Jerod Santo:** One thing I want to talk about, which is tangential to this, is taste. I think, especially with the rise of AI code gen, we are seeing now that our ability to have and to influence good taste in software engineering is becoming more valuable, and our ability to execute the technical bits by hand becoming less valuable. And so you wrote on this - this is one of my favorite pieces that you've written of late... "What is good taste in software engineering?" Because of course, I can just assert that we must have good taste, and we all like our own taste to a certain extent, don't we? ...and so we assume that we all have our own good taste... But how do you know if you do? How do you know what it is? What does it even mean, good taste in software engineering? And in this post, you really try to get down to the nitty-gritty of answering those questions. And the way you go about that is by really proxying taste to values. Can you start there, and explain kind of what that means? Why your taste comes from your values, and how we can actually decide what our values are.
+
+**Sean Goedecke:** Yeah, sure. So this approach definitely comes from my background in analytic philosophy, where if we're not sure what something is, we often try to do what we call conceptual analysis on it, sometimes I think a little bit bombastically, called conceptual engineering... \[laughs\] Perhaps much like software engineers call what they do software engineering... It's a little bit aspirational.
+
+**Jerod Santo:** Right.
+
+**Sean Goedecke:** But the idea is to take this kind of concept that you're not sure about, like taste, and try to explain it in terms of another concept that you have a better handle on, in this case values. So really, you're starting from this point where people use the word taste to mean a lot of different things; perhaps they're not sure what they mean themselves when you get down to it... But if you look at taste as an expression of your values, a lot of things become hopefully a little bit easier to understand.
+
+And by values I don't mean necessarily your ethical values. I mean the things you care about when you sit down to do software engineering. So for instance, you might care about reliability, you might care about performance, you might care about accessibility, you might care about having your code well factored, you might care about having your code not be spread in too many different places... You can list these kind of things forever. And I call them values because I think there aren't definitive answers to a lot of them. Exactly how important is readability compared to reliability? Well, it depends on the situation, and it depends on you. It depends on what you value personally. So I think most software engineers, they end up with this kind of constellation of values where they really care about reliability, they care about accessibility a little bit, or perhaps the other way around... It's sort of what they're willing to trade off and what they're not willing to trade off.
+
+When somebody's sort of set of values match your set of values, you might say they have good taste... I take it one step further, I say that your set of values can be a good or bad fit to the project you're working on itself. And I say that when your set of values is a good fit for the project you're working on, that's having good taste. And the ability to kind of be a bit flexible about your values to fit the project you're working on - I think that's the ability to have good taste in general. Which is - if you'll forgive the reference... It's a very Aristotelian kind of way of looking at it, where the good person, for him, the good man, has this ability to fit his taste to the exact situation with a nicety. I think the same thing kind of applies in software engineering. It's about being responsive to the particulars of the situation.
+
+**Jerod Santo:** \[34:10\] Well, that's interesting. I'm finding that my own values - and I do think that fitting them to the project is smart... I think that's a good way to go about it, because that's ultimately what you're there to serve, is what you're currently working on... Not necessarily yourself, or some sort of ethical high ground. Although there are, I think, values that people consider foundational, and they aren't willing to budge on certain values, and I think that that's a fair way to stand in the world... But man, I'm sure seeing certain values that I've held for many years - so I've been in the game, Sean, for two decades, writing software. And not in large organizations like yourself, so we have different perspectives from that. Always small teams. Lots of times by myself. And I find that because of the advent, specifically in the last six months even, of coding agents that have gotten to be good enough that I'd rather use them than write myself, even though I love writing software, and I still feel like I'm writing software, but I'm not just doing the writing anymore - I feel like some of the values that I've considered bedrock for years of myself are starting to maybe change, or like I'm ready to change them... Because maintainability - okay, in the small; not in a large, but in the small - seems like... Is it all that important? Or will it be all that important, when a rewrite is 10 seconds away, by a skilled developer? And I'm not saying it's here today, but I'm just seeing the trajectory, and I'm just wondering if all my values are going to change. Have you had any sort of those thoughts as this stuff is starting to percolate? Because it's getting to be pretty good at this point, and I imagine it's not going to get 0% better from here.
+
+**Sean Goedecke:** Yeah, absolutely. I mean, you would have more context on this than me having been in the industry longer, but certainly for me the last three years and the last year even have been the largest, the most rapid time of change I can recall for my entire time in the industry. And not even just because of AI. I mean, since 2023 with the death of the zero interest rate regime - this is also something I write a lot about - what it means to be a software engineer has completely changed. And now with the advent of genuinely capable AI agents writing code, it's changing again. And it's sort of continuing to change.
+
+And yeah, a lot of the old ways of thinking about things no longer apply. You almost want to list the things that haven't changed, because the change has been so extreme. Although I do think it depends where you are. One thing I believe is software engineering is -- there's so many different types of work wrapped up in it. And there's definitely people who are working on things for which AI agents are still not yet a good fit, or not yet good enough.
+
+**Jerod Santo:** Oh, for sure.
+
+**Sean Goedecke:** And I think a lot of those engineers kind of look at discussions like these and they think "What are these people talking about? AI agents are just not there yet." But for somebody like me, who works -- a lot of what I do is kind of web servers, and protobuf mangling, and fairly standard kind of forms of software engineering work... The AI agents are here, man. They're certainly good enough to help, and in many cases they're good enough to do the entire work. And it's completely changed what it means to do the job.
+
+**Jerod Santo:** Yeah. I've even experienced that in my own work where there's certain areas of what I'm doing, specifically using - I won't say bleeding edge, but just like modern frontend technologies, and specifically CSS technologies, SVG etc, where I find that they completely lack the ability to do what I want them to do.
+
+\[38:02\] And then I'm doing something else, writing some sort of static generator in Go, and it's like "I didn't write any of this code." I've helped it refactor along the way, but it's totally capable of doing this with just a little bit of direction from somebody who knows what they want, and they have some taste in the way you should go about building the system, to just like guide and say "No, don't do it that way. Do it this way." They're like "Okay, I'll do it that way." And in those areas, it's very capable.
+
+And so yeah, it's just not evenly distributed, both in like technology, and like verticals, but then also across the industry and what you're working on. Everybody's mileage is going to vary, and I think we're getting very small mileage right now, and I just can't imagine that it's going to stay right here... Because I've been calling the AI winter for a little while, and the plateau, and every time I think we've hit one... Granted, a lot of it's been tooling around the models that's been getting better, more than the models themselves right now... But I just know how much investment is in making this get better, and so it's just like... It's like percolating and pending. And so a lot of my hard-earned values and beliefs are kind of like teetering right now. Like, "Slow down to go faster" I've been saying for years, and it's like, I don't know, maybe it's just "Go faster to go faster" now. Or at least it will be in a couple of years. I don't know.
+
+**Sean Goedecke:** Yeah, your example is really interesting to me, about needing to write your own sort of CSS, and SVG code, but having the Golang stuff be automated... Because to me, it's been almost the reverse.
+
+**Jerod Santo:** Has it? \[laughs\]
+
+**Sean Goedecke:** Yeah, I've kind of used AI agents to write most of my frontend stuff, certainly for my blog... But at work, sometimes when I'm writing Golang code, I find that I have to kind of do it by hand, because the agent's not quite there yet for what I want to do. But I think that kind of just reflects, you know, you're going really deep in one area and I'm being very shallow. And when you're writing your Golang static site generator, that's a fairly well understood problem. And when you're doing things that are less well understood, it's just kind of like "Yeah..." It's not so much that it's good at one technology and bad at another, it's just kind of hard to... You know, you can be doing very different things.
+
+But to your broader point, I absolutely agree. I think there's going to be a new set of engineering values that comes in when this technology stabilizes... And it's very hard to say in advance what those values are going to be. I mean, how much engineers are going to need to manually review the work of AI agents as well, I think, for instance, is still very much up in the air.
+
+**Jerod Santo:** Yeah, that's a really big question, is how much -- right now it's made code review extremely valuable, right? Your ability to look at a written piece of code, whether it's from a human or an agent or whatever, and judge appropriately "Is it good enough? Does it need to be rewritten? Where are all the problems with this? Is it going to pigeonhole us? Is it fine?" That, I think, is peak software engineering skill today, and maybe it will continue to be. I mean, there's people that don't think that the agents are going to get very much better at all from here... And if they stay right where they are, it's absolutely a huge value. Maybe not -- so readability, I should say then, is a value that's going nowhere, right? But if we get past this threshold of what we actually have to care about is the inputs and the outputs, and we don't ever have to read it, or very rarely... It's like, readability is for the birds at that point. Who cares? It was a means to an end. So just so much as just weighing in the balances right now is kind of unnerving.
+
+**Sean Goedecke:** \[41:45\] I do think readability is never going to go away. I think it may go away for like Python and Golang and JavaScript, but I think if it does go away there, it's going to kind of recede back into the prompt, in a sense. So for instance, writing readable, coherent assembly code used to be part of the skill of any programmer. When compilers got good enough, that kind of stopped becoming part of the skill set, and instead you just had to write readable C code that would compile into assembly, and you didn't care what the assembly looked like, or whether it was readable. We might be heading towards a world where you don't care whether your Python or JavaScript code is readable. But I think in that world, you'll still care a lot about whether your prompt is readable, or whether the instructions you give to the agent is readable.
+
+Some people are saying - and I don't know if I believe it, but some people are saying that English is kind of the new programming language, that your expression of the spec is sort of your programmer work product. And whatever it is that the programmer or whoever is giving to these AI agents, I think it's always going to be important that that's readable. It's always going to be important that that's understandable by humans, because humans are going to have to update that when they want changes to be made to the product.
+
+**Jerod Santo:** Yeah, 100%. And we definitely see that being formalized in things like spec-driven development at this point... Although we're not at the point where you write the spec and that's all you write. You're still going beyond the spec, but maybe we get to a place where all we're reading are specs and user stories, and stuff like this. I think that's kind of a bland, boring world myself, because... Gosh, I mean, I don't like writing documentation today. A spec is a necessary evil, to a certain degree... Whereas for some reason - and maybe it's just because I've been writing code for so long - I'd much rather read and write code than read or write specs. But I guess we do have to change with the times, don't we, Sean?
+
+**Sean Goedecke:** Well, we'll see. I mean, the ideal specs for AI agents might end up looking a little bit like code. They might end up being kind of technically specific enough to... But I think it's hard to look at the future and say it's getting less interesting to be a software engineer. I mean, the last three years have been a period of such intense change... Whatever you might say about it, it's very interesting to be a software engineer now. There's a lot of things going on.
+
+**Jerod Santo:** It's very interesting right now.
+
+**Sean Goedecke:** For good or ill. So yeah, maybe we're headed towards a world where it's much more boring to be a software engineer, but certainly, if it continues like it has been, that's not what's happening.
+
+**Jerod Santo:** Yeah. Well, it's a bit ironic that what these things seem to be best at are the most artistic aspects of white-collar work... Whereas we were hoping that they would do the dishes and the laundry, instead they're doing the art and the videos and the code. We're writing the docs and they're writing the code. I'd prefer if they wrote the docs and I got to write the code. But you know, I guess take what you can get.
+
+**Sean Goedecke:** It's the great paradox of working with large language models, that they're terrible at the things that ordinary computer programs are really good at, and they're really good at things that ordinary computer programs are terrible at. It's a different way of doing it.
+
+**Jerod Santo:** Yeah. Well, that's where the advent of the tool has been so amazing. Again, when I talk about the software around the models getting so much better, the fact that GPT-5 can't do -- they can't count how many R's are in strawberry, or whatever, famously... But it doesn't have to now. It knows that when a certain set of things it's not good at, it's going to actually shell out effectively to a Python program or whatever it is, or Perl in the case of maybe counting the number of R's in a string... Which is going to get it right quickly and every single time, and then it's going to use that response. That was such a sea change, I think, where it's like, they don't have to be good at everything. In fact, we can limit the domain that we use the models into this very narrow corridor, and then just equip it with all the tools it needs to get the answers. And that's where I think we've gotten in the last year or so where it's like "Okay, this is really good now."
+
+**Sean Goedecke:** \[45:58\] Yeah. I mean, this is something that -- this was anticipated, for sure. I think in 2020 Leopold Aschenbrenner wrote "Situational Awareness." Maybe not 2020. Check that. But years and years ago, long before AI was the thing that it is today. And one of the things he wrote about in that was what he called unhobbling, which is the idea that even if -- once you have a really strong model, you can improve the system without improving the model, because there's so much scope for making the tool better; giving the model the ability to like do things, and then act on those things... There's been so much engineering work in systems like Copilot Agent, Cursor, Claude Code, Codex... All of these things. And that's why I'm kind of skeptical of an AI winter, because even if the models stop getting better, which they might, there's so much clear low-hanging fruit to be picked in the tooling.
+
+**Jerod Santo:** Yeah. That's a good call. And had I known about the tool calling as an option back in 2020, I wouldn't have been calling for this plateau so easily... Because I just wasn't seeing the models get that much better and I kept thinking "Okay, we've hit a plateau. We have to move beyond transformers to a new paradigm." But no, you can just keep equipping these things, and narrowing their scope, and allowing them to -- what did you call it? Situational awareness? To know when to call a tool, and then just make the tools better, better, better, better, and more context etc. and you get huge wins from that without having to have any sort of transitional step change in the actual abilities of the model. So I did not see that coming, but apparently I should have, had I read that particular paper. Or I don't know, maybe you were writing about it back then and I didn't read your blog back then.
+
+**Sean Goedecke:** Not in 2020...
+
+**Jerod Santo:** Yeah, okay.
+
+**Break**: \[47:47\]
+
+**Jerod Santo:** Assuming that code review stays relevant, you wrote a good post called "Mistakes I See Engineers Making in Their Code Reviews." I think it was back in September. And I think that's something that our audience would certainly love to hear, because code reviews right now is such a valuable skill, and doing it poorly is not good for your career. So what are mistakes that people are making, and how can we make our code reviews better?
+
+**Sean Goedecke:** Yeah, sure. So this was a pretty opinionated post I wrote, and I was honestly shocked that the reception was so positive. I thought people would be very negative on it. So maybe it's less controversial than I think... \[laughs\] But the idea that I have here is that -- well, one, you shouldn't leave many comments. This is a thing that I think happens a lot, particularly in these big tech companies, where you'll have code reviews that have 50 or 100 comments sometimes. And I think that's just an order of magnitude too many. A good code review should have like six comments or fewer... Because you just can't track more than those things. There's no way you have more than six interesting things to say about a pull request. Past that, it just becomes kind of line item comments. And the problem with line item comments is that they suck up all the attention that should be spent on a higher level discussion.
+
+So instead of commenting in 30 different places and saying "Oh, here you've used map filter where you should be using reduce... You should leave one comment saying "This is a pattern that you should adopt." And that way the person can kind of like treat it a bit more holistically. That's something I believe.
+
+The other thing that people do, I think, is that when they review a pull request, they read through the diff and they review the diff. And I think that's a big mistake. I think the most interesting and most useful comments you can give in a code review are not about the diff. They're about the code that wasn't written, or the areas of the codebase that haven't been touched, that ought to have been touched.
+
+The highest value code reviews I've ever seen have been comments like "You just don't need to do this, because we have this system here that you could use instead." Or sometimes "You've built this system in this place, which I can see why you did that, but it's much, much better to build it here, for these general high-level reasons." And I think if you're just kind of -- if you're doing a lot of code review and you're just skimming the diff and you're reviewing the diff like a large language model would, you're gonna miss out on those kind of interesting comments. So I think those are the main two things I would say, which is don't leave too many comments, and don't just review the diff.
+
+**Jerod Santo:** Yeah. So that second one is, I think, super-powerful. And I certainly am guilty of just reviewing the diff, and so I'm feeling very seen by this one. And I coached in basketball, Sean... I've got young boys playing basketball, I've been coaching for years... And so as you coach, you try to figure out how to become a better coach, obviously. Otherwise, why are you doing it? And similarly, when I first started coaching, I would just literally coach and review. I mean, that's what a review is, right? Like, you're actually watching, and you're giving advice and opinion based on what you see. And I'd review what they were doing. "You didn't do this right this way... You should put your foot here..." etc.
+
+\[54:17\] And what was way more powerful when I realized I need to be looking at it from a bigger scope and actually telling them what they could have done instead, not what they did right or wrong, but actually they're not seeing this part of the court, or "Here's a better way where you could avoid this thing altogether." I mean, that's kind of what you're talking about here, where it's like, don't look at what they're showing you, look at the entire system, and tell them what they're not seeing. That requires some scope though, and you've got to really understand the system, don't you?
+
+**Sean Goedecke:** Yeah. I mean, for sure. But that's why it's so valuable, because you're bringing to bear your understanding of the entire system. That's why code reviews with that context are so much better, because they're harder, because they incorporate kind of a broader understanding.
+
+**Jerod Santo:** That makes tons of sense. So how long have you been working at GitHub?
+
+**Sean Goedecke:** Oh, five years... Four years.
+
+**Jerod Santo:** Four or five years. And in that time I assume you've done code reviews, so you've done political influence... You've done the things -- you're not just writing about the stuff, you're actually doing the stuff, right?
+
+**Sean Goedecke:** Yeah, yeah, I'd say so...
+
+**Jerod Santo:** \[laughs\] I'm just making sure. I mean, obviously, right? But I want to make sure that's the case. Sometimes we just have color commentaries, and they're not actually out there doing the work. So that's probably one of the reasons why you're so good at it, is because you're actually living the life.
+
+Have you ever been unable to convince inside GitHub something that you want to happen or change, and you thought "I know what I'll do. I'll write a blog post about this. It'll make the front page of Hacker News, and then maybe some of my superiors will read it and I can influence change in that way"? What about that, Sean?
+
+**Sean Goedecke:** Oh ,man... If I've ever done that, it's by accident. I try really, really hard to do the exact opposite of that. I do not write about things I've worked on in like the last year... I try really hard to not give examples from GitHub, from the company I'm currently working at... I am so, so paranoid about writing some blog posts and then having my teammates or my manager read it and think "He's talking about me. Why wouldn't he just come talk to me?" No, no, I don't want to do anything of that.
+
+**Jerod Santo:** It's like the most passive-aggressive thing you can do, is like blog vaguely, veil it, that you're blogging about one of your co-workers, or your boss, or something...
+
+**Sean Goedecke:** Exactly. And they do read my blog, many of my co-workers.
+
+**Jerod Santo:** I assume they do. That was kind of -- my question was facetious, but... I guess the real thing is there, is like at a certain point when you have a blog like yours, which has become very well-read, you have more influence; not just inside of your political team there at GitHub or wherever you're working, but actually you're starting to change the minds of your colleagues across the entire industry, which is going to like boomerang back into GitHub. It's kind of weird. Even if you don't want to, you can't avoid it at this point.
+
+**Sean Goedecke:** Yeah, it's scary... I mean, because I still -- you've been working in the industry twice as long as me. Many of my colleagues have been working in the industry longer than me. Plenty of people have experience I don't have. I'm writing from where I'm at. But I don't claim to have all the answers. I speak confidently because that's how I speak, but there's always the caveat of "This is just from my experience. This is just my opinion." So it's a little bit scary to think about people reading my posts and thinking "Oh, that's how things are. I'm going to change the way I work based on that." I'm like "Well, I hope that works out for you." I mean, it worked out for me, but it's not a responsibility I'm entirely comfortable with.
+
+**Jerod Santo:** \[57:57\] Well, you're very good at what you do... This is why guys like me, who've been doing it twice as long as you have, still want to read what you have to say... Because not only do you have a good head on your shoulders, but you're very good at explaining it in ways that are just easy to read, and it resonates. I find myself nodding along to what you're saying in agreement more often than disagreement, so... Keep up the good work.
+
+Are there any posts you have - and you said before the show started, or maybe it was on the show, I don't know anymore - that you had 13 or so drafts going on. Here we are, November 5th, 2025, getting towards the tail end of the year... What are you thinking about right now? What's on your mind? What do your drafts look like? Are there percolating thoughts that maybe you want to workshop here on the show, or maybe just get out there in some sort of a non-finalized form? The thing about podcasts, Sean, is people give us the benefit of the doubt because we're just talking. Once you write it down and publish it, people take it very seriously, and it's kind of like encoded forever. Now, we do have transcripts, but still, there's way more bandwidth in a podcast, and you have more opportunity to just talk and not feel like every word has to be the perfect word. So we've been thinking about recently that we can maybe just put out on the table and take a nibble on.
+
+**Sean Goedecke:** I don't think I have anything super-controversial in the queue. I tend to write my controversial posts as quickly as possible and I publish them. If something's sitting in drafts, it's probably because it's insufficiently controversial, and I'm trying to think of a way to make it a bit punchier.
+
+But, I mean, one thing that I've been noodling on for a while is my take on the idea that getting the main thing right is so important. This is a draft I've had sitting in my drafts for a month, that -- like, you can be like a grinder, you can be like heads down, 100% doing what you think is like good work, 12 hours a day, and you will get out-competed by somebody who is completely half-assing what they're doing, but what they're half-assing is the right thing, and what you're doing is the wrong thing. And I'm still trying to kind of articulate exactly how I feel about that, but... There have been times in my career where -- maybe this is something that I haven't really expressed in blog posts because I am a bit uncomfortable saying it, but there have been times in my career where I have not worked very many hours in the day. Some days I work a lot of hours, but some days I don't. Some days I work very, very few hours. If I'm just not feeling it, I don't force myself to grind.
+
+And one thing I've sort of grappled with is "Do I deserve my job in those periods where I'm working comparatively fewer hours? Am like cheating my company out of my labor?" And one thing I console myself with is "Well, maybe I'm only doing -- this day I only really got four hours of work in, but it was all in the right direction. So that's more useful than 12 hours of work in random directions." I don't know about that... It's sort of uncomfortable talking about how many hours you work on the internet, particularly when your boss is gonna read it.
+
+**Jerod Santo:** That's completely fair. I would say -- my response to that would be it depends on what your company is hiring you to do. That is, of course, contextual to that agreement, but assuming it's a typical tech company who writes and sells or rents software as a service like GitHub does, they're probably hiring you to move that organization in the right direction, right? They want you as a laborer, to put your thoughts and your experience and your intellect towards progress. And certainly, they probably wouldn't be interested in you grinding in the wrong direction. I mean, this is why we make fun of the 10X developer a lot, just because it's a meme at this point, and of course, there's people who are tech bros who act like they are these things, and I don't think that's the case... But I do think software is a place where it's very easy to move in the wrong direction... I mean, you can become a 10X-er by just being a 1X-er, because a lot of people are just doing the wrong thing, and actually... What's the opposite of progress? ...like, moving backwards. It's so easy to be off and actually be doing harm, even unbeknownst -- I mean, most of the time unbeknownst to yourself.
+
+\[01:02:23.26\] And so being on the right path and doing the right thing is like -- that's the name of the game. And grinding in the wrong direction isn't getting any value. Sometimes, however, you don't know you're grinding in the wrong direction. And so you've just gotta grind until you find out this wasn't right. Other times you're like "I just can't work right now, because my mind isn't right." \[laughs\] And it's like "I've got no problem with it." But of course, that's up to you and your employer...
+
+**Sean Goedecke:** I'm comfortable in my relationship with my employer, but...
+
+**Jerod Santo:** That's great. Great to hear. It sounds like with good reason. I do know that when you're on the main thing, though, you can make major progress really quickly... And you can accomplish a lot in a small amount of time, because it's so highly leveraged. Like, what we do is highly leveraged. And so that's the point about leverage, is when you wield it correctly - massive gains. But of course, when you wield it wrong, you're going the wrong direction. So...
+
+**Sean Goedecke:** Yeah, sometimes I think about it like this, where particularly the more senior you get, if you do one genuinely useful thing a week, that puts you ahead of, I think, most engineers. When I say genuinely useful, I mean like a contribution that would not have happened without you. That is in the right direction, and that actually does realize a lot of value. You can do one of those a week. There's a sense in which you can do one of those a year, and you can still be very much in the black. But of course, you can't do that without doing a lot of different things. You have to try things to see what works. Maybe if you were the sage high up on the mountain, with perfect information, you could do three hours of work a year, and be super-impactful... But I don't think any human being can do that.
+
+**Jerod Santo:** Well, it's like that old commercial where the guy saved a nickel... Do you remember this one? He's running around,
+
+"I saved us a nickel. I saved us a nickel." He's like running around the office, and nobody cares. They're like "Dude, it's a nickel." Nobody cares. And then finally, he walks past like three of the higher ups, and he's saying "I saved us a nickel.
+
+I saved us a nickel." And the right person heard it, and he's like "We do that operation 17,000 times a day", or something like that. The nickel was being saved at a scale. Nobody realized it. I can't remember what the commercial was even about... But I do remember that, because when I talk about leverage - like, there's your thing.
+
+I don't know if you saw the post, Sean, recently, the TikTok intern who saved them 300K a year rewriting a very small part of their service from Go to Rust. Did you see that post?
+
+**Sean Goedecke:** No, I didn't. But it doesn't surprise me.
+
+**Jerod Santo:** Yeah. So there's an opportunity there, where this person comes in for an intern... Of course, they were assigned this task, I believe... And it's like this one particular endpoint of a thing, where it's just -- like, 95% of TikTok is written in Go, I believe, according to that post... And so lots of Go. And Go's a relatively fast programming language and everything, but because of the nature of this particular call, they benchmarked it, they realized the GC pauses or something was causing it to use way more compute than it needed to... And so they just rewrote -- their internship was basically rewriting this one thing in Rust, so they could hyper optimize it... And it saved them like 300K a year in compute. And those are the kind of cool wins you can get when you do the right thing.
+
+\[01:06:02.09\] Now, it was an intern, so is it a unique contribution? Like, probably a lot of people at that org could have written that particular program. It's just cool that an intern got to do it and write about it. But it's an example of what you're talking about, where it's like, if you can bring 300K a year in value in a couple of months - well, that's way more than they're paying you.
+
+**Sean Goedecke:** I mean, it's easier to do that in some roles than others, for sure. I can't talk a lot about it, but one of the really interesting things about working for Copilot, working in like the Copilot area at GitHub has been just the ridiculous growth that Copilot has had financially. And you're very rapidly in a position where you're making technical decisions on which hundreds of millions of dollars are kind of flowing through... And talk about leverage - your mistakes and your wins will effectively cost $10 million here or there. I have made mistakes that have cost a lot of money, I have done things that have recouped a lot of money... It's a really, really interesting position to be in, to have that amount of leverage. It's kind of scary.
+
+**Jerod Santo:** Yeah. It is a high stakes game, which makes it all the scarier, knowing the status of the industry and the lack of sometimes resiliency. The lack of resiliency, and the overwhelming desire to go faster, faster, faster, and to keep up, and to pass. And how much is on the line now. I mean, if you look at our lives now, how much of our lives are informed or completely driven by these digital things that operate them, whether it's our paychecks, our bank account, our voter registration, our votes, our education... Whatever it is. Our scores in our school, our relationships... You name it. Our medical history, as well as diagnoses. Everything's on there now. Everything's in there. And we software engineers are the ones that are like -- we're twiddling those bits, and twiddling the wrong way and things can go poorly. But also opportunity, right? Opportunity to help a lot of people if you do it well. What else?
+
+**Sean Goedecke:** What else? I have this -- this is one of those things where... What is it? Was it W.H. Auden who said that enjoying your own handwriting is like enjoying the smell of your own farts...
+
+**Jerod Santo:** \[laughs\]
+
+**Sean Goedecke:** I think the equivalent of that for blogging is writing posts about blogging, and writing posts about the experience. So I try to do that as rarely as possible, but I have this draft I've been working on about how to write posts that blow up on the internet, and how to write posts that will hit the top of Hacker News, and stuff.
+
+**Jerod Santo:** This is genuine advice, or this is sarcastic advice?
+
+**Sean Goedecke:** No, no, genuine advice.
+
+**Jerod Santo:** Genuine. Well, you mean, you have the ability to do it clearly, you've done it on a repeated basis... So your advice -- you have the clout. What is your summary? Give us the quick list. What do you do? How do you do it? Are you trying to do it, first of all?
+
+**Sean Goedecke:** Yeah, I'd say so. I wouldn't say I'm click-baiting, but I want people to read what I write, and one way to do that is to have it appear in these aggregators like the Changelog, like Hacker News, all these other places. It's hard to say, but a big part of it is finding the exact right tone to hit. You sort of want an academic article, but a little bit more colloquial, but not too much more colloquial. You want to be like way, way, way away from the kind of default GPT-4 style of snappy sentences, but you can't make it too -- it's a fine line to walk. You want a post that people can comment on without reading the post, if you know what I mean. You want a title that's going to kind of have people --
+
+**Jerod Santo:** \[01:10:26.03\] Make the point.
+
+**Sean Goedecke:** Yeah. The title has to make the point for you.
+
+**Jerod Santo:** Right.
+
+**Sean Goedecke:** But the title can't be too punchy, because then you end up in a situation where you have more comments than upvotes on the post... And as soon as you get more comments than upvotes, it triggers the Hacker News flame war mechanism, and your post just gets annihilated off the front page.
+
+**Jerod Santo:** Okay...
+
+**Sean Goedecke:** So you have to kind of -- you've got to be really careful about doing things like asking questions in your title, but you also want your title to be punchy enough, because most people aren't going to read the article. So it's this kind of fine line to walk. The whole process -- I did write a blog post about this, which I really enjoyed writing... I think I called it "Impressing People You Don't Respect."
+
+**Jerod Santo:** \[laughs\] That's a good title.
+
+**Sean Goedecke:** Well, I didn't call it that initially. I was afraid to call it that, and I turned it down, and I've since kind of punched it up a bit. When you're trying to get your blog read, it's a really interesting thing, because I don't care about volume of readers. What I care about is quality of readers, if you know what I mean.
+
+**Jerod Santo:** I do.
+
+**Sean Goedecke:** If the right 10 people read a post, I wouldn't care about the other 90,000. It's getting it to those people who are going to give you really interesting feedback, who are going to email you with their own experience... That's why I write the blog posts. But you can't do that unless those blog posts get seen by a lot of people. And in order for them to get seen by a lot of people, you have to kind of appeal to audiences that perhaps aren't going to give you that interesting signal, that useful feedback.
+
+So there is this weird dynamic where you have to kind of be a bit clickbaity, even though the people who you're really writing to don't care about the clickbait ,because otherwise it's not going to get in front of them... If you know what I mean.
+
+**Jerod Santo:** Right. I do know what you mean.
+
+**Sean Goedecke:** I didn't anticipate that being a thing...
+
+**Jerod Santo:** So you have to attract the people that you don't actually care about them reading it, because they will amplify the signal of it loud enough so that it will reach the people who otherwise might not have seen it, and now they're the ones who are actually the high quality readers that you're trying to reach in the first place. Is that what you're saying?
+
+**Sean Goedecke:** Yeah. Although I do care about all of my readers. And if you read my posts, I love you.
+
+**Jerod Santo:** I'm just over here trying to figure out which category am I, what group am I in over here?
+
+**Sean Goedecke:** To give like a concrete example that I think might make it seem a little bit less arrogant to say this... A lot of the people who are going to upvote your post on Hacker News aren't going to read it. They're going to maybe -- they'll maybe click through and like skim the first sentence, they probably are just going to upvote based on the title. I don't really care about those readers, because they aren't readers. They don't read my posts. But I do want them to upvote, so that my post gets shown to people who are actually going to read it. That's the kind of thing I mean.
+
+**Jerod Santo:** But you don't want to also across the line, so that the person who actually is the high quality reader, they look at your title and they think "Ah, this guy's a hack. He's just trying to get to the top of Hacker News. I'm not going to read this thing."
+
+**Sean Goedecke:** It also doesn't work. Like, super-clickbaity titles are not typically that successful on Hacker News. Sometimes they are, but in my experience, it works better to be a little bit lower key.
+
+**Jerod Santo:** \[01:13:46.29\] Well, I would say we have similar aims here, and I think one of the great things about the Changelog, which we've been doing for a very long time - and we have a niche audience; we're software developers, and so it's a very small group of people... And even among software developers, there's bigger shows, or whatever... But we have all the best listeners. Our listeners are -- and of course, yes, I'm stroking their ego right now as they listen to this, and that's part of my goal... But the other part is just to be honest and tell you they're kind, they're thoughtful, they send us great emails, they're insightful... And so one of the reasons why I love the Changelog specifically, and the audience that we have here - and I'm sure there's tons of crossover with your blog, Sean, so I bet you also have these same people; you've got me at least... Is that they are all the best things that you want in a listener. Like, they don't flame, they don't hate... We get very little -- when we do get critical feedback - which we do, and we appreciate, because that helps us get better - it's always well stated, and clearly with positive intent. We have awesome listeners, and so hopefully you pick up a few readers from this episode. But I think there's a serious crossover going on.
+
+**Sean Goedecke:** As a Changelog listener myself, thank you. I'm going to take all that as praise to me specifically as well, but... \[laughs\] Yeah, I agree. You have a you have a great audience. And I certainly really, really treasure the messages I get from readers, the thoughtful messages.
+
+One thing about blowing up on Hacker News is that you do get a lot of very negative feedback.
+
+**Jerod Santo:** Yeah, totally.
+
+**Sean Goedecke:** You do have to have a thick skin. My initial post about how about shipping projects in large tech companies where I was talking about how the process being fundamentally political... There were a lot of comments on that, and have been since basically saying "This man is clearly a psychopath."
+
+**Jerod Santo:** \[laughs\]
+
+**Sean Goedecke:** "I would not want to work with him in any capacity. I feel bad for his colleagues. He will destroy any company he works at." I get a lot of comments like that too, which I have to just \[unintelligible 01:15:50.05\]
+
+**Jerod Santo:** Yeah... Those - I just laugh at stuff like that. That's good stuff. We've been on Hacker News a fair bit of times, and we've got thick skin at this point. You have to. Or you just have to not read it. Now, it's not quite the most icily of software world, but I think that's reserved for the R programming Subreddit, where they'll also tear you apart, but with less intellectual ways of doing it... I don't know if that's a way of saying it --
+
+**Sean Goedecke:** I was really surprised that Reddit was worse than Hacker News. I thought Reddit would be better than Hacker News... It's not. Reddit is so much worse. The comments are so much more negative.
+
+**Jerod Santo:** Yes. And Reddit has a culture of piling on, especially on jokes, where if one person disses you in a funny way, there'll be reply after reply after reply of like them one upping each other, in like a downward spiral of dirtiness.
+
+**Sean Goedecke:** It's hard to put yourself out there. I have people like -- like, my resume is public on the site. Any post that gets sufficiently popular, there's a handful of people who go and find my resume and tear it apart. They're like "Oh, he's a staff engineer, but staff engineer at like Zendesk and GitHub. Like, what does that really mean?"
+
+**Jerod Santo:** "What does that really mean...?" \[laughs\]
+
+**Sean Goedecke:** It's like, "I don't know, I'm pretty proud of where I am."
+
+**Jerod Santo:** I was going to say, those are good jobs, dude...
+
+**Sean Goedecke:** I think so, but... People on the internet are always going to criticize. But I think it's useful having my name out there. It's useful it being me and it not being anonymous. I think that adds to the value, even though it's kind of hard \[unintelligible 01:17:18.24\]
+
+**Jerod Santo:** A hundred percent. The only problem is I've been pronouncing your name Sean Goedecke in my head, and now, after years of reading you, I know it's Sean Goedecke. And so when I say that the next time you're on Changelog News, I will be sure to pronounce it right. Sean, keep writing, man... Keep writing software, keep writing prose about software, we'll keep reading them... Anything else you want to say to me or to the audience before we call it a show?
+
+**Sean Goedecke:** No, that's it. Just thanks to everyone for reading. As long as people keep reading, I'm going to keep writing.
+
+**Jerod Santo:** Awesome. We'll let you get back to your day... This is the end of my day, but the beginning of yours, so I'm sure you have some writing and some work ahead of you. I'll let you go. Thanks, Sean.
+
+**Sean Goedecke:** Thanks so much.

Fostering open source culture (Interview)_transcript.txt

@@ -0,0 +1,349 @@
+**Adam Stacoviak:** We are back with Arun Gupta. It's good to see you, Arun. It's been a couple years.
+
+**Arun Gupta:** Yeah, it's been about a year and a half or so.
+
+**Adam Stacoviak:** You know, I was really encouraged by our conversation we had at All Things Open... I love All Things Open, by the way. It's been a while. Going back later this year. Arun, you are at Intel... You wrote an awesome book... You've been in open source for - would you say a very long time? Is that how you'd describe it? A very long time? Or do you say the years?
+
+**Arun Gupta:** Well, over a couple of decades, let's say. So it's been a while. It's been a while.
+
+**Adam Stacoviak:** Okay. Gotcha. I like the way you say it in your book, of course, too. The book I'm talking about is "Fostering Ope Source Culture." We just got our copy today, so we don't have super-deep depth, but...
+
+**Jerod Santo:** There it is.
+
+**Adam Stacoviak:** Arun is holding it up. Covering open source culture, business alignment, OSPOs, of course, internal events, inner source, external communities, employee enablement... And then also getting started. So this is a book for anybody really trying to do what you've done in your career, which is jumpstart and prime, worthy corporations to really foster and share the joy of open source, but help others do it well. Where should we begin? What made you write this book?
+
+**Arun Gupta:** Yeah, I mean, this is my fifth or sixth company where I built that kind of open source culture. My experience really goes back to Sun Microsystems... Gosh, 2003, 2004. I mean, literally over two decades ago... Where we were taking the entire company from closed source to open source. And back in those days, the Java EE reference implementation was GlassFish. And as we were building that reference implementation, we were putting that out on the internet, and simple things like -- we would have like a hallway conversation, and somebody reminding that "Hey, you can't have a hallway conversation. Let's send an email out on the public mailing list, so that people in the community can understand what's going on over there." Those are simple cultural nuances. And we did that over a couple of decades ago. And since then, I spent a couple of years at Red Hat, a couple of years at Couchbase, a few years at Amazon, and about two years at Apple. And now at Intel for about three years. Across all these companies, I've been building developer communities, I've been bringing that open source cultural change, and I recognize that pretty much the recipe is rinse and repeat, in that sense.
+
+And as I was moving across these companies, I recognized -- I kept doing the same things again and again... And I was also talking to a lot of startups, a lot of other enterprises, a lot of other friends. I'm part of several open source foundation boards. As I talk to senior leaders across the enterprise, I recognize these are very similar recipes. And that frankly said "Okay, how do I scale this element?" That's why I decided to write the book.
+
+So half the book is about my theory in terms of how do you build that open source culture, why that culture is important. And the other half, the more interesting part I would say, is about 40-plus case studies by 55-plus contributing authors on how and why they have built that open source culture.
+
+**Adam Stacoviak:** I always love a good case study, that's for sure... Especially whenever it's applicable. Sometimes case studies are just case studies for case studies' sake... Because hey, you've got to have white papers, and somebody on your page, or whatever. But wow... I was digging into some of these, and... Like I said, I just got my copy today, so I haven't gone through a thorough depth with it yet, but excited to do so.
+
+**Jerod Santo:** So Arun, you've been at Intel for a few years now. I think when we met you had just joined, or maybe you were there six months, or something like this... And you really came on at Intel to foster this open source effort, or I don't know if you call it a community, or this OSPO, this office, perhaps... Can you speak about your experience there, and maybe how you've applied your learnings inside of Intel, and what's changed as a result?
+
+**Arun Gupta:** Yeah, totally. So I joined Intel about three years ago, April of 2022. And I was hired to build the open ecosystem team. And that's sort of what I built for about a couple of years. Now, for the last few months, I lead the broader developer programs effort at Intel. So for the open source culture part of it, the open source program office was part of my team. My team had the DevRel folks, my team would sponsor the open source foundations, my team would sponsor these open source events...
+
+\[08:08\] We would do speakerships across the industry, we would recognize people internally at Intel who would do the chop wood/carry water work, which is So so essential in terms of keeping an open source project thrive. Because oftentimes code is glory, or the chop wood/carry water is what keeps the project going.
+
+So I think those were some of the key elements that we did. We participated very actively in the inner source practice. We ran a lot of internal open source summits, enablement, advocacy... All sorts of stuff that we did, frankly, to energize the internal engineering community... And also telling the story in a credible manner, in an authentic manner to the outside developer community, why Intel cares about open source.
+
+We have a very long history... I mean, I was looking at our first commit back to open source was back to the GCC compiler 40 years ago. And over the last 20 years, Intel has been the top corporate committer to the Linux Kernel. We are one of the top contributors to Kubernetes, OpenJDK. We are among the top three contributors to both PyTorch and TensorFlow. So in that sense, it's the hidden jewel that we never talked about it. So my goal was to really find those stories inside Intel and tell those stories in a credible, authentic manner to the outside world.
+
+**Jerod Santo:** Yeah. Yeah, I recall that from our conversation, because it was news to me. I just thought "Intel... There's not an open source story there." And it just sounded like it was always there to a certain extent, but nobody was telling that story. It was too far inside, to use your guys' Intel Inside...
+
+**Arun Gupta:** \[laughs\]
+
+**Jerod Santo:** It was too inside for a lot of us devs to even know that Intel was doing open source. So the company was already bought into the overall idea. Does your book also talk about the process of advocating - not just "We should use open source because it's great for our business", but then taking that next step of "Let's not just be users of open source. Let's participate, let's contribute, let's support etc." Was that already established inside of Intel, or did you have to bring that to them?
+
+**Arun Gupta:** No, I think it was very well established inside Intel. We just had to kind of set up some formal programs where that is more recognized. And once that is recognized, then it becomes a lot more sustainable. I'm at Intel, but prior to that I was at Apple, and prior to that at Amazon. And those are the companies where we had to do a lot of grind work in terms of why open source matters. Like, sure, you can leverage Kubernetes to the best of your advantage... But if there are issues, if there are challenges in the project, how do we enable our engineers to participate in the open source community? Because you can't expect to win an NBA game by standing on the sideline. You've got to get into the field, get under the three-pointer line, and shoot something over there to score. And that was the whole narrative there, that "Hey, your problems are very unique to you." You can't just file an issue in the GitHub community and say "The community (quote-unquote) will resolve it." Because there is no nebulous community who's going to care about your issues. Open source is all about understanding that fabric, understanding that dynamic, doing the chop wood/carry water work, so that when you need help, others will be willing to help you out over there.
+
+So a lot of it was -- Kubernetes is built by 80,000 developers across 5,000 companies. There is no need to invest into any such engineering effort. One or two maintainers is all it takes. Either hire the top maintainers, or groom inside the company for somebody to become a maintainer, so that then your issues can be resolved. And then, of course, once you become a maintainer or you hire a maintainer, then you don't put them into dark; you're only going to work on my issues, because that's what brings the credibility to the maintainer as well. That's what brings authenticity to the maintainer as well... That in addition to solving your own engineering challenges, they're continuing to help the community, because that helps them expand their network, their skill set.
+
+**Jerod Santo:** \[12:24\] And what's the best way that a large org - because you're mostly working with enterprise. I mean, your history is all enterprise. Of course, there's small business and startups as well, but for enterprise to engage, obviously, like you said, hire existing maintainers, or groom future maintainers of the project... What are the most -- that to me seems like the obvious one. Like, that's probably the most impact you can have. So set that one aside. What are the best ways that enterprises can engage with their open source communities?
+
+**Arun Gupta:** A lot of different ways, actually. Oftentimes, we think that just the code contribution is the only way you can engage. One of the chapters in the book is really talking about 10 different ways by which you can contribute to open source communities in a non-code manner. Coding is just one part of it. Project management - you know, these projects are heavy on project management. Bug triage, reviewing the code. Technical documentation. And again, these are more on the technical side. But then let's get into a little bit more non-technical side. Usually, these projects sit in an open source foundation. And these projects require security audits. These projects require infrastructure, CI/CD infrastructure. These projects run their events. So you need all sorts of different skill sets in order to sustain these projects. So sponsor those foundations, participate in the governing boards... There are working groups. There are special interest groups like SIGs, that drive the technical discussions in these projects. So encourage your employees to participate in those working groups, so that they can actually continue to have an impact over there.
+
+For example, Intel allows me to be their rep on the governing board of CNCF, the Cloud Native Computing Foundation, and Open Source Security Foundation. And about a couple of years ago, when there was an opportunity that, okay, I could stand up for an Open Source Security Foundation governing board chair... So I talked to my boss, that "Hey, you know what? This is not an additional job that I'm taking up, because it's going to consume a non-trivial amount of my time." She helped me out. She said, "Okay, what do you need?" So then she basically allowed me to hire a person who I can offload a lot of my work in terms of internal open source strategy, so that then I could take over this role.
+
+So I think kind of creating those incentive mechanisms, recognizing the employees, putting that as part of their annual focal or quarterly insights, or their OKRs, goes a long way. Because at the end of the day, we humans are always looking for incentives. So creating those incentive mechanisms, being very deliberate about it, and saying "Hey, your bonus really relies upon this, or your individual performance kind of relies upon this" goes a long, long way... As opposed to "I don't know why you're contributing to this open source." If you start getting down to micro level, that this pull request that you reviewed does not help to my bottom line, that does not get you anywhere.
+
+So I think step back, think at a macro level... How does it help your effort? How does it help the community? Because end of the day, the sustainability of the open source project is super-critical. So for example, at Apple pretty much all of their services run using OpenJDK. So the sustainability of the project is super-critical to them. Because if you were to internally fork the project, OpenJDK is built by thousands of developers. The community knowledge, the intellectual knowledge it brings across that diverse set of developers and use cases is what makes JDK strong. You can't have that kind of complexity, that kind of engineering expertise from just one company. That's the beauty of open source, and that's where the management, the middle management, everybody across the chain has to understand that "Okay, this is why contribution to open source is critical."
+
+**Jerod Santo:** \[16:33\] What's the hardest pitch to actually sell on the inside of these? Because you're preaching to the choir here, and so we're very much like nodding along in agreement... But certainly people get pushback. And some pushback probably appropriately, because it's a lot to commit a lot of resources to something that you don't actually kind of have to commit to. Like, "Yeah, it'd be nice, but we do have a quarterly report to get done, and we have things to hit, and things to sell." So what are the kind of pitches that struggle to get through, and how do you overcome such things?
+
+**Arun Gupta:** I think the biggest \[unintelligible 00:17:13.12\] that I've seen across these companies is just purely lack of awareness. When I was at Amazon, or when I was at Apple, either of those places, "Oh, we've always done it this way, and we were fairly successful. Why pivot? What is this open source thing? Who cares about it?" And again, granted, this was about a few years ago, and a lot has changed over those years. Amazon has definitely become a lot more open source friendly. Not so sure about Apple... But this was purely lack of awareness.
+
+So in that sense, the strategy that we saw worked at Apple was that "Hey, all of this business around iPhone, Mac... We don't want to touch that at all. There is no need for that." Then we kind of create those isolation areas, that "Hey, this Kubernetes thing, this OpenJDK, this PyTorch, this all heavily lives in the open source world. This is completely undifferentiated heavylifting for you. And if we make that work better, that only runs our infrastructure in that much more scalable a manner. That helps us cut down our cost of running that entire infrastructure."
+
+One developer, maybe two contributors, two maintainers goes a long way building Apple's credibility. So for example, when I joined Apple - and I've been on the CNCF board for about eight years. When I joined Apple, there were about 40 people that joined after I joined, because that's sort of how the open source world moves. "Oh, this person is big in open source, and he's joining Apple, and I don't know what is going on in Apple..." That builds trust and credibility. So in that sense, it becomes like a hiring magnet.
+
+So I think all of those things really take time to simmer. And I would say the first few months were a little challenging, because you walk in there and you start questioning everything, you start asking everything... And inertia is always hard, right? You know, Newton's law of motion is always kick in; that thing is moving in a certain direction, you're trying to change the direction, it's like "Wait, wait, wait... What's going on here?" So I think in that sense, that was a bit challenging. But I was talking to my team, my ex team from Apple, and they're saying "Hey, we have--" Because that team now exists. I built the first open source program office at Apple. And they were saying "We have such a wonderful relationship with engineering, with legal, with marketing... They understand the value that we bring, a lot more streamlined operations..." So in that sense, I would say a bit of a patience is required, that "Okay, things are going to work out", and perseverance is required.
+
+\[20:01\] Understanding what the management wants, understanding how you're going to operationalize it across middle management to the engineering level is all that it takes. And being very patient, being very perseverant, and being very clinical about it, that okay, I'm not trying to change the world, I'm not trying to boil the ocean. Here are a few simple steps that would make it work for us. Because in open source, we talk about self-interest. "Does that matter?" "No." That's exactly what works, actually, as a matter of fact. I'm not doing this for the charity. It's enlightened self-interest. I'm participating in the open source community, so that when there are issues, I can just jump in and fix my issue, and that really reduces my downtime of an application.
+
+So I think those are some of the pitches that I've seen that seem to work better, and be more acceptable. In case of Amazon, for example, Amazon is always, always keen on working with the customers. So when I joined Amazon back in 2017, we were launching Amazon EKS. I joined in March or So and then I learned about "Hey, we're going to launch EKS" at re:Invent that year. So that whole next seven, eight months was helping Amazon understand "How do you work with the open source community?" You don't just walk in -- I mean, AWS, in that sense, is the biggest hyperscaler. You just don't walk into an open source community, "I want it done this way." So teaching sort of the dynamics of the open source community, helping them build those open source relationships goes a long way.
+
+**Jerod Santo:** There's kind of two angles into open source as an enterprise. You have "Let's support, embrace, extend", not extinguish, but extend, "the stuff that we currently are already using", right? Like "We already depend on this thing. It's core to our business. We're going to actually put some money into it, or put some engineering effort into it." And then the other one is like "Let's take something that we have built, that's internal and proprietary, and let's open source that." And I think that there's different concerns and different pushbacks to those two different types of embracing. On the actual "Let's open source some stuff front", certainly there's people inside of large enterprises that say "Yo, we can't do that. This is our competitive advantage." What do you say when people say "This is what we sell. Let's not give it away", for instance?
+
+**Arun Gupta:** Yeah. My first question anytime a team comes to me that "We want to open source this", my question is "Why? Why do you want to open source this? What's in it for you to open source this?" Are you looking to gain mindshare? Are you looking to get more engineering contribution? Is that a mandatory requirement? ...say the partners are telling you "We would not adopt this technology until it is open source, because then we are not just relying upon you, but we can fork it if we need to, or we can contribute to it directly." So I think there are several reasons. And if I go through the book, essentially... You know, this is the book that we were talking about. There are several Why that we talk about. Is it a faster innovation? Is it a more sustainable codebase? Because then, essentially, you're not just reliant upon your engineering teams. You could do fun things, but then in the project you can also create Good First Issue, where you can encourage developers from around the world to help contribute over there. It could be cost savings, as a matter of fact.
+
+For example, I may not want to run CI/CD infrastructure, and a hyperscaler may be interested in contributing credits for you to give those infrastructure. It may be a strategic initiative for the company.
+
+\[24:04\] So for example, when Google launched Kubernetes, it was made very clear - if you keep it as a Google project, even though open source, nobody would care about it. That's the reason they ended up contributing Kubernetes to a foundation, so that it's now a neutral governance body etc. And then everybody else kind of jumped onto the bandwagon and became a success.
+
+It could be a compliance and security, as a matter of fact, or community building. Or it could be sheer market visibility, that "Hey, this company--" Apple, for example, launched Swift, and they just want market visibility, because the only way to build developers around that is getting Apple's name out there and bringing all those students and those developers who can then contribute directly to the language.
+
+The reason Intel contributes to these open source projects -- I mean, we contribute to 300+ open source projects. You pick an open source project, and we likely contribute to it. The reason we do that is because our customers, for example, who use our Intel Silicon - whether it's a hyperscaler, a laptop, a network, an edge, whatever it is, they download these open source projects, and they expect them to work in an optimal manner. So the Why really is "Because our customers expect these open source projects to leverage the latest chipset. That's the reason we contribute back to these projects."
+
+I think whether you are contributing to an existing project, or you are open sourcing an existing project, the Why, the business alignment is fundamentally critical on why you want to do that.
+
+**Break**: \[25:50\]
+
+**Adam Stacoviak:** You mentioned having some case studies... I don't know if you have any case studies around this, but the question really is is aside from Kubernetes - that's a good example of Google open sourcing it, not holding onto it; they gave it to the Foundation. What are other good examples - and maybe you've written about them in your book, I'm not sure - that are beyond Kubernetes, that are examples of a large corporation giving something...? Or even a smaller startup even; not even so much a corporation. But what are some other examples beyond Kubernetes?
+
+**Arun Gupta:** Yeah, I think the book has about 40+ case studies, and these are case studies from companies like Mercedes-Benz, Bloomberg, Toyota, GitHub, Infosys, Dell, Amazon... A wide range of companies. Fidelity, Johns Hopkins, UC Santa Cruz. So from a wide range of verticals as well. And these case studies really talk about why they care about open source culture, and what have they done within their company to foster that open source culture.
+
+So for example for Bloomberg, one of their big corporate philosophies is philanthropy. And for them, contributing to open source fundamentally aligns with that corporate philosophy. And as a matter of fact, what they also do is for developers who contribute to open source, they do matching dollars. Now that "Okay, hey, we're going to contribute dollars to an open source foundation of your choice."
+
+So in that sense, lots of these -- so the beauty of these case studies is no matter where you are in your journey... Are you starting? Are you an expert? Because the wide range of case studies that have been incorporated into the book - there is something for everybody. And frankly, I reviewed each and every of these case studies. I did not write them, because they were written by 55 contributing authors... But even to date - and I'm sure over the next several months and years - whenever I read these case studies, it's like "A-ha! I missed this nugget. And this is how I can apply it to my day-to-day life." So I think in that sense, there are lots and lots of different case studies. So I would encourage readers to go through the book and say "Which vertical do you align with?" Oh, maybe I'm in the pharma industry, so I will take a look at Johns Hopkins. Maybe I'm a retailer. I'll take a look at Amazon. Maybe I'm a GSI. I'm going to take a look at that.
+
+I'm a device manufacturer. I'll take a look at the Dell case study. Or I'm into a technology company. Several examples in there. So I think in that sense, hopefully the diversity, the quality, and the quantity of case studies gives everybody a nugget that they can pick up and apply into their day-to-day work.
+
+**Jerod Santo:** \[32:18\] I think going back to the Why is really important and well said by you, because if you have a solid Why that aligns with your business incentives, everything else kind of makes sense. And if you don't, then you're kind of perpendicular and you shouldn't be doing this in the first place. And so if you can find the why right away... I think Swift is another good example; you brought that one up, Arun... Where it's like, Apple put a lot into Swift. And if they just wanted an awesome language for iOS or Apple platform engineers, they probably didn't need to open source it. And of course, there's non-open source languages of the past that are just fine for that. But if they want a general-purpose programming language that's adopted by everyone around the world, it had to be open source. So that was their why. It's like "We want this to be more impactful than merely the next best language..." Not the next best. THE best language that they think, for developing Apple apps. And so that's another good one where it's like, that was a huge decision. Because think of how much money they had in the Swift just engineering hours. It was like five years of just Chris Lattner working on it, and then they brought a team in, and then like all this stuff... And then they got all this stuff rewritten in Swift... So that's a huge investment that someone said "Yeah, we're going to go ahead and give it away to the world, in order to accomplish our business goals." And I think that's totally fine. It makes complete sense. Not everything is merely altruism. It's cool with Bloomberg that they have that in their mission statement or whatever, like "Philanthropy is one of our core incentives at Bloomberg." Not every business does that. But if you can find your why, then the details have to be figured out and managed, but they're just details.
+
+**Arun Gupta:** Correct. No, I think you're putting it very well. Different companies have different Why's, and each company has different Why's for different projects, as a matter of fact. But locking that down, writing it down is super-important, because oftentimes it's brewing in your head. So usually, when any leader comes to me that "Okay, we want to open source this project", my first question is "Why?" And there is corporate altruism, as you said, in some cases; more often than not, it's an enlightened self-interest, that "A-ha. Okay, this is where it's going to serve me, and in the process it's going to serve the world as well." It's totally a fair gesture, as a matter of fact.
+
+I think, Adam, you were asking... Intel contributed, for example, oneAPI. So we created this oneAPI that's going to be a unified API across different accelerator architectures. So we created this for our benefit, but then over the period of time we realized "Hey, this could benefit the world as well." So a couple of years ago, we created this thing called as UXL Foundation. Uniform Acceleration Foundation. So that's the foundation that we created, and brought a whole bunch of members together over there, and now that thing is taking a life of its own.
+
+So in that sense, the whole idea was, if this is indeed an API, a unified API, then everybody needs to partner in and everybody needs an equal stake at the table, so that they can define what the life of it's going to look like. Because if you want that collective engineering effort, there is no other way without going out in the open source, and that diversity of thoughts is what I really enjoy in the open source journey.
+
+**Adam Stacoviak:** \[35:42\] We're steeped in open source, of course, but how does this reflect into the other open source, a.k.a. inner source? I was just leafing through the GitHub case study really quick while we were having this conversation. I was just looking at how essentially their inner source -- GitHub began as a social coding platform... We all know GitHub, I don't need to tell you what that is. But their practice of inner source really became this sort of internal -- basically, open source behind the firewall inside GitHub. How does that manifest in a lot of these companies? Is that truly open source? What are they calling it? Is it licensed? How do you manifest inner source? What begins that? Is it just simply "Our software and you can contribute", or is it -- how does that manifest?
+
+**Arun Gupta:** I think the biggest reasons that I've seen inner source happening are multifold. First is you want to avoid duplication of code. How many loggers do you need in a world, right? In a company, if everybody is writing a same ETL app, there is no point doing that code duplication. So how can I create that discipline where inside the company, when you're not within the firewall, you're not going outside, it's only engineers from the company that can contribute to it, so hopefully you can be a bit more open. So in that sense, how do you publish that repo so that others in the company understand this ETL app is being built, and we can all contribute to that? Again, very much open source philosophy - that 80% of my task is already done with that one ETL app. I just need to put 20% more effort to customize it for my need. And if I contribute it back upstream inside the inner source repo itself, that'll do my job. So that's one.
+
+The second part of it is - you know, a lot of the companies have certain stricter requirements, what you can or cannot do in open source. Inner source basically is bringing that open source practice inside the firewall. So that's really a good way by which you can coach your employees, train your employees on all the open source skills, with much less risk. Oftentimes companies are under a trepidation that "Oh, you know what? If this person wearing my company hoodie commits something out in the public, it's going to look really bad. That one person will define the whole company", which is generally not true at all. Everybody makes mistake; giving people a chance to recover from the mistake is exactly what defines them to be more successful... But it is what it is. So inner source in that sense provides that learning opportunity platform, where they can practice all of the open source practices inside the company. And GitHub definitely has a very excellent case study on that.
+
+There is a case study by InnerSource Commons \[unintelligible 00:38:46.19\] over there as well, because they really call out in terms of all the fun things that have happened within Inner Source Commons. Because that provides common terminology, they run an Inner Source Commons gathering on a regular basis... They have lots of inner source case studies over there... So you can start looking at it on all the fun things that are happening in the inner source space.
+
+And then the third reason that I've seen it is you could see potentially inner source as a stepping stone. A stepping stone to going to open source. For example, a project started, you want to see how the internal collaboration is going to look like... Because oftentimes teams don't recognize, don't realize that "Hey, I've built a project. Before I release into open source, I don't know what level of support I would need. Do I need 30% more engineering if pull requests start coming in? How am I going to manage them? If people file issues, if they are not aligned with my roadmap, how does that work out?" So again, inner source kind of gives them that feeling as a stepping stone to going to open source, that "Okay, this looks like -- okay, I understand my engineering efforts are not going up by 30%. By only 10%. And maybe when I go to open source, they will double up to 20%." So it helps them with their resource planning, and their roadmap, and all of those things.
+
+\[40:10\] So in that sense, inner source is very, very effective. And Intel has a very large inner source practice. We have hundreds and thousands of repos here, which are run on an internal platform... A very nice nomenclature. So it kind of calls out, "Are you looking by a language? Are you looking by a technology? Are you looking by a business unit?" So you can start filtering your repos. And that really significantly improves the discoverability of the entire element that we were talking about earlier.
+
+So I think in that sense, inner source is definitely a key factor that allows you to be on that open source journey in a much more constrained environment, and giving executives that feeling that "Yeah, nothing is lost here. It's all happening within the company." And most importantly, giving them a comfort feeling that when these people go out, they're trained on these tools very well.
+
+**Adam Stacoviak:** Yeah, it seems almost like a proving ground for open source at large. It's like, if you don't know what open source is, or - Jerod, back to the question you asked before, like what are the hard questions or what are the pushbacks you get from management? I feel like if you can prove "This is how inner source works. Now imagine how it works for the world."
+
+Our safe world, behind our firewall, controllable, all those things... It's still unclear to me though how you begin. Is it simply like a private repo on GitHub? Is there a way to dashboard this stuff? How do you truly spark this? It seems like it's bottoms up, where it's "I've got this thing I think this team or these other teams could use." I still don't understand how you manifest inner source. Is it just behind authentication and authorization? Is that what it is?
+
+**Arun Gupta:** I'll give you two examples. One of them is from Apple. The IT team from there reached out that folks were recognizing the ETL example that I gave earlier, that there are six different ETLs that are happening. It's a waste of resource, waste of time, and a waste of engineering resource. Now, how do we consolidate all of them together? Essentially, what we did is from ground up we built up an inner source org, and we built up a nomenclature that "Okay, folks, as you are creating these new projects--" And even within the company, there could be sometimes a firewall, right? Like, oh, IT team only want to collaborate among themselves. Which is fine. Whatever the structure is. So we created an internal org on a GitHub platform, or whatever your Git platform is. We created an internal org and said that "Okay, folks, before you create a new project, take a look at that org." There is some discoverability mechanism, there is some tagging mechanism. You can start looking at dashboards in terms of how healthy the project is, how many other teams across the company have adopted it, what is the bug fixing rate, what is the pull request approval rate... Because that kind of shows the health of the project in that sense. Because if anybody is coming and looking at an open source project, that's what they would look at. Why not adopt the exact same practice inside the company?
+
+So that's sort of how we started, ground up. It took a while for us to gain traction. We ran an inner source event internally. A lot more curated people, because then they're coming to learn about inner source... And we had to intentionally and very deliberately build those relationships and connection. So it took a time, but then the practice kicked off.
+
+\[43:45\] I came to Intel... At Intel there was already an effort that was happening on that. So we have 100,000 plus repos. Those were all put onto the platform, they were all on a wide range of CI/CD systems, wide range of Git repos, wide range of orgs... So all of that was automatically then thought about that "Okay, how do we bring this into inner source again, improve the discoverability, cut down the cost on running this wide variety of CI/CD systems, cut down the cost by doing this wide range of orgs etc.?" So that really allowed us to minimize the cost, and then also improve the engineering time that is required. I already found 70% of my work done, and I'm just going to contribute to that project, and that really cuts on my engineering time. And now, I'm also able to exchange that mind, that knowledge with another engineering team who's having a similar issue. So in that sense, it continues to build that intellectual knowledge, as opposed to intellectual property for your work, and you continue to make progress. So I think there is no one right answer here. It really depends upon where you are in the journey and how you want to start about it.
+
+**Adam Stacoviak:** Yeah. I like that. I mean, it seems like some of the key characteristics are organization, some sort of org that says, "Okay, we're going to make this a thing. This is not a thing... We're making it a thing." That's easy. You mentioned nomenclatures... So how do we describe this stuff? And maybe each company is a little different with the way they describe it. Maybe there's certain terms across all inner source organizations, across different companies, so it just makes sense. And there's some sort of shared knowledge or shared nomenclature.
+
+And then you've got sort of visibility. I don't want to recreate the wheel... I don't want to go out to the open source land, because I can't play there. I've got to play in business user land, not open source user land. And so where can I find that stuff? And then gauge my own desires. Like, are they maintaining it? Does it solve my true problems? Can I actually contribute all these different things that you would consider?
+
+I love though how this manifests to say "These are the open source culture mechanisms you would care about." I love how that has the possibility, I suppose, of getting people to truly understand what open source at large works and how it works, versus simply just how it works internally.
+
+**Arun Gupta:** I think one of the things that we don't talk often about is as new developers are getting on board, they're coming fresh out of college. I mean, a lot of the developers understand open source really well, but for developers who don't understand open source well, who have that fear of failure that "Oh, it's going to spoil my career. It may not look good on my career", or whatever, one of the best keynotes I heard at KubeCon was this person said "Hey, I burned down a 10,000 node Kubernetes cluster at Spotify. And a week later I did that again. And my manager said, "Go give this keynote and how you recovered from it."
+
+So I think inner source in that -- I mean, it's a bold move, and I really love it. I admire it, because that's how I am. But not everybody is like that. So in that sense, inner source really gives that safe space for people to like "Okay, I'm going to refine my skills in my own comfortable environment, and I'm going to try this out. I played with the CI/CD system, I know how to git fork it, I know how to git merge it, I know how to rebase, and I'm learning internal tools..." So that's from the engineering perspective. On the other side, from the inner source team, doing that advocacy, providing those tools, running those workshops, running those internal hackathons, that "Okay, let's get together. Let's figure out what this project is. Let me guide you how this works." And that opens up the avenues and the opportunities for the engineer that when I go out in the public, it's very much like that. It's just that the person on the other side is from a different company.
+
+**Adam Stacoviak:** Mm-hm. That's cool.
+
+**Jerod Santo:** So you dedicate an entire chapter to internal events. You just mentioned a couple of things right there... And while we're talking inner, it was a surprise to me to see a chapter on internal events, because I'm not in the know, but apparently you are. And so tell us why that is like a whole chapter of your book, and why that's so important inside of organizations in order to foster this open source ethos.
+
+**Arun Gupta:** \[48:12\] I think it's super-important. And I'm, again, looking at the chapters here... The types of internal events that I talk about in that book is like workshops, where you are bringing people and teaching them that "Okay, let me tell you how to be an effective Git user." Hackathons, where you bring a project and say "Let's hack on this together. So now that you've learned the skill, let's apply this to a real time project." Because just the workshop and hackathons - just a very different nature. Bring those external guest speakers. These are luminaries across the industry that you can bring them on. That way you get to hear their story and you feel inspired about it.
+
+Internal projects, where you may say that "Okay, I'm just going to run a hobby project", and on a daily basis, I'm feeling the need that I need to do this again and again and again. So every knowledge that I've learned so far, I'm going to spin up an internal project. Maybe in inner source, as a matter of fact. Having those topical summits.
+
+I remember when we did the Open Ecosystem Summit at Intel a couple of years ago, there were about 2,500 people that showed up, and a lot of internal discovery happened. Like, "Oh, I did not know you were working on this project. I did not know you were working on this effort." So that really allowed them to connect with each other and then join hands.
+
+Very ditto experience when we ran the first Kubernetes Summit at Apple. There were 1,000 plus people that showed up at Apple to that Kubernetes summit. And Apple, as you understand the company culture, it's very walled across different views. But for the open source topic like Kubernetes, they were all able to share information, and they recognized that the technical challenges are very similar, because they know exactly how the Kubernetes is deployed, they know the internal terminology, the entire architecture... So they can talk at a very much elevated level in that sense. And that essentially brings that internal community. There's a particular chapter dedicated on how do you organize the hackathons. How do you go about doing those hackathons.
+
+So I think it's super-important, the relevance of those internal events, because those internal events help you prepare to be a better citizen out in the open source world, they help you bridge the gap, connect to the existing open source elements... So a lot of fun. I think it's very, very relevant. And I've seen the importance of these internal events all across. And in that chapter, there is case studies by BlackRock, Infosys, Intel, and SUSE in terms of how they have leveraged these case studies, how they have leveraged internal events.
+
+I remember SUSE talks about something like Hack Week, where they will have a week dedicated to hackathons. No work done during that time, but essentially bring a project in, hack on it... And that incentive by the top management is really big time, and it encourages employees that it's okay to do that. And that is what is fundamental.
+
+**Jerod Santo:** You mentioned the thousand people that showed up to that first Apple Kubernetes event... You remind me of an old show we did with a late great Dan Kohn. One of my favorite episode titles ever, Adam, if you remember this one. We named it "Kubernetes brings all the cloud natives to the yard."
+
+**Adam Stacoviak:** I love that.
+
+**Jerod Santo:** Because we were just so surprised at how many -- like, the growth of KubeCon, and just how excited people were for Kubernetes back in the late teens. It was amazing. Like, KubeCon doubled in size every year for the first five years, or something crazy like that. And it was probably a surprise at the time that a thousand people showed up, but it probably shouldn't have been, because man, people were just interested in Kubernetes, for sure.
+
+**Adam Stacoviak:** Yeah, for sure. That was a good title, Jerod. Good callback.
+
+**Break**: \[52:05\]
+
+**Adam Stacoviak:** One thing that -- you were talking about hackathons; I was just leafing through that section of your book there, and I want to pull out this... If you don't mind, I want to quote your words. You said "Hackathons and coding sprints are like HIIT." Is that HIIT? Yeah, it is.
+
+**Jerod Santo:** High Intensity Interval Training.
+
+**Adam Stacoviak:** That's right, High Intensity Interval Training.
+
+**Jerod Santo:** Okay, now we're going to get the athlete to come out...
+
+**Arun Gupta:** \[laughs\]
+
+**Adam Stacoviak:** That's right.
+
+**Jerod Santo:** Arun the athlete.
+
+**Adam Stacoviak:** "Hackathons and coding sprints are like HIIT training for open source developers. They provide time-boxed and concentrated bursts of engagement that push participants to deliver impactful contributions within short timeframes." I think that's a great analogy, because I've always struggled to understand how you can do a hackathon at a company. I get it for, "Hey, come" and it's sort of like -- because the thing we struggle with as human beings (obviously, we're human beings, right?) is that we desire and thrive in connection. You can't expect innovation or things to happen unless you connect with others. And I think this whole part here on like internal events is enabling me inside of a company to get to know somebody else. And even if I go and just meet one person, or two people, now I've got more connection in the company, I've got deeper roots, so to speak. So I'm not leaving anytime soon if I feel connected, if I feel invited. And then you take a thing like a hackathon, which I was like "How do you do that well at a company?" Well, you're just sort of taking the inner source you may already have spun up, and you're saying "Go out, have fun. Meet people. Do fun things. Innovate."
+
+**Arun Gupta:** Correct. When I was writing that chapter, I was thinking in terms of "What do I do in my daily life? "And Jerod, thank you for the reference. I've been a long time athlete, and this morning, as a matter of fact, my workout was a HIIT workout. I'm a lifetime runner, so I've been running for a very long time... But in order to maintain that running -- you know, you can't keep running every day, because that leads to a burnout. Some people do, which is great for them... But for me, I want to run, but then I want to also get equally strength-trained, so that I can be a more effective runner. And if we bring that analogy essentially to open source, that's exactly what it is. In open source, new tools keep coming all the time. And in order to understand those tools, that high intensity interval training is what is required. Because what I'm doing is for one hour, I'm all in; I've cut down all noises and distractions from my life. I just want to learn this tool. I want to fail in that tool again and again and again, so that I understand the way this works, the way it doesn't work, and what are the boundary conditions, what are the normal -- happy path, all of that. So once I understand that, then I can go back to my normal running, because then I can say "A-ha, I'm stronger in this area and I can apply that skill into my day-to-day life."
+
+So I think it's super-important in that sense, on taking that intense exercise. It could be a workshop, it could be a hackathon, whatever it is. And in hackathon - again, Adam, as you called out, it's really a good way to engage with other people. Because no matter which business unit you are from, you are sitting together for a common purpose. And that pair programming concept is very extreme in that sense, right? Because you say "A-ha, I'm thinking it this way." And that diversity of thought immediately comes on to you, and then you are able to make progress on that together. Just the fact that, "How do I do a code review? Or if somebody has done a code review on me, how do I accept that feedback?" Because don't take it personally, take it in the technical sense that "Yeah, this is a really genuine feedback" and accept it.
+
+One of the things I talk about often is conflict resolution. It's a big one in open source. So often we get confused between personality conflict and task conflict. That "Oh, this person has given a feedback, he lives in this part of the world, and that's why he must be giving me this feedback." Keep that aside. It's not about personality conflict. Think of it as a task conflict, learn that element, and see how that makes you become an effective coder, is what this all should be about.
+
+**Jerod Santo:** Yeah, I feel like we could have an analogy for the fitness junkies out there. Like, you have HIIT classes like a hackathon, Pomodoro, that's like Tabata... Coding bootcamp... Well, I guess that's just a bootcamp. That already has an analogy. But we could have a bunch of these. I like that.
+
+**Adam Stacoviak:** Well, that's somebody coming in and educating them what to eat.
+
+**Jerod Santo:** There you go.
+
+**Arun Gupta:** Yeah. See, I mean, in the book - I'm looking through the chapters here... So for example -- because I wrote this a while back. The hackathons are like HIIT. The guest speakers are like the mentorship and coaching that athletes receive from seasoned experts.
+
+**Jerod Santo:** Okay. Yeah, sure.
+
+**Arun Gupta:** \[01:00:04.04\] That's what guest speakers are. Athletes build muscle memory through repetitive training, and hone their skills through simulated game scenarios in practice. That's your internal projects, essentially. And topical summits really dig deep into a particular topic. And these are like-minded individuals, experts that get together. And these are more like "Hey, I'm going to a bootcamp here", and all sitting together, really discussing that topic. So I think, yes, in that sense, you're right, Jerod.
+
+**Jerod Santo:** \[laughs\] Well, you beat me to it. I didn't know you already went through such a fleshed out exercise. But yeah, good metaphors, for sure, for those interested in such things as fitness.
+
+**Adam Stacoviak:** That's good for you. It's good for you.
+
+**Jerod Santo:** Not for you, Adam.
+
+**Adam Stacoviak:** Oh, I like fitness. I'm down with the fitness.
+
+**Jerod Santo:** \[laughs\] So Arun, share more of your fitness bent. You said you run a lot... Give us some stats. How long have you been running? How far do you run? How fast do you run? Etc.
+
+**Arun Gupta:** I've been running for 40 years. For a long time. For a very long time. I've done 35 half and full marathons around the world.
+
+**Jerod Santo:** Wow.
+
+**Arun Gupta:** I think 2021 was my most running year. I ran 2000 plus miles that year, which is an average of five and a half mile every day. So that was the most extreme running year. Now I'm on a much more easier scale. It's about 10K every four or five days a week, and the other two days are lifting. So I do work out every day. The days I don't work out, I feel very -- like I'm not very myself. So I work out every day... This morning I had a seven o'clock meeting. That means I got up at five, did my one-hour workout... I have a full pull-up cage in my garage. It was a full-on 55-minute workout. I pushed to the limit, all sorts of stuff. Box jumps, ball slams, bench press, pull up bars... All sorts of crazy stuff. So today was my running day because it was raining, but otherwise -- like, yesterday it was 10K, the day before it was a 10K... And all of those stats are available on my Strava.
+
+**Jerod Santo:** Oh, that's cool. So you do that all by yourself, or do you have a partner in crime? Sometimes for me I get more motivated when I have an accountability partner, or something.
+
+**Arun Gupta:** I'm all by myself. My schedules are very early in the morning, and I've got to go drop my son to school, get the breakfast ready for the whole family in the morning... So I do like really early in the morning.
+
+**Jerod Santo:** Nice.
+
+**Arun Gupta:** Most of the days I'm getting up around 5, 5.15, I hit the road by 5.30 or so, do an hour workout, and back in time for the first seven o'clock meeting.
+
+**Jerod Santo:** Very impressive. Very impressive.
+
+**Adam Stacoviak:** Yeah. I find if I have -- I like an accountability partner, Jerod, but I find if I have to coordinate with somebody else, it slows me down. I can't -- I've got to spend an extra half hour, or just some buffer in there to deal with the "Hey, how you doing today? Let's get going."
+
+**Arun Gupta:** Exactly.
+
+**Adam Stacoviak:** You almost need somebody who's like super-strict. Don't talk to me... Don't even say my name. Just get to the work.
+
+**Jerod Santo:** Yeah. You want somebody who's in the same mindset as you are, which is like "Neither one of us want to be here, but we know we should be."
+
+**Adam Stacoviak:** I can't find that person. That's what I'm trying to say. I agree, but I can't find that person.
+
+**Jerod Santo:** For me it's like if I'm going to get up early, I have to have something to where if I don't do it, I'm letting somebody else down... Because I'm fine with disappointing myself. I do it all the time. But to disappoint someone else is like just harder on me... So I'll get up for them. But not for me. Arun, apparently you're self-motivated, I guess.
+
+**Arun Gupta:** \[01:03:58.19\] \[laughs\] Yeah. One of the very exciting things we did last year was we went -- I took the entire family to Kilimanjaro, and we all hiked the summit. So my younger son and I are big into fitness, but I trained my older son and my wife... But I'm really, really grateful that we could actually summit Kilimanjaro. And then literally in about a couple of weeks we are again going on scattered trips, but we are going to Patagonia. Well, my son and my wife are going to Patagonia, and then I'll be joining my son on a beautiful expedition to Antarctica.
+
+**Jerod Santo:** Wow.
+
+**Arun Gupta:** So all of those really require a high level of fitness. I'm grateful that we're able to do that.
+
+**Adam Stacoviak:** Absolutely.
+
+**Jerod Santo:** So provide a comparison for us... Difficulty and pain of summiting Mount Kilimanjaro, or authoring a book.
+
+**Arun Gupta:** Oh, authoring a book was easy. Authoring was very easy.
+
+**Jerod Santo:** Really?
+
+**Arun Gupta:** Yeah. And my experience -- see, frankly, in 2023 December is when I did all of my writing. I pretty much did all of my writing within two weeks.
+
+**Jerod Santo:** You just cranked it out.
+
+**Arun Gupta:** Yeah, yeah. Just two weeks --
+
+**Jerod Santo:** Most people say that writing a book is like the hardest thing they've ever done.
+
+**Arun Gupta:** No... No.
+
+**Jerod Santo:** No. Weaklings.
+
+**Adam Stacoviak:** It's too easy.
+
+**Arun Gupta:** I mean, this was my seventh book too, right? So I've done that for a while.
+
+**Jerod Santo:** Oh, this is your seventh book. Okay, so you have some experience.
+
+**Arun Gupta:** I've been writing for a while, and particularly for this book, I've always been practicing this. So it was more about structuring my thought and putting it into a chapter format. A lot of the heavy lifting I would say was really working with those 40 plus companies, 40 plus case studies, getting their management approval, PR approval, helping them build a use case, and all that; kind of refining their case studies. That's where the biggest heavy lifting was. So that's where we spent a lot of our time. Kilimanjaro was easy as well, as a matter of fact, because I run regularly, I lift heavily...
+
+**Jerod Santo:** What's hard for you? Tell us something what's hard.
+
+**Arun Gupta:** I think I take things as they come, and just live in a peaceful, mindful manner. I would say the only hard part was the last day, where they woke us up at 10 in the night, and we hiked all night to reach the summit at seven in the morning, and then come back the same day. So we started like 10 p.m, we hiked up, we came down, took an hour break for lunch, and then hiked further down at a lower elevation. So it was about 5,000 feet up and about 6,000 feet down in the same day, in a stretch of like about 16-odd hours. And that was tiring. We slept like a log after that.
+
+**Jerod Santo:** Down sometimes is harder than up, isn't it?
+
+**Arun Gupta:** Down sometimes is actually -- down is always hard. A lot more impact on your knees. And the grade is pretty intense, so I think that took a bit of a toll on the body.
+
+**Adam Stacoviak:** I feel like the down is almost more dangerous than the up sometimes, because you can -- you almost have a faster pace because you're naturally being assisted by gravity. But then your knee buckles, your ankle twists, your sure footing may not be sure footing...
+
+**Jerod Santo:** You slide down on your butt...
+
+**Adam Stacoviak:** Yeah.
+
+**Arun Gupta:** Yeah, yeah. And particularly because when we were going up, it was all during night. Because 10 p.m we got up, 11 p.m. we started the hike, and we didn't see sun until 6:30, 7 in the morning. So it was all hike during the night. So with my headlamp, all I'm seeing is the shoes in front of me, and I'm just chugging along as they're going through \[unintelligible 01:07:39.29\] whatever. So you just chug, chug, chug.
+
+**Jerod Santo:** Well, why did you start that late? Did you want to see sunrise, or?
+
+**Arun Gupta:** Oh, the whole idea was to see sunrise at the summit. That was one. And the second reason was you don't want to stay at the summit, because Kilimanjaro is 19.3441 feet... So at that elevation, usually afternoons are a bit tricky; the thunderstorms can appear. So they wanted you to get back from a higher elevation before noon kicks in.
+
+**Jerod Santo:** Gotcha.
+
+**Arun Gupta:** \[01:08:13.23\] Going back to the book, one of the elements that we missed talking about is the relevance of open source foundations. And that's pretty big, actually. We kind of touched upon that topic earlier in the podcast... Open source foundations like Apache, Eclipse, Linux Foundation of course is big... And I'm myself on the CNCF governing board and OpenSSF governing board, and the chair for both the governing boards.
+
+It's very important that enterprises understand the relevance of joining these foundations. And there are different tiers of membership... Because joining those foundations really helps you understand all the good things that are happening in the foundation, connects with similar minded leaders... And again, it becomes a hiring magnet, essentially. Why Intel is joining over there? Intel really cares about the open source culture, and other open source leaders get influenced that "Okay, I want to work at Intel and I want to build my open source career."
+
+I remember when I joined Amazon, one of the first events we did, we got up on the re:Invent stage and we said "Hey, build your open source career at Amazon." And this was back in 2017, and that got a lot of applause, because essentially, Adrian Cockcroft, my manager who hired me at that point of time, he was big into open source. I've always been big into open source as well. Bringing two large open source leaders that are credible in the open source community and saying "Hey, we're joining this --" I wrote the six-pager for Amazon to join Cloud Native Computing Foundation, and bringing that credibility and saying "Hey, we are joining in there" really goes a long way in, again, fostering that open source culture.
+
+**Adam Stacoviak:** I'm glad you mentioned that, because I think foundations are really important. It's always a grab bag of how you perceive membership, right? Like, you might say "Oh, this is necessary because you've got to support it in cost", but then you think "Well, is that pay to play? Does that get me a seat at the board, so to speak? Does that give me access no one else gets?" And the usual answer, the obvious best answer is probably no, unless you're at the wrong foundation, and that's just like a different foundation...
+
+**Arun Gupta:** Correct.
+
+**Adam Stacoviak:** But I think that's really important to bring up, because I think that the foundations - they give the neutrality, right? They give the neutrality, the formation... And gosh, what has happened with the Linux Foundation over the last 15 years, I'd say, has just been tremendous. I feel like there's some angles where it could be good or it could be bad, like centralized control, or centralized organization, so to speak... But a large majority of the most important open source work is being done under the Linux Foundation. And that could be seen as a good thing and potentially as a bad thing for centrality, in terms of centralizing things... But I think it's mostly a good thing that we've got a worthy and organized body that can do that and have such great structure. They've proven to do it well over the many, many years.
+
+**Arun Gupta:** That is correct. I mean, different foundations operate very differently. They play a very critical role in terms of growing these open source projects. They provide this central functionality like CI/CD, security audits, legal support, marketing support, event support... So a lot of advantages of these open source foundations.
+
+Some foundations give you a seat at the governing board, depending upon what tier of membership you take. Most of the foundations that I'm aware of, they usually have their technical oversight committee or technical advisory committee. Those seats are all elected. Those are not pay-to-play at all. The maintainership is, again, meritocracy, so douocracy, and how do you rise up to become a maintainer. So in that sense, the administrative part and the technical parts are separate. Like, I'm on the governing board, so we are responsible for the administrative, financial, legal aspect of it, but we have zero say in terms of what the technical roadmap of the foundation is going to look like...
+
+\[01:12:23.13\] And I like that separation of concerns in that sense. But if the technical folks need some budget support, for example in OpenSSF, they come over to the board. But we're also trying to create swim lanes over there, where a certain level of funding is not required to be seeking permission of board every time, because we want to create swim lanes where people can go faster on their own.
+
+**Jerod Santo:** Can you tell us more about douocracy? You just said douocracy. What's this?
+
+**Arun Gupta:** Yeah, so think about the concept of you do things in open source, and that really helps you rise up in open source. And it really could start with just being a lurker on an open source project, that "Okay, I'm just observing what's going on in the open source project." Then occasionally jumping in, that "Okay, hey, I'm participating in the working group", and then I'm starting to participate in the conversation, or I'm jumping into the Slack channel. And first maybe I'm asking my questions, but then slowly, as I gain more knowledge into the project, then I can start answering other questions.
+
+Then maybe I jump into sort of the overall aspect of looking at other people code, reviewing people code. Then maybe I jump into the element of "Hey, I'm going to start contributing code." And then I'm contributing more serious code. Maybe I'm contributing deeper sections. And eventually, that douocracy is that whole contribution ladder where you rise up from being a user, to a contributor, to a committer, to a maintainer, and different projects at different hierarchy levels or different contribution ladders. But that douocracy is the more you do into the project, the more you are known into the project, and the higher your chances of getting a leadership position is.
+
+It's not like "Hey, Intel is part of the CNCF governing board, and Intel must have a maintainer into the project." No, it's purely, purely douocracy. You do the work into the project to rise up to that level of maintainership.
+
+**Jerod Santo:** I like that, douocracy.
+
+**Adam Stacoviak:** How does one begin? So we've got a diverse listenership in our audience... I would say from individual contributor, to leadership, and everywhere in between. What is a good on-ramp? I know you probably described some versions of this in your book, but what are the best ways to begin this culture shift of inner source, open source, organization etc. if it doesn't exist? Or maybe it's been tried before, but it hasn't been successful. What are some ways to begin, for our audience?
+
+**Arun Gupta:** Yeah, one of the last chapters, that's what I was looking at, in the book... It's really about getting started on your culture change journey. A lot of the times if you continue to do things in a certain way that you've always done, and if you're not challenging the status quo -- because change is the only constant. If you're not challenging the status quo, then -- that's the first thing you need to do. How does open source work for you? You need to overcome your legacy mindsets. That's an important element, and there's a chapter in the book that talks about that part of it. And then the book really gives you like a 10-step framework on how you can jump onto that cultural journey. I mean, the diverse audience is great. Having a top-level executive sponsor - because you need that somebody either at the executive level, top executive level, that says... And it doesn't have to be at the ELT level, or the CEO must say that open source is our strategic direction. It could be like "Hey, this is a group where a lot of the open source technologies happen", and that executive stands up, that "Yes, this is how I'm going to define the culture of my org, and I'm going to actively encourage, allow them to contribute to open source. This is what the incentive mechanism in the companies are."
+
+\[01:16:22.00\] Identify your stakeholders. Not just within your team, but outside your team. If there are other business units that are doing similar work, if there are folks in marketing that care about, if there are folks in recruitment that cares about. Because at the end of the day, they want to bring the top talent over there. So how you can start connecting with that.
+
+You need to have -- I remember when I joined Apple, I think second or third day, I was having a call with somebody... And my God, that person was screaming at me. "I submitted this pull request so many weeks ago! It has not been approved yet! It's been stuck into legal!" "What can I do to help?"
+
+**Adam Stacoviak:** Very calmly, yeah.
+
+**Arun Gupta:** It was just a listening session, a venting session, essentially... Let that frustration come out. And it's very important to just listen with it. Again, we talked about task conflict/personality conflict. Think in terms of task conflict. What is a task not getting done? Because that person has probably been trying really hard for many weeks, months, and is not getting through... Just understand the task, just get going on that.
+
+Identify sort of the non-core open source areas, or maybe the core open source areas where you want to make attraction. Define those walls within which you can do more open source or less open source. I think that's an important area. Create those working groups, for example. At Intel we have several of those working groups where we teach on the best open source practices. Anybody want to go to an open source project, they can come for education enablement. OSPO does a good part of that, essentially. Definitely having an open source program office goes a long way.
+
+I don't want to give away all the tips from the book, so I definitely encourage you to take a look at the last chapter of the book that gives you that entire framework on how you can get started with it. And there is a section in that chapter which talks about sort of my journey on how I went about building that exactly in these different companies.
+
+**Adam Stacoviak:** I was thinking as you were reading that, "Arun, don't give all the goods away", but at the same time, will this be -- will this be the book that gets handed to someone when they say "How does open source work? How can I make it work for my company? I hear about this thing called inner source... What are the frameworks?" I feel like you've, in two weeks - as you said, you wrote this book quickly, in two weeks. And a lot of the work was case studies and external work and whatnot. But I feel like you've really organized a lot of your wisdom and thoughts over your decades of experience to really give us the initial steps, if not the full-on steps of what it takes to change the culture inside of companies, whether it's inner source, or open source, or whatever.
+
+\[01:19:03.04\] So I think you've done a tremendous job. I can't wait to read the entire book, personally. I've only leafed through the table of contents and some of the stuff with you along this conversation here in my morning here... But I think this is a great framework from what I can see. I mean, I'm really proud of the work you've done here, and I appreciate this being in the world to give that guidance. Because crowdsourcing is the plain term, but open source is the future innovation. I think, actually, if I can leaf through it quickly... Page two. I want to read this, if you don't mind. You might know what I'm not going to mention, but on page two you mentioned Mark Andreessen's really famous phrase when he said "Software is eating the world." And then Jim Zemlin, the executive director, and I believe the founder of the Linux Foundation, he said "Most of that software is open source." And then of course, our very own Arun Gupta says here in 2024, "Open source culture sustains innovation."
+
+And I feel like what's happening in LLMs, and AI, these models being open source... We're seeing a massive change in accessibility to software, accessibility to models, accessibility to all these things... And the innovation that we're seeing in our world. I mean, open source has won. And that innovation - you've given a great blueprint and on ramp for so many organizations to say "How can we really be a part of this, or put it to work for our company, or to truly understand what it is?" So I really appreciate this work you've done here for the open source innovation that will come.
+
+**Arun Gupta:** Thank you. Thank you. I'm really excited about how the book came about to be. I really hope this book is a fire starter for people who want to jump onto their journey, kickstart their journey... Or if you are far along in your journey, look at those case studies and see what other fun things you can find out.
+
+**Adam Stacoviak:** Anything left we have not asked you? I know that you've got a lot to share. Have you shared at all? What's left? You can even plug something if you wanted to.
+
+**Arun Gupta:** No, I think it's real good. I think in terms of everything around the book, we have talked about it. Open source culture is such an interesting topic, and I've been working on this for such a long time that every time you talk, a new story comes out, a new perspective comes out. And I would like people to think about, "Would you go to a potluck and not take a dish of your own? Would you go to a communal garden and just eat all the vegetables, and not plant your own vegetable?" So that's why open source is like a potluck, or a communal garden. So contribute, make it sustained, for yourself and for others, and that's the critical element.
+
+**Jerod Santo:** A perfect note to end on. Thanks, Arun.
+
+**Adam Stacoviak:** Thank you, Arun.
+
+**Arun Gupta:** Thank you so much for having me here

LIVE from Denver with Nora Jones! (Interview)_transcript.txt

@@ -0,0 +1,539 @@
+**Jerod Santo:** Hello, and welcome to the lovely Oriental Theater in Denver, Colorado. Thank you so much for being here. I'm Jerod...
+
+**Adam Stacoviak:** I'm Adam.
+
+**Jerod Santo:** ...and we've been doing the Changelog Podcast for a very long time, but we've never done this. So thank you for coming, this is truly a special time... And we wanted to have a special guest with us, a Denverite, and a really cool person... And luckily, we found Nora Jones. Let's hear it for Nora.
+
+**Nora Jones:** Thank you.
+
+**Jerod Santo:** Nora is the founder and CEO of Jeli, which is an incident management and learning platform. Then we had her on the show a few years back, in an episode called "Learning from Incidents." That was February, 2022. Since that time, Nora and the Jeli team have been acquired by PagerDuty, where you now work...
+
+**Adam Stacoviak:** Love that.
+
+**Jerod Santo:** And your title's really cool.
+
+**Adam Stacoviak:** And long.
+
+**Jerod Santo:** And long.
+
+**Adam Stacoviak:** Wait for this.
+
+**Jerod Santo:** So I'm going to read it for us all. You're a senior director and the head of pricing and product strategy and growth.
+
+**Adam Stacoviak:** That's three jobs.
+
+**Nora Jones:** Yeah. Too many hats.
+
+**Jerod Santo:** \[laughs\] So let's start with the acquisition. So tell us about Jeli... You're part of a much bigger organization now, but just briefly the Jeli journey, starting it, and then take us to the acquisition. I'm sure you could do that in 30 seconds.
+
+**Nora Jones:** Yeah, absolutely. So I started Jeli in 2019, and prior to that I had spent my entire career as an engineer dealing with incidents. I was working at Netflix, I worked at a company called Jet.com, which got acquired by Walmart... I worked at Slack for a bit... And I was kind of doing similar roles in each of these companies. And what I really learned through these companies was the power of post-incident reviews, and how they could really rally an organization to drive change, and focus on understanding other people in the organization's perspective. And what I really started realizing was how much could be learned from how we speak to each other during incidents... So how we talk to each other in Slack, when there's gaps in communication, who we bring in, why we bring them in, what they're doing... And that can really be used to build more expertise within organizations, and really understand where discrepancies and misalignment lie in a productive way.
+
+So Jeli was a product that we built based on real problems I was experiencing. It was really a tool I wished I had had in some of those situations... And so that was where it was born. And I'm very, very nerdy about incidents and that topic, and reliability, and I've been a PagerDuty user my whole career... And so the acquisition came up; it wasn't something I was looking for, but I've always admired PagerDuty... And I actually was like "This could be really good for the industry, actually having these products together and interconnected." So it made sense to me. And yeah, here we are.
+
+**Adam Stacoviak:** How did you feel about the acquisition process? You said you didn't plan for it... How did it come about? Did they reach out to you? Were they fans? Were they users?
+
+**Nora Jones:** Yeah, no, I mean... So the first two integrations that we built for Jeli, like before we even had customers, were Slack and PagerDuty. So we were like really tightly integrated. And the integration we had with PagerDuty was getting on-call schedules. And we were using the on-call schedules to figure out if people were actually supposed to be there during incidents, or if we were pooling in people that weren't on-call. Because that can indicate there's a knowledge silo, an expertise gap, if we're constantly relying on people that the business didn't actually plan to be there. So that was the first integration that we built with PagerDuty.
+
+So I spoke at one of PagerDuty's conferences in - I think it was 2020. It was like their first remote conference... And so we had always kind of been friendly, and maintained a relationship for a few years. And then they reached out and asked if like we wanted to have more serious conversations. And so I started exploring it, and it just kind of made sense. We had been doing Jeli for five years, and it was either we develop our own on-call solution and expand that way, or we join the most powerful player in the industry with that. And so it seemed like a really unique opportunity.
+
+**Adam Stacoviak:** \[08:12\] When you looked at that acquisition, as a founder, you think exit, right? Because that's a good thing. You built something, you want to profit from that, you want to see your work be recognized... Paydays are great, obviously. But how do you also think about the -- one thing you said in your early part of it was how you thought about the product and what the industry could leverage from the fact that Jeli moves to PagerDuty. They've got a much bigger distribution than Jeli did, so it's obviously going to be good for the industry. How do you weigh that choice between personal gain and the industry of people you know and love, and the product they need to succeed? How do you weigh the difference when you consider that acquisition?
+
+**Nora Jones:** I mean, I don't necessarily think those are separate things, you know? Like I mentioned, I built this product because it was something I wished I had had in my jobs. And so that was really like kind of my North Star, was getting it out to everyone. And so after the acquisition, seeing the usage of something that I had built get automatically included in people's toolkits when they start a job was what I had really dreamed about. And so I think it was mutually beneficial from that perspective.
+
+**Jerod Santo:** Does Jeli as a platform still exist as part of PagerDuty today? Is it a tab, or is it a subsection?
+
+**Nora Jones:** Yeah, so it's within PagerDuty. So one of the things -- so I spent the first year helping integrate the platform, helping integrate the cultures... It's very different bringing a 20-person company to a 1,200-person company. And some of these people had never worked at an organization bigger than 30 folks, so you know, there was some adjustment there. And then one of the things I noticed was within PagerDuty's pricing and packaging there were parts of the incident lifecycle that were almost limited to the enterprise plan. And so they immediately put Jeli on the enterprise plan, but I started realizing there's really more of the incident lifecycle that should go downstream, to all plans. And so I launched that project within PagerDuty, so Jeli is actually available to all plans in PagerDuty right now.
+
+**Jerod Santo:** Is it called Jeli?
+
+**Nora Jones:** What's that?
+
+**Jerod Santo:** Like, do they opt into Jeli, or is it called something else now?
+
+**Nora Jones:** It is called Jeli in the user interface. It'll probably eventually just be called post-incident reviews, but you know...
+
+**Adam Stacoviak:** Well, you're in charge of product, so you can change that if you want to, right? You can keep it or change it.
+
+**Jerod Santo:** Just keep the name. Oh, you're saying you're in charge of product, so you can decide if that gets renamed.
+
+**Nora Jones:** \[laughs\] Yeah. I mean, I think at first it made sense to keep the name, but I think it evolves over time. I think if the name was something more relevant to what it did, it would probably make sense to keep it. But if someone's in their PagerDuty UI and sees the word Jeli, they're not going to know what that means.
+
+**Jerod Santo:** Yeah, totally.
+
+**Nora Jones:** As much as I would hope they do.
+
+**Jerod Santo:** We focus on the name, but I think it's a proxy for this handing off of a thing that you built. And the attachment as somebody who built something, that you have for a company or for a product. And the name kind of represents like an era, or the thing, right? Jeli is the thing that you built... I ask these things because I'm always curious about the emotional attachment of a founder to the thing they built... And sometimes you have to like part with that. It's like "I'm selling the thing, it's gone now." But it sounds like this is a good case, because it's going to continue on, maybe with a different name, maybe not. But you didn't really have to let go.
+
+**Nora Jones:** \[11:57\] Yeah, it's integrated. There's more integration that's being worked on. It has a product manager managing it; I handed off my baby... But it felt like the right thing to do, you know? And we have a lot of plans for it too, so I'm excited to see how it evolves.
+
+**Adam Stacoviak:** What was the response from PagerDuty customers? As you said, it became an enterprise first, and you're like "No, it needs to be for everybody." And then so you expanded to all the plans. So every PagerDuty customer has access to Jeli, whether you keep the same name or not... What was response to have this -- I know it was already out there as a company, but you've obviously tapped into maybe an audience that never heard of Jeli, or heard of it less, or maybe took it less seriously... But now you're acquired by PagerDuty, the biggest in the industry... I mean, we call it PagerDuty because it's pager duty, but it's also a company.
+
+**Nora Jones:** Yeah.
+
+**Adam Stacoviak:** So what was the response like from the existing customer base of PagerDuty having this new tool?
+
+**Nora Jones:** Yeah, so when we got acquired, 90% of our customers were PagerDuty's customers. And the very first customer I sold to with Jeli - giant enterprise company, which is very rare for startups... And so all of a sudden, most of the companies that were using us were large enterprise companies, which is a lot for a tiny startup. You're kind of trying to act a lot bigger than you are. Someone wants to talk to the salesperson, I'm like "Let me go put on this other hat, and I will be your salesperson for a second."
+
+**Adam Stacoviak:** Ed Chambers?
+
+**Nora Jones:** Yeah, exactly. But yeah, and I just actually lost my train of thought.
+
+**Adam Stacoviak:** Sorry about that.
+
+**Jerod Santo:** Good job, Adam.
+
+**Nora Jones:** No, no, no. I forgot the question.
+
+**Adam Stacoviak:** Well, you were saying putting on different hats. You were going from like founder, to sales to janitor...
+
+**Jerod Santo:** Because of enterprise customers.
+
+**Nora Jones:** Oh, you were asking how customers reacted.
+
+**Jerod Santo:** Yeah.
+
+**Nora Jones:** Because we shared so many of the same customers, those customers reacted really well. They were already using both of us... So it's like a new integration, that will just make it work even better together, eventually like one piece of paper, no multiple contract cycles... And then the other thing is PagerDuty was originally created to page folks. And PagerDuty has done a lot more than that over time as well, but I think we were also filling a critical area of the platform that hadn't been invested in much. And so a lot of customers were like "Oh, this has been on our user voice request list for years, and now, it's finally happening." So yeah, we've had a we've had a positive reaction.
+
+**Jerod Santo:** How did it work once you were on the inside with regard to your personal drive and initiative? Because you've kind of been to the top of that mountain -- like, starting a startup and building and scaling and getting enterprise adoption and then making a sale like that... That's like a huge accomplishment. And now you're inside the behemoth, right? The PagerDuty. Is it like "Well, now I'm going to climb the corporate ladder?" Or how did you think about it? Because you have a new title, I noticed, since the one that you got when you first joined. So you've climbed, it seems, a certain bit. But is that what you decided for your own ambition's sake? "I now want to move up PagerDuty", or "Now I want to have Jeli rolled out to more people"? Or how do you look at that side of growth for you?
+
+**Nora Jones:** \[15:37\] Yeah, I mean, I've been reflecting on this a lot recently, and there's a lot of stages of grief involved in it, too... But I had been an engineer my entire career prior to starting Jeli, and so I actually had grief from being a CEO and not being in the code. I would want dopamine hits, so I'd go update our marketing site every now and then, just to feel like I was doing something, even though I was doing all these different things... But I also realized part of the reason I started Jeli was after incidents in an organization, a lot of the time every role is impacted in some way... But also, a lot of the time if there is a post-incident review, it's just done by engineering. And they're not going and talking to marketing, or sales, or customer support necessarily. Like, it's usually done in a vacuum. Or even executives. And I started realizing, as an executive, as the CEO of Jeli, that I had my own sharp end. I had my own area of expertise that -- you know, when I was an engineer, I always kind of felt like "Oh, executives don't get it. They don't get--" But then when I was an executive, I was like "Oh, I didn't get it back then." I was able to see a different side of the coin. And so I'm bringing all this up because now I get to be in a new role in a large organization, where I've worked at large organizations before, but always as an individual contributor engineer. And so from my perspective, I'm like "Oh, I'm learning another side of this coin."
+
+So it's really interesting to me to have played all these different roles... So that's kind of where I'm at right now, is like I'm learning a lot. I'm getting to run pricing at a large public company right now, and it's fun, and it's challenging, and I'm just kind of enjoying where I'm at right now. I'm taking it a day at a time.
+
+**Adam Stacoviak:** So you have pricing, product, and growth.
+
+**Jerod Santo:** What else is there...?
+
+**Adam Stacoviak:** That's three jobs, right? But I mean, you're over it all. So you're over three divisions, or three different silos...
+
+**Jerod Santo:** Or are those just goals? What are those things? Are they goals? Are they divisions? Product, pricing, and growth. Are those mandates for you, or how does that manifest?
+
+**Nora Jones:** Yeah, good question. So the role I have is head of product strategy and growth, and I think that product strategies should really drive pricing, and drive how you organize yourself, how you ship your product... And so I'm really helping with the forward-looking roadmap, along with my colleagues, and taking input from them, and then leveraging that to drive pricing. So I do have those three different units under me, and I am wearing a lot of hats right now. I am hiring, so if you know anyone, please let me know.
+
+**Jerod Santo:** And is Denver like a locale, or is it all remote and you just happen to be in Denver?
+
+**Nora Jones:** It's all remote, I happen to be in Denver. There's like 50 PagerDuty employees that live in Denver.
+
+**Jerod Santo:** Really?
+
+**Nora Jones:** So there's a lot of folks here.
+
+**Jerod Santo:** Is there an office or anything?
+
+**Nora Jones:** There isn't. We always say our office is the brewery. \[laughs\]
+
+**Jerod Santo:** There you go. Let's go meta for a second, and give you an opportunity to get a drink. I've seen you glance at your drink a couple of times...
+
+**Nora Jones:** Sure.
+
+**Jerod Santo:** Lowercase m, meta.
+
+**Adam Stacoviak:** Not the real Meta.
+
+**Jerod Santo:** I still despise them for trying to steal one of my favorite words, meta... Grok and Meta have both been stolen by billionaires, and I'm not happy about that. In podcasting there's an interesting psychology to it, because intellectually you know there's going to be listeners later... And whether it's 5, or 5,000, or 50,000, you come into it and you're like "This is going to go out to people." And so there's that first five or ten minutes where you're thinking about that... And there's a really cool thing that happens, because in reality it's just three people on effectively a Zoom call. And those 5 or 5,000 or 50,000 people aren't there... And so you kind of just forget about it. And it's neat, because the conversation becomes very relaxed, and you might forget there's an audience... In fact, sometimes people afterwards say "Hey, I can't believe I said that. Can we X-nay that, or whatever?" That doesn't happen on stage. That's my meta point right there.
+
+**Adam Stacoviak:** \[20:08\] That's the meta point.
+
+**Jerod Santo:** I'm very well aware that there's some people here...
+
+**Adam Stacoviak:** Y'all are there, right?
+
+**Jerod Santo:** We're happy you're here...
+
+**Adam Stacoviak:** We can barely see you.
+
+**Jerod Santo:** I'm reporting back that so far it doesn't feel yet like just the three of us sitting in a room... I guess because we're not \[unintelligible 00:20:17.11\]
+
+**Adam Stacoviak:** In all honesty, I love this format. I'm loving it. We should do more of it. What do you think?
+
+**Jerod Santo:** Yeah, I think so.
+
+**Adam Stacoviak:** We should do more of it.
+
+**Jerod Santo:** I didn't want to be the first one to say AI...
+
+**Adam Stacoviak:** I was going there.
+
+**Jerod Santo:** Okay, go ahead.
+
+**Adam Stacoviak:** Well, with you...
+
+**Jerod Santo:** Okay.
+
+**Adam Stacoviak:** Let me pull my notes, actually... I guess, first, let's start with this - obviously, AI is the latest breakthrough for everybody...
+
+**Jerod Santo:** He said it.
+
+**Adam Stacoviak:** ...and it's almost a mandate for every organization to somehow figure out "How do we leverage it? How do we enable it?" And some of it has been generative, some has been simple summarization... What has been your position with being in charge of product to think about this new tech that's burgeoning and infiltrating and welcoming? How are you thinking about it for PagerDuty?
+
+**Nora Jones:** I mean, I'm thinking about it in a lot of different ways right now. I think every single company in the world right now is trying to figure out how to use AI, how to develop policies around it... And for B2B SaaS companies, they're trying to figure out how to monetize it.
+
+**Jerod Santo:** Yeah.
+
+**Nora Jones:** And everyone's in the early days of it, but also everyone's kind of going through that, because that's how the market is reacting. There's expectations around that. I think especially within software reliability, my take is on focusing on when the agents fail, because that will be the case. And PagerDuty has been the one that you call in those situations... And so where my head's at right now is how do we keep being that for folks, even in this new age? And how do we help them with the parts of PagerDuty that are tedious, or annoying? For example, we're developing a shift agent that takes care of your schedules for you, because I know that is the largest complaint...
+
+**Adam Stacoviak:** Beautiful. I love that. Shift agent.
+
+**Nora Jones:** Yeah, shift agent. So it's things like that. I'm not a huge fan of AI that I feel like takes away the necessary human tasks, but it's like the things that feel tedious or boring. And I don't remember where this quote came from, but it was like "I don't want AI to make my art for me. I want AI to do my laundry for me, so that I can make more art." That's kind of how I think about it as well. But I use it every day, multiple times a day, and it's been huge.
+
+**Break**: \[23:03\]
+
+**Adam Stacoviak:** One thing you said recently, and this is on X... You said "The complexity of novel incidents", which I thought was pretty cool, "will only increase." Can you describe that?
+
+**Nora Jones:** Yeah. I mean, we are going to have incidents that we have no idea how to predict right now, and they are going to involve things that no one internally wrote. And so we also have to figure out how to debug them, and I think that is going to be challenging. I think the level of incidents we have are going to reach this rate where they're very challenging for a little bit, we're going to understand them a little bit more, and it's going to normalize, but I don't think we've hit that point yet. So I think we're all kind of like bracing ourselves for it.
+
+I even saw... Someone was using -- I think it was Repl.it. And Repl.it deleted the entire codebase during an incident. It was like a new -- and the agent's apologizing, like "Oh yeah, sorry, I really shouldn't have done that." \[laughs\]
+
+**Adam Stacoviak:** That's bad.
+
+**Jerod Santo:** That's normally when you would fire somebody.
+
+**Adam Stacoviak:** That's straight up out of Silicon Valley, okay? Season six, that happened. The TV show.
+
+**Nora Jones:** Yeah.
+
+**Jerod Santo:** So when we were in Seattle -- when was that, May?
+
+**Adam Stacoviak:** Yeah, May.
+
+**Jerod Santo:** For Microsoft Build. There was a -- I don't know if it was a demo, or a video... There were some claims that I didn't substantiate one way or the other, that they had agents on standby to resolve your incidents for you. Or at least to do the first level of triage on an incident, so that that pager goes off maybe later, maybe not at all. And I thought "That seems really awesome." Like, I would not -- it would be nice if nobody was on pager duty. Like, that's a world that we all would appreciate. It's not a reality yet, but there's starting to be people saying that that kind of stuff is going on. From your purview inside PagerDuty, is that going on? Are there agents doing that kind of stuff?
+
+**Nora Jones:** Yeah. And I will caveat this, and just to put it simply, I think there's a few different types of incidents. There's the well understood, which are probably not even incidents anymore... And I'm sure all of us have worked at organizations where someone is urgently doing or fixing something that they've probably fixed before, and maybe there could be more automation around it. And so that's where I think AI could really help, is like those interrupts that -- and we've all seen it, we've all worked with someone where you're like "Why are you still spending your time doing that? There's better things for us to be doing right now." And so I think that's where AI can really help.
+
+And then I think there's partially understood incidents, where it's like this kind of feels familiar, but not totally. I think AI can maybe assist there, but it shouldn't take the wheel. And then with the new and novel ones, like I was talking about, that needs to be pretty human-driven. And so I think it's just AI will really, really help with the kind of tedious, repetitive incidents and automation around them. That I think would save us all a lot of time and allow us to focus on the things that really need our focus.
+
+**Adam Stacoviak:** Are you at a level where you're still involved with the incidents? How much do you talk to the user, I suppose is the question? And two, can you give some details on the specifics of the novel incidents? Like, a terrible one, or just like when you're like "Wow, that would have never been in my imagination before."
+
+**Nora Jones:** \[29:41\] Yeah, so I'm not in incidents anymore, but I'm talking to our customers and our users every single day, and it's still one of my favorite topics. I definitely do disaster tourism in our incident channel, just to see and understand... But I know better than to try to like hop in, you know...
+
+But I think a lot of the new and novel is going to be around leveraging agents, leveraging generative AI in parts of production, in areas that we don't understand. I think that's really where a lot of it's going to come up. I'm very curious about the company that had their codebase deleted... Like I would love to read that incident review when it comes out.
+
+**Adam Stacoviak:** "Undelete that, AI. Undelete it now."
+
+**Jerod Santo:** Yes. "Undo, undo, undo." Are there any -- this could be historical as well, not necessarily new and novel, but any fun incidences that are no longer under NDA? You don't have to name any names or anything like that, but certainly there are probably situations that have become legend amongst your teams over time, maybe internally, or maybe just like anonymously, externally, crazy stuff that's happened. Can you think of anything? It doesn't have to be like the craziest ever, but...
+
+**Nora Jones:** Like recent?
+
+**Jerod Santo:** No, just whenever.
+
+**Nora Jones:** I mean, yeah, I have my favorite incidents...
+
+**Jerod Santo:** Okay.
+
+**Nora Jones:** I always joke that when I'm old, I'm just going to be reciting incident numbers... Because 712 still sticks in my head, and 659. I remember those incidents, because I basically slept in the office that night, trying to resolve it. A lot of the incidents were trying to figure out in a very simple way how a piece of technology was supposed to work. And a lot of the times we had put the technology through the paces and almost leveraged it to do the things that it wasn't meant to do. Consul, for example, Kafka... Those have been in some pretty hairy incidents in my experience. And a lot of the time it's because we implemented them without having deep subject matter expertise in them. And then we also saw the power and the flexibility of some of it, and used it for things that it wasn't necessarily meant to be used for. And then a lot of the folks that designed those things don't work at the company anymore. So then you're trying to figure out live what the intention was of this decision, and what the expectation was, and what's actually happening... And those can get pretty complex.
+
+**Adam Stacoviak:** Can you take us tactically into -- I know you spent you said the first year kind of integrating... When did you begin this kind of newer role? Was it a year after you integrated and you started forming a team? Take us tactically into forming that team, strategy, rejiggering PagerDuty product, pricing... How did you rethink a lot of what was going on there?
+
+**Nora Jones:** Yeah, so when I joined PagerDuty, we had like two main product orgs. We had incident management, and then we had another product org that was doing automation, and AI, and it was like kind of focusing on other use cases... So like CS users, or FinOps, you know? So we obviously went into the incident management organization... There was no team around post incident reviews. And they did have an integration team that was in charge of a bunch of different integrations... So like ServiceNow, Datadog... That team was also in charge of the Slack integration, but it had been very underinvested in. I obviously had the team I came in with, and then I also had the chat experience team is what we called it, which was our Slack and Teams integrations.
+
+\[34:06\] The very first task we had to do was -- we were obviously selling Jelli standalone prior to the acquisition... The very first task we had to do was to get it into one of the existing PagerDuty plans, which was the enterprise plan. And working at a startup, I was using products that get things into skews really quickly, and I didn't realize the complexities of it in a large organization. When that was deemed my first task, I'm like "Cool, that'll take like 20 minutes." That did not take 20 minutes.
+
+**Jerod Santo:** Sure.
+
+**Nora Jones:** So I learned a lot about that process... And then it was also taking the existing PagerDuty Slack app and bringing all the Jeli functionality into that Slack app. So that took about a year to really fully integrate, and it's live, so you should check it out... And then, midway through the year, that was when I kind of had that realization that our pricing wasn't totally matching our value. So I just kept sharing it in meetings over and over again until I think I annoyed enough people that they were like "Okay, you can go do it." And then I ended up in the head of pricing role after that.
+
+**Adam Stacoviak:** You have a new job now.
+
+**Nora Jones:** I have a new job now. Yeah, exactly. So after launching it in all three product lines, we went through a reorg internally, and aligned our organization a bit more to the customer value, which was around six months ago... So my new pillar of the org is product strategy, growth and pricing. So I have pricing analysts under me, I have engineers working on growth, onboarding... I have folks working on insights and analytics, so really driving what are patterns we're seeing, and how do we feed that back in so that we can help you with the well understood... And then - yeah, pricing analysts and product intelligence folks as well. So they are looking at how users are using the product, which areas they're getting the most value out of, to really understand where we should keep investing in.
+
+**Adam Stacoviak:** Growth at Jeli, OG Jeli was probably different than growth at PagerDuty.
+
+**Nora Jones:** Yeah. We didn't have growth...
+
+**Adam Stacoviak:** Can you go into some of the details of how it's different, how it's changed? Is it scary? Is it just way bigger? Do you have more money to spend? Bigger distribution because you have much more customers? What's it like?
+
+**Nora Jones:** I mean, yeah, we have way more customers. At Jeli I had personal phone numbers of all of our customers. That's not able to be the case at PagerDuty. I think customer relationships are so important, but obviously, I can't have relationships with all of our customers now. So I also have to like use more discernment, and I try to form relationships based on where I'm seeing usage. For example, we have an incident roles function, or capability in the product... And we obfuscate a lot of data, but I could see people were creating a manual scribe role. And we actually have automation that, that does that for you, so that no one's like typing the entire Zoom call, which is... That's not a great use of time. Like, that would actually be a good use of AI. So I'll reach out in those situations, and... That's sort of really how I think about it.
+
+I'm thinking about it from the customer relationship perspective, and I'm also thinking about it in terms of like an entire incident lifecycle, from like detecting, to triaging, to mobilizing a team, to resolving, to learning. If I see customers that are only here, I try to go talk to them. Because when you're only in the detect and like the noise phase without having standard processes all the way over here, you're going to end up in this hamster wheel... And you could actually drive a lot more improvement and save a lot more time.
+
+\[38:15\] So that's really how I think about it, is just in terms of customer value, which I think is where my background really helps. Like, I really love being in this role, because I've been a user. So it's nice for that to come full circle.
+
+**Adam Stacoviak:** How much code do you write?
+
+**Nora Jones:** Zero. I mean, just in my free time, but not at work. Yeah.
+
+**Adam Stacoviak:** Is it because AI is writing it for you?
+
+**Jerod Santo:** I was gonna say, how much code do I write? Way less than I wrote a couple months ago... If you set AI aside and imagine Open AI hadn't done what they did with ChatGPT, November of '22, and everybody freaked out and it changed the course of the industry... What would the product roadmap look like? You're looking at product at PagerDuty, right? Like, what -- your initiatives certainly changed. I'm not sure how multi-threaded you guys are in terms of things you can work on at the same time. But everybody has to have an AI strategy, and that's because everybody has to have an AI strategy.
+
+**Nora Jones:** Yeah.
+
+**Jerod Santo:** So that drives product. We were actually joking about it last night at the meetup... It's kind of like when one cool kid in middle school gets Air Jordans, whether they're ugly or not, it's like, everybody has to. What's your Air Jordan strategy? How are you going to get the Air Jordans? And it's like "I've got to figure something out. I don't know. My friends are going to ask." And that's kind of AI right now, like where it's like "What's your strategy?" You have to have one. So that's driving a lot of the product roadmaps, right?
+
+**Nora Jones:** Totally.
+
+**Jerod Santo:** If that wasn't something you had to think about, you would have more freedom to build kind of what you would like to, or... I mean, maybe you haven't thought about that because you're too busy thinking about what you have to think about, but in a world where you didn't have that necessity, what would it look like?
+
+**Nora Jones:** I mean, I think it has driven some of our product roadmap, but it's also made us faster in a lot of ways.
+
+**Jerod Santo:** Sure.
+
+**Nora Jones:** I don't spend a lot of time coding in my organization because that's not my best value to the customers, or to the company, or to the folks I'm managing. But what I have found AI useful and is not just like driving the product roadmap it's like actually aligning folks. If I have an idea or if I've been seeing a lot of customer data, I can mock up something really quickly, and have a conversation with our engineers, our designers, and our executives way more quickly than I could have otherwise. We don't have to spend a lot of time in a lot of design cycles getting that initial wireframe; we can just start to visualize. And so I think it's been really actually helpful with remote work from that perspective, especially when you can't get in a whiteboard with someone... So I do think it's, it's made us faster. I think that we would be slower --
+
+**Jerod Santo:** Sure.
+
+**Nora Jones:** I would certainly be slower without it. I think in terms of --
+
+**Jerod Santo:** Just like Air Jordans. I mean, you run faster with the nice Jordans.
+
+**Nora Jones:** \[laughs\] What's that?
+
+**Jerod Santo:** I was making a joke. It was bad... Keep going.
+
+**Adam Stacoviak:** You run faster if you have Jordans.
+
+**Nora Jones:** Yes. Yeah, exactly.
+
+**Jerod Santo:** Sorry...
+
+**Nora Jones:** But yeah, in terms of the product roadmap, I don't think -- there's never going to be a world where we have zero incidents. So I think those are some of the wrong problems to go after with AI, is like trying to get to like zero incidents. It's more to handle the more well understood stuff, so that engineers can focus their time on things that only they can do, rather than things that agents or AI can help with. I think the whole industry is going to get it wrong in some cases. We are very new at this, and we're all kind of like stumbling toddlers with AI right now. And so we're going to mess up, and I think we just have to reflect on "Here's a good use for it, and here's a not so good use for it", rather than trying to use it for everything. And so that's my biggest worry, is that people are going to get overly dependent on it for things that they should not get overly dependent on it for.
+
+**Adam Stacoviak:** \[42:31\] I agree with that.
+
+**Jerod Santo:** Yeah. And earlier when you were talking about novel incidents, or like the really tough ones where you need that one subject matter expert to come in and connect the dots that nobody else could connect, right? Do you -- I don't want to use the word "fear" necessarily... Are you concerned that increased adoption of AI tooling reduces gradually subject matter expertise?
+
+**Nora Jones:** I'm definitely concerned about that. I think there's ways that it increases subject matter expertise. I started Jeli before all of this was really happening, and my honest vision for Jeli was there is so much data and how you communicate that you can learn what Adam's expertise is, and what Jerod's expertise is just by how they're talking. Like "Oh, we noticed we pull Adam in for every issue we have with this particular codebase." And so we start making this map of Adam's expertise, and we can start to download how Adam knows what he knows, and how he does what he does...
+
+**Adam Stacoviak:** Or doesn't.
+
+**Nora Jones:** Or doesn't. And making that searchable. So I always wanted to put AI in Jeli, but then everything changed. So I actually think it could really enhance subject matter expertise, because it has a look into your whole organization... I think the challenge is with hallucination with that. Like, is it going to make up stuff that it's not actually seen? And I also think that humans could stop trusting their own judgment sometimes. I'll always test myself sometimes. Like, if I feel like I'm personally using AI too much, I will ask it a question that I know I have a lot of expertise in, just to see how it does, you know? And I feel like that helps me come back to Earth sometimes, just to make sure I'm staying grounded. But I think we all have that personal responsibility, but it's also -- it's going to change the way we talk to each other. I think a lot about the language aspects of it, and I think about like kids that are probably learning with that now... How are they going to talk to each other in the workspace, in a certain amount of time? Anyway, those are things that I...
+
+**Jerod Santo:** Yeah. To that point, there was a recent headline about changes in language, and how ChatGPT in particular is already affecting the way that we talk. For instance, the word "delve" is like skyrocketing in use, because it routinely says delve. Like "Let me delve into that for you." And now more people are saying delve. This happens all the time in humanity. You hang out with people and you start to talk like those people, and we're just hanging out with ChatGPT.
+
+**Nora Jones:** Right.
+
+**Jerod Santo:** And so we're going to start talking... "It's not just X, it's Y." It always says that, too.
+
+**Nora Jones:** Yeah, exactly. And you know, I've always loved using em dashes, but I also find myself not using em dashes...
+
+**Adam Stacoviak:** No more em dashes...
+
+**Nora Jones:** ...and I'm so upset about it, because I've been using it for years...
+
+**Adam Stacoviak:** Me too!
+
+**Nora Jones:** ...and now I'm rethinking, "Are people going to think I wrote this with AI because I'm using an em dash?"
+
+**Jerod Santo:** Right? So you purposely remove them, because you don't want to look like this is AI-generated.
+
+**Adam Stacoviak:** Same.
+
+**Nora Jones:** Yeah.
+
+**Adam Stacoviak:** \[46:02\] Can we go into -- I'm not sure how this manifests for you, but as a leader, it seems like maybe you all have some tooling that lets you glimpse into someone's capabilities. I imagine that's a great tool as a leader, because you want to spend one-on-ones with folks... Some of the team you know really well, because you brought them with you from Jeli, but some you're bringing on... And as a leader, you want to know what their strengths and weaknesses are. Do you have that kind of insight? Is that what you're talking about, where you can sort of like probe into your team to say, "Okay, Jerod's really good at this, and Adam's really good at that", and you have the AI kind of like learn it for you, and you can sort of prompt whatever it might be to learn more about your team.
+
+**Nora Jones:** I haven't been using it for that perspective, but I do -- you know, as I build out my org, and I'm interviewing folks right now, and I've interviewed folks that are really great, and clearly experts in a certain position, but I'm also thinking about the expertise of the current team I have... And if there's these four different areas I need expertise in with pricing, for example, if I already have someone in that bucket and I'm interviewing someone that also has the same expertise, that's not going to maybe be valuable for me to hire that person, even though they're great. And so I do think that that could be helpful in terms of just like creating teams that work really well together, and in looking at where someone's strength is.
+
+I have someone on the pricing team that loves to go into the data, data expert, but is not as proactive with talking to sales, and talking to customers... And so as I was looking for my next person, I was like "I really need someone that is doing that in the organization. I don't necessarily need them to be a deep expert in the data, because I already have that. But I know that just because I interact with my team a lot. I think if you are running a large org, that might be helpful. I also -- I use Granola, which is really, really great. I don't know if either of you use it...
+
+**Adam Stacoviak:** What is it called?
+
+**Nora Jones:** Granola.
+
+**Jerod Santo:** Never heard of it.
+
+**Nora Jones:** It's a note taker during meetings... And I have found that so fantastic, especially for one-on-ones. It just helps me remember them, it helps me actually be present, so that I'm not having to take notes during it... It prompts me of things to follow up on... And so I feel like it's actually made me a better manager, because I'm able to be more present, and recall more.
+
+**Adam Stacoviak:** How is -- I hate to keep going to AI, man... But how is -- I was just saying I hate to keep talking about AI, kind of, but I'm curious how AI is helping you not hire, or let go of people. Can you be candid about how it's shrinking or growing your teams?
+
+**Nora Jones:** Yeah, and I know I probably sound like a broken record, but I really think it takes over the tasks that people were probably doing manually, that they didn't need to be doing. So I think it increases roles, in that it is going to completely change certain jobs where someone's entire job was manual data entry.
+
+I spend a lot of time in like Tableau, and Periscope, and unfortunately Excel now... And when I get stuck, or break something, I actually use AI to help me fix my query. And that's been really nice for me, in just terms of speed.
+
+**Break**: \[49:49\]
+
+**Jerod Santo:** What are you excited about, or what are you interested in, whether industry or elsewise?
+
+**Nora Jones:** What am I excited about...?
+
+**Adam Stacoviak:** Well, she's golfing now... She's a golfer...
+
+**Nora Jones:** Yeah, I've golfed twice, so...
+
+**Adam Stacoviak:** You're a golfer.
+
+**Nora Jones:** I think I'm a golfer now.
+
+**Jerod Santo:** Yeah, for sure. No imposters.
+
+**Nora Jones:** And you know, I was talking to Adam earlier. He was like "I mean, I just hit the ball. I don't really golf."
+
+**Jerod Santo:** You said that?
+
+**Adam Stacoviak:** I did...
+
+**Nora Jones:** And I really resonated with that. I can hit the ball, and it feels good... But yeah, what am I interested in? I'm really interested in how linguistics are changing with AI, and also just more remote-first cultures. We really lean on text communication now more than ever, and I think things can easily be misinterpreted over text, which actually can cause a lot of --
+
+**Adam Stacoviak:** Sure. That's a good response for people. In text, they would say like "Sure."
+
+**Nora Jones:** Yeah. Or "K".
+
+**Adam Stacoviak:** And I'm like "What does he mean by sure?"
+
+**Jerod Santo:** I say that all the time. It means yes.
+
+**Nora Jones:** When I joined PagerDuty, everyone reacted with the ACK emoji, which is a very engineering thing to do, but... You know, I had people that weren't in engineering join PagerDuty, and they thought people were mad at them, because they're reacting with the word ACK in Slack. And I'm like "Oh no, they're not mad at you. They're just giving you a thumbs up."
+
+**Jerod Santo:** Do they know there's a thumbs up emoji they could use?
+
+**Nora Jones:** What's that?
+
+**Jerod Santo:** I said there is a thumbs up emoji they could use.
+
+**Nora Jones:** I know, exactly.
+
+**Jerod Santo:** You're using the at?
+
+**Nora Jones:** ACK, A-C-K.
+
+**Jerod Santo:** Oh, ACK.
+
+**Nora Jones:** Which I think the history of that usage at PagerDuty is because you're acknowledging an alert. And so people use it internally more than I've ever seen at any other organization to just be like "Okay, cool." But they're not upset about anything. Anyway, I'm thinking a lot about how these nuances -- we all have our own perspective of how to communicate and organize, and it really impacts the organization, it really impacts how products get shipped and developed, it impacts how incidents get resolved... And so I've been looking more into that lately. I've been following this guy on Substack, Adam Aleksic... He's a linguistics graduate from Harvard, and he's been really just analyzing how the world is changing its communication, but I'm really interested in how that impacts organizations, and how it impacts products.
+
+**Adam Stacoviak:** We are live before a studio audience, they are hearing this... But can you share what's on the horizon? What's something new for PagerDuty that you're excited about, that you can maybe tease a little bit, or just spill the beans on?
+
+**Nora Jones:** Yeah, I mean, we are completely revamping our... I know this probably doesn't sound as exciting for folks, but it's very exciting to me. We're completely revamping our pricing and packaging, so that you're not going to have to worry about seats as much as you do. It's a big transformation, and it's really focused on value that we're providing. That's the little teaser.
+
+And then other things I'm really excited about are seeing how Jeli evolves within PagerDuty. There are a lot of insights that we have in Jeli, and combined with the insights of PagerDuty I think can really drive change to the beginning of that incident cycle. And so I'm pretty stoked about that, and how that can help onboard new engineers, and how it can help ramp other folks up, like maybe ramp up a junior SRE because they can see how someone solved the problem. I'm excited about how it might level people up, and create this new culture of engineering. Yeah.
+
+**Jerod Santo:** \[56:23\] Well, Nora, you are smart, kind, and fun to talk to, so we really appreciate you joining us... This has been really awesome. Thank you. Let's hear it for Nora Jones.
+
+**Adam Stacoviak:** Nora Jones.
+
+**Jerod Santo:** Wanna do a little Q&A?
+
+**Nora Jones:** Sure.
+
+**Jerod Santo:** We do have nine minutes, as I've been told, and he's correct... So -- we also have a microphone right here... And we have a little treat. Keychains. So the stickers are out there, free to have, but these keychains, these cost extra. To get a keychain, you have to come to this microphone and ask a thoughtful question for Nora. You can judge Nora, whether it was a good question or not, after you answer it.
+
+**Nora Jones:** Keychain-worthy or not.
+
+**Jerod Santo:** So if anybody has a question, please come on up for a keychain. Please say your name, and what kind of --
+
+**Nabil:** I'm Nabil. So I kind of have a two-part question, and it's about the acquisition, basically. So it sounds like a great match and everything, and so I imagine most people thought it made sense and everything, but I was just curious, from Jeli's side, before the acquisition actually happened, were there people that were against it, and maybe what were their rationale for that?
+
+**Nora Jones:** Were there people that were what?
+
+**Nabil:** Were there people from Jeli that didn't want the acquisition to happen, that were against it? And then maybe after the acquisition, if you're able to share that, was there a percentage that just kind of took their equity and left soon after, or whatever? I'm just curious.
+
+**Nora Jones:** I mean, I'll answer that as best as I can, but I think all acquisitions have challenges that come up. I think overall it was a very positive experience. I actually was -- I love reading research, and I was actually reading a bunch of research on what makes a successful acquisition, and a lot of the time it's if the founder stays. And so I also was trying to really focus on being happy myself, because I knew I was modeling that to my team, and figuring out those ways...
+
+There were definitely people at the organization that hadn't worked at a large company before, and it just wasn't a fit for them, and so they moved on, and that's totally okay. I think that makes sense in those situations. And then I think there might have been some people that were initially like "Oh..." There's always -- when you're a 20-person company, it's a really cool energy that you can't replicate in a 1,200-person company. So I think there was grief involved with it from everyone's perspective, including my own. I've spent a lot of time grieving not being a founder anymore, and it not being my baby...
+
+I think the hardest thing is when I see a problem, and I'm like "That is not my responsibility anymore. I'm just going to say my words, and then move on." But I think there were also some people that have been surprised by how much they like it, that I think will probably stay at the organization for a long time.
+
+**Adam Stacoviak:** Thanks, Nabil. Anything else?
+
+**Nabil:** No, that's it.
+
+**Adam Stacoviak:** Thank you, sir.
+
+**Nora Jones:** That was a good question.
+
+**Nora Jones:** Good catch. Oh. Too soon. I thought you had it.
+
+**Adam Stacoviak:** James.
+
+**James:** \[59:47\] Hi, folks. I'm James. One of the things that we know from other industries - avionics, nuclear, the list goes on - is after each incident, there's an incentive to make the system more automated, more safe... And that reduces the average number of incidents, but it increases the worst incidents. How does PagerDuty think about proactively helping organizations not just react to incidents, not just do their own thinking about incidents, but instill knowledge about how to design systems, how to design their organizations to prevent the catastrophic ones?
+
+**Nora Jones:** Yeah, it's a good question. I don't think that the right approach after a gnarly incident is to set up automation. I think the right approach organizationally after a gnarly incident is to talk about it, a lot, and try to figure out how it happened. Not why it happened, but how. And have these productive conversations where we're all sharing our piece of how we thought things were supposed to happen, so that we can learn from each other, and so we can learn about how this actually unfolded. I think until you have that understanding, automation is going to cause issues, to your point. I've been a big believer in that for a while.
+
+When we were at Jeli, myself and Laura McGuire and \[unintelligible 01:01:20.14\] we wrote what we called the "How We" guide, which stood for the "How We Got Here Incident Analysis Guide, which is free online, but it talks about how to extract the most learnings after incidents, so that you're better prepared in the future. And I didn't mention this aspect of right when I came in, but one of the most important things to me when I came in was that we were using Jeli internally, and also practicing what we were preaching. So we use Jeli for every single incident now, and we are looking at how we communicate during those incidents, which is driving what we do or do not change.
+
+I think with some incidents you might not need an action item at the end of it. It doesn't mean that you didn't learn from it, but the action that you take might be understanding your system more. So I think the best thing that we can do as PagerDuty is model that for folks, so that we can help them do it themselves.
+
+**Adam Stacoviak:** Thanks, James.
+
+**James:** I'm really here for the keychain, because I have a toddler. \[laughter\]
+
+**Adam Stacoviak:** Thanks, James.
+
+**Jerod Santo:** There you go, James.
+
+**Kendall:** Hi, Nora. I'm Kendall.
+
+**Nora Jones:** Hi.
+
+**Kendall:** How long until you're twitchy and go do another thing?
+
+**Nora Jones:** Say that one more time.
+
+**Kendall:** How long until you get twitchy and go do another thing?
+
+**Nora Jones:** And go do another company?
+
+**Kendall:** Yeah.
+
+**Adam Stacoviak:** Are you itching to leave and do something else? When will you get the itch?
+
+**Kendall:** I mean, there's probably golden handcuffs, and there's probably an appropriate answer, but I'm curious what you're actually thinking.
+
+**Nora Jones:** I am having fun right now. I'm doing things that are challenging to me, and I had never thought I would be doing pricing... But it is activating both parts of my brain that I really love to activate, which is psychology and math. And it's very much an art and not a science, which is what I really -- that's where I really thrive and gravitate towards. So I think as long as I'm learning and feeling like I'm adding value, and feeling like I can make a difference, I don't think I will get an itch. But I'm also playing with different technologies that I can't necessarily always play with at work, and I... You know, I haven't found something that really captures my interest enough yet. But yeah, I'm having a lot of fun, so...
+
+**Kendall:** Thanks.
+
+**Adam Stacoviak:** Thanks, Kendall. We've got three more.
+
+**Jesse:** \[01:03:59.00\] Hi, I'm Jesse. You said your first customer at Jeli was an enterprise customer, and that's kind of rare. How did you get that? Was that a previous relationship, or just \[unintelligible 01:04:06.22\]
+
+**Nora Jones:** Yeah, good question. So like I mentioned, I have been an engineer most of my career, and it was within the reliability community. So most of my network was in reliability. And so I was really building out in the open with people I knew. And I also previously worked for Netflix, I was working for a company that got acquired by Walmart, and I was working for Slack... And so I really only knew enterprises. Looking back, I thought it was kind of an accident, but it was like "Oh, I was designing for something I needed in my organizations." And so I think that's just how it happened. And as we were building out in the open, we were looking for design partners, and we had slots for five. And I tried to get five very different types of customers, because I wanted to figure out who was finding the most value out of it... We did one with a startup, and it went horribly, actually. And then it went really well with this enterprise. And I remember talking to our investors about it, because if you -- if you have an investor and you're telling them that the first customer you're signing as an enterprise customer, they're gonna try to talk you out of it, because they don't think you're ready for it yet. And it was way more organic for me to build for enterprise, rather than to build for startups.
+
+**Jesse:** Yeah, that's awesome. I think there's something to \[unintelligible 01:05:41.29\]
+
+**Adam Stacoviak:** Thank you very much for that question.
+
+**Nora Jones:** Yeah, thank you.
+
+**Adam Stacoviak:** Yes. Matt...
+
+**Matt:** You've talked passionately about a lot of the things for post-incidents, and my expectation is that there's a -- knowledge transfer and communication is hard... And I'm curious what your thoughts are on where PagerDuty and the aspects that are implemented there integrate with things outside of PagerDuty to help facilitate better knowledge transfer and communication across different divisions within an organization? Because not everybody implements all the things that you expect, and so I would like to know what your expectations might be for how an organization might better leverage the product to make that vision a lot more usable.
+
+**Nora Jones:** So if I'm understanding your question correctly, how can customers that are using PagerDuty best set it up so that they're creating a learning environment?
+
+**Matt:** Yeah. Because these things are hard... And how do you help leverage making that easier?
+
+**Nora Jones:** Yeah. So one of the things I really appreciate about PagerDuty is how they've developed their integration system. We don't expect customers to use us for everything, but you do get the most value out of it if you're connecting it to all the other tools that you're using. There's over 700 integrations, and you can also build your own... And I think that is the best way to get the value out of it, because then we can understand all the signals that you might be seeing, and really help you prioritize if something is a really urgent signal that deserves human attention, versus not.
+
+I think that's where I see a setup go awry, is when folks don't spend the time to set it up that way, and then it just becomes more noisy than it needs to be. But if you are really taking a look at your overall company, and being like "We use this product here, we use this product here", and really connecting it to PagerDuty, it's going to be a lot more valuable of a system that is not just learning from all that, but it's also surfacing a lot of that to you as well.
+
+**Matt:** \[01:08:19.16\] And do you see that as like a good way to interact or get that information to the parts outside of just the engineering, and expose that understanding?
+
+**Nora Jones:** Yeah, so...
+
+**Adam Stacoviak:** Can you restate that question, please?
+
+**Nora Jones:** Yeah, so the question was "Do you see that just for engineering, or are you exposing it to everybody?" So that is something that was always on my mind with Jeli, is just like not keeping these learnings in a vacuum with just engineering. And it is challenging. It is not an issue a tool can totally solve. It has to be a cultural change. It has to be an engineer going and talking to the person that launched a campaign, and learning more about their world. And I think we really try to facilitate that. We show you all the people that were involved in an incident, even people that were just lurking... Which can be really interesting sometimes. Like "Hey, why did legal hop on to lurk?" Go DM that legal person and ask "Hey, I'm just reevaluating this incident. I'd love to know what your interest was in it."
+
+And I think what we really try to provide with the platform is a way to make that less awkward... Because it's like you're just looking at the data, and the data points say that these different folks joined. So it's worthwhile to talk to them to get this holistic view of what really happened. And so I think that's how we're really facilitating it.
+
+**Adam Stacoviak:** That's cool. Thank you, Matt. Next?
+
+**Jim:** Hi, my name is Jim. I wanted to ask you - you've thought a lot about reliability... So when you were early in your career, did you study and reference? What's your velvet underground of reliability?
+
+**Nora Jones:** I really started -- I actually started my career in hardware. I studied computer engineering and I went directly to home security in a smart home company. It's terrifying having that as your first job, as like a reliability person, because if you mess up, someone's house might get robbed. So that was stressful.
+
+And then I saw this job opening at an e-commerce company for a software engineer. That wasn't technically my role. I was doing less web development and doing more hardware... But I looked at some of the traits that were needed. It was like a developer productivity and reliability role, and I was like "I do a bunch of this now", and so I just applied. And I joined the org, and they were having incidents multiple times a day, every day. The company was growing so quickly that there wasn't even a place to do our incident reviews, so we used to do them in the hallway every morning... And it was like the hallway next to the bathroom, so people were going like in and out of it... It was bananas.
+
+And I was really trying to get creative. Like, my boss at that point was like "Do whatever you want. I don't care. We can just try things." And I found what Netflix was doing, which was called chaos engineering. And I started reading a lot about it, and I was like "Wow, this is like really, really interesting", because there is a big organizational psychology to reliability, too. It's not something that can just be fixed by fixing code.
+
+And so I started experimenting with it, and I was really influenced by Netflix... I started -- actually, I stumbled into the field guide for human error, which is in the aviation space... And so all of those things I think were really fundamental to my passion now, and what I ended up following to learn more about it. And someone brought up other industries earlier... I also read post-incident reports from other industries as well, because I find it very fascinating... And it is relevant a lot of the time, too.
+
+**Jim:** Thank you.
+
+**Adam Stacoviak:** Is your book still for sale?
+
+**Nora Jones:** Is my what?
+
+**Adam Stacoviak:** Your book.
+
+**Nora Jones:** Yes, it is. Yeah.
+
+**Adam Stacoviak:** Go buy the book.
+
+**Jerod Santo:** There you go.
+
+**Adam Stacoviak:** We'll link to it in the show notes.
+
+**Jerod Santo:** Let's hear it one more time, Nora Jones.
+
+**Adam Stacoviak:** Nora Jones, y'all.

Leading leaders who lead engineers (remastered) (Interview)_transcript.txt

@@ -0,0 +1,621 @@
+**Adam Stacoviak:** Lara Hogan - hey, nice to have you here on the Changelog, finally. I'm a big fan of yours. I've been reading and paying attention to you for so long, and I think it was back when I logged Laura Hogan on mentorship and sponsorship. This is the first time I've heard the idea of sponsorship and really dug deep. We logged that through our news feed, and then I think I subscribed to your newsletter immediately after that. It was super-cool.
+
+**Lara Hogan:** Oh, that's so nice. That's so nice, thank you. I'm so excited to be here.
+
+**Adam Stacoviak:** Yeah. And you wrote that post back in 2017. I didn't find it until 2019, so - sorry about that.
+
+**Lara Hogan:** It's evergreen, you know? It's the kind of thing that I could talk about all day long, every single day, for the rest of my life. It's the kind of thing where it seeps into everything that we do, this idea that by default we all tend to give advice to each other, but actually that's not the most powerful tool in our toolbox. And there's all of these other tools, for example sponsorship, that we can use every day.
+
+So yeah, I find myself talking about it... My poor partner is sick of it at this point, like "Alright, I get it, Lara..." \[laughs\]
+
+**Jerod Santo:** So you've teed it up... What's the difference then? How do you break them out?
+
+**Lara Hogan:** Yeah, totally. So if we step into a room with someone and we're so excited to support them, we really wanna see this person grow and learn, we see so much potential... Instead of defaulting to giving them advice to sharing our perspective, sharing what we've seen work and not work, instead it's really important to consider what are the opportunities that we can basically put our name online for on behalf of this person. Where can we throw their name into the ring? Where can we give them the opportunity to do visible, valuable work that's connected to business goals?
+
+We often forget that we can do this, because we already have folks in mind, or we have folks with more experience, or we have folks that are friends that we're already thinking about whenever a new opportunity comes up... But when it comes to helping people grow, sponsorship is actually the one -- you know, studies have shown it's correlated to career growth, not mentorship. Mentorship is just like "Here, let me show you all the things that I know to help you avoid pitfalls in the future." That's not growth. That's getting unblocked, or maybe that's avoiding major issues, but that's not growth.
+
+So in a lot of my work I'm trying to help folks, especially engineers, get out of the habit of believing that the best thing that we can do, the value that we can provide is sharing our knowledge, and instead think about these opportunities and where can we sponsor people through these opportunities.
+
+**Adam Stacoviak:** You said it was correlated to what? When you said that it was correlated to something, I didn't hear what you said there. What was it correlated to do
+
+**Lara Hogan:** I don't know, I forgot already...
+
+**Jerod Santo:** Career growth, I think you said?
+
+**Lara Hogan:** Oh, yeah, career growth. Thank you.
+
+**Adam Stacoviak:** Gotcha. I wasn't sure if it was like a research, or something like that. I wanna make sure I earmark that in my brain, because I'll forget if not.
+
+**Lara Hogan:** Tons... Yeah, there's a ton of research. I mean, sponsorship is one of those things that's been studied for decades, and all of the studies show that people get further in their careers, they have more opportunities to do even more visible, valuable work, they're more likely to get a stretch assignment, they're more likely to get raises, they're more likely to have other people higher up in the ladder know their name... You know, people \[unintelligible 00:07:05.10\]
+
+**Jerod Santo:** Yeah.
+
+**Lara Hogan:** There's so much data that says that that's where we should be investing our energy when it comes to supporting other people as they grow... And that's also what we should be asking for when we're thinking about our own growth... Like, who are the people that know our work and might be able to vouch for us for the big, juicy leadership opportunities that, you know -- like, I wanna start public speaking, or I wanna write a company blog post, or I wanna open source this project... Who are the people that can maybe vouch for me and help me put my name in the ring for those opportunities?
+
+**Adam Stacoviak:** Yeah. In some ways too it almost forces you to not be in your own way.
+
+**Lara Hogan:** Absolutely.
+
+**Adam Stacoviak:** Because so often do we just not nominate ourselves, because we're like "Oh, I'm sort of an impostor here, I'm not actually that good. I don't really do this well. Some people pay attention, but let's just face it, I'm not that good." And then somebody close to you nominates you or suggests you or sponsors you or advocates for you or refers you, however it is, and you almost feel like you have to do better because you love them, or you care for them, or you appreciate them, or whatever it is, their position in your life, and you're like "I have to live up to what they believe in for me", so it almost makes you do better even.
+
+**Lara Hogan:** \[08:13\] Absolutely. I have this sneaky trick question at the beginning of my workshop on mentoring, coaching and sponsoring... And my sneaky trick question is I have everybody who's in attendance share one thing that a manager has done for them that has skyrocketed their growth. And I don't tell anybody the difference between these three skills before we have this little intro... And everybody, nine times out of ten, they describe someone who's been a sponsor to them.
+
+And when I get to the sponsorship part of the workshop, I'm like "Guess what, everybody? Not one of you mentioned mentorship when you were describing something that a manager has done for you that has skyrocketed your growth." One person mentioned coaching, everybody else mentioned sponsorship.
+
+Jerod and Adam - think about a time when someone's done something for you that skyrocketed your growth. What was that?
+
+**Adam Stacoviak:** Well, I'm glad you asked, because I have that earmarked for something to mention... I've shared this before on this show at least once...
+
+**Jerod Santo:** Thank you for asking.
+
+**Lara Hogan:** \[laughs\]
+
+**Adam Stacoviak:** We have a show called Backstage, where Mireille Reece, Doctor of Clinical Psychology who I co-hosted Brain Science with - she asked me a lot of questions about my past. And then very specifically, back in the day when I was in the military, I was in what they call my MOS training. It's training for my job in the military, essentially. And I just hadn't really considered being a leader -- I mean, I had leadership qualities, I was a strong person in my friend groups... I wasn't like this shy person. But I just never really considered being the leader.
+
+So the drill sergeant says "Stacoviak! You're first squad leader now." Out of nowhere. Because I guess the other person was messing up. And I kind of correlate that to sponsorship. It's kind of by force, really, so I'm not sure... Maybe you can help me judge if that's truly sponsorship or not... But my drill sergeant believed in me enough, or just picked me by random number or whatever it might be, but he's like "You're now a leader." And I was like "Um... Okay..." What did I do? I didn't do anything special. And then from that day forth I just started to learn specific things to be a leader in the military... And all these things because that person believed in me, in many ways by force, I was suddenly a leader... You know what I mean?
+
+**Lara Hogan:** Yes...!
+
+**Adam Stacoviak:** So I had to live up to that. And that was many years ago, and I'm still -- I think since that moment I was like "Wow, I can be a leader."
+
+**Lara Hogan:** That's absolutely sponsorship. That is 100% the definition of sponsorship. And that's the kind of story that I hear all the time in these workshops from people. Like, "I don't know why this person believed in me. I don't know why this person had faith in me. I didn't think I could do this thing. But they did, and that skyrocketed my growth." 100% sponsorship.
+
+**Jerod Santo:** Have you been on the receiving end of that in your career?
+
+**Lara Hogan:** Me?
+
+**Jerod Santo:** Yeah.
+
+**Lara Hogan:** Oh, yeah. Oh, my goodness... It's funny, because I also participate in the interest too, so I was trying to pick a new example... And there's just so many. I think about the difference between a time when my manager went on parental leave for six months, and at the top of the doc for like "Who to contact in different situations while I'm out" he put "If you've got any questions, ask Lara." He was a VP, I was not. And I was not prepared for that was like -- he didn't ask me first... It was sponsorship. As you said, Adam, by force. Sponsorship by force. But again, it connected me to so many different people within the business. I didn't know what cap-ex or op-ex was. I didn't understand how head count worked, and it threw me in the deep end in a way that really, again, skyrocketed my growth.
+
+Or, I talk about the person who was my manager - actually at the same company - when I was on deck to get a promotion to director. I was leading a web performance team, I was leading a product infrastructure group, and my director was in the meeting, vouching for me, and all the other directors in the meeting were saying "Yeah, but maybe she's not technical enough. Like, frontend... I don't know. She doesn't have enough background, she doesn't have other experience." And my director was like "You know, she wrote a book about web performance, right? I think we can say that she's tactical enough to have the job of a director, like you all." And that's sponsorship, too. That was behind closed doors. I didn't know about that till much later, but that's also -- sponsorship can be invisible to you also.
+
+**Adam Stacoviak:** \[12:00\] Yeah. The thing too with any sort of any opportunity - it comes with more connection.
+
+**Lara Hogan:** Yes...
+
+**Adam Stacoviak:** Just blatantly. Not even like network connection, but just more connection. Whenever you're in conflict, the recipe to disperse and mitigate conflict is more connection, not less connection. So you would think that anybody in anything really would just thrive on more connection, so what you're providing with sponsorship is just a greater social network, a greater work network, more connection to more people that know -- whether you're good or not doesn't really matter. It's your chance to now show up and prove or showcase who you are, and that connection can pay dividends over time.
+
+**Lara Hogan:** I could not nod more enthusiastically about that. And it benefits you as a sponsor. You look so good by sponsoring someone else that goes and kills it on whatever they're working on. And you're building a new leadership bench. The only way that your organization - your team, or whatever - is gonna survive is if you have a group of strong leaders there to pitch in... And the only way to get them to the place where they can pitch in is to provide them with sponsorship opportunities.
+
+**Adam Stacoviak:** Well, the way I discovered some of your work was through your book, obviously, Resilient Management, and you wrote an excerpt for the A List Apart website, which we're super-huge fans of in like forever, basically.
+
+**Lara Hogan:** Mm-hm. Same.
+
+**Adam Stacoviak:** My entire career has been like "If A List Apart showcases it, it's awesome stuff." But you wrote an excerpt there... And I also wanna mention that you voiced your own book, which was super-cool, by the way. I didn't expect that, and then I'm like, "Oh, that's really cool."
+
+**Lara Hogan:** \[laughs\]
+
+**Adam Stacoviak:** But I've read that, and it kind of frameworks a lot of this - leaders, managers, you've got the idea of mentoring, coaching, sponsoring, and then delivering feedback. It's sort of like all the different hats you wear as a manager. How did you come up with this framework? Was it something you observed, was it something you defined yourself? How did you know what you know to now teach others?
+
+**Lara Hogan:** Yeah, so I learned a lot of what I know today from this person Paloma Medina. She was the Etsy learning and development director there. So basically everything that I talk about today, I learned in some way or was influenced in some way by Paloma's work.
+
+When it comes to mentoring and coaching, two of the skills, I've definitely learned the difference between those from here. But really, I started getting interested in sponsorship when I was trying to help the staff engineering group at Etsy start to add people who were not meant to the group. I was trying to figure out what are all of the different ways that we build really homogenous leadership teams, what are the contributing factors... And I started doing a lot of reading about in-group bias, I started reading about all of the different ways that we might start to measure the success of changing the demographics in a group over time... And I came across the work of someone named \[unintelligible 00:14:44.08\] who's done a ton of work on sponsorship and research on sponsorship. And she had this quote that stuck with me ever since then... It's about women, but I think we can apply it to any member of a minoritized community. Members of minoritized communities are over-mentored but under-sponsored.
+
+I started to realize, if I wanted to try to enact change within this staff engineering cohort, I needed to help these folks understand the difference between the two things. Understand that what they were doing was taking members of minoritized communities out to coffee to teach them what they know, but that's not the way to make this group more diverse. The way to make this group more diverse is to help make sure that the people getting the next opportunities to do visible, valuable work are those members of the minoritized communities, and support them in that way. That was really a pivotal moment for me in understanding the power of this work, and also trying to spread the good work about it to folks who are in positions of privilege and authority who can do it.
+
+**Adam Stacoviak:** Yeah. That's a challenging position to be in too, sponsoring folks. It's almost like "Do they need my help? Maybe..." Back to that self-doubt, potentially, or the impostor - do they need my help? Maybe they just need my advice. They didn't even ask for it. Should I even give it?
+
+**Lara Hogan:** Right, right.
+
+**Adam Stacoviak:** You almost to some degree can be paralyzed by the ability, but I guess if you know you can help somebody, you know you can help somebody.
+
+**Lara Hogan:** \[16:03\] Right. There's a few there, of - what if this person fails at the thing that I'm sponsoring them for? And that's real, that's valid. There's always that risk. But then what I try to do is coach people through like "Okay, what happens? Let's play that out." Like, what's the worst possible scenario? And then what can we do to prevent that scenario from happening? What are the check-ins? What are the other forms of support that this person might need to be successful? How are we measuring success? How are we making sure that we are making this an accessible leadership opprotunity to those who don't already have the implicit power or privilege that we do?
+
+There's a lot to unpack there, but it's a reasonable fear, and I want to make sure people understand that that's not an abnormal one, but it's totally surmountable.
+
+**Jerod Santo:** Yeah. Because when you vouch for somebody, you're lending your reputation. And so if they fail, then your reputation is harmed, and the next time somebody turns to you for advice -- maybe they don't turn to you anymore, because they're like "Well, you let me down the last time." I like what you said there with regard to the support, because in order to hedge your own risk - well, what do you do? You don't just provide the vouching or the opportunity, you don't just open the door, but you also give them what they need to walk through the door, and make it through to the other side. That's really powerful.
+
+**Lara Hogan:** And that might also be advice, but it's really important to ask the person what they need, rather than just assuming that what they need is still more of your advice, because that may not be it. One of my most mind-blowing moments was realizing that what someone needed was access to a meeting where decisions were being made... So I invited them to shadow me at this important meeting. And honestly, it opened up so many doors. Just the act of being there, silent, with me. There's so many different ways that support can manifest, and advice is just one.
+
+**Adam Stacoviak:** Yeah. The one shadow I can see happen -- well, actually, I got to witness it first-hand was when I sat down with Sid Sijbrandij from GitLab, and he is die-hard on his shadows going everywhere.
+
+**Lara Hogan:** Nice.
+
+**Adam Stacoviak:** So my show, Founders Talk, is really one-on-one, and so we sell it as one-on-one. So when you have more than one in the room, it just can kind of put some performance operations there, potentially; like, you might perform because someone else is there, or act a certain way...
+
+**Jerod Santo:** It changes the vibe.
+
+**Adam Stacoviak:** Yeah, it can really change the dynamics. So as a believer in those dynamics and the one-on-one, I wasn't trying to advocate that they shouldn't be there, but just for the reasons of the show and how we did it... And I understood where they were coming from, the shadows came anyways, and that was totally cool with me... I agreed to it beforehand; he didn't just show up and I was like "Okay, how do I deal with this?" I agreed to it. I thought it was a good idea, given what that CEO's shadow program does. And I just think that's super-cool to do that, because you get to really, like you said, you get to open those doors and just being there silently, but you get to witness somebody doing their thing. And sometimes you can just learn something by just osmosis. You just sort of see that happen, and boom - you've gained this new opportunity because you saw it happen, and now you're thinking "Well, i could probably do that, too."
+
+**Lara Hogan:** Absolutely. Absolutely. And you get to see so many different ways that leadership manifests. That's something else I think a lot about, is there's not just one way to be a leader, and to be able to have access to a diverse group of leaders shows you all of the different ways that you might wanna hone your own approach to leadership too, and there's something really powerful, it sounds like, in what they're doing, allowing them to not just witness them performing, or being a part of a podcast, or being a CEO, but all of the people they're interacting with, too.
+
+**Adam Stacoviak:** Yeah.
+
+**Lara Hogan:** That's cool.
+
+**Adam Stacoviak:** You can see a new side of that person too, and gain - or probably lose - some respect from, depending on how you act...
+
+**Lara Hogan:** There's always that risk... \[laughs\]
+
+**Adam Stacoviak:** Yeah, there is some risk there, but one thing I love particularly about Sid is that he seems to be the same person no matter where he's at. He seems to showcase the same values... And that could be a filter, I don't know, because I only really see him in, I suppose, professional spaces - and I've met him personally, too - but I've never observed him anywhere where it's like "Okay, this is a different Sid." Or "He's angry, so now he's this way." Or "There's pressure, so he acts this way." And he always seems to be the same kind of cool, calm, collected, calculated Sid that thinks, but then also has empathy and compassion in his words and responses, and also seems to be a good person to admit when he's wrong and willing to change. I think even witnessing that face-to-face in a shadow program is enough to change somebody.
+
+**Lara Hogan:** \[20:18\] Totally. I talk a lot about the idea of our default approach in a lot of my workshops and coaching sessions, just because we all do have a default approach to leadership... How it looks to other people, it's a default approach. But it's really important to switch it up when the context calls for it, so I'd be really curious - maybe it's a question I can ask him someday... It's like, "Okay, in what situations does that approach not work for you, or not work for your team? What are the opportunities that you have to switch it up based on what's needed?" It's hard.
+
+**Adam Stacoviak:** Yeah.
+
+**Jerod Santo:** So you've mentioned the difference between sponsorship, mentorship and coaching... I wanna dive into the coaching bit a little bit, but put a pin in that first, because you're also talking management and leadership, and I wonder if you differentiate those two, or if you don't. And if so, what are the differences?
+
+**Lara Hogan:** Great question. \[unintelligible 00:21:01.23\] tradition of trying to define these two very broad, mixed-up categories... \[laughter\] Honestly, just like mentoring and coaching tend to also be really conflated a lot of the time. I do find a distinction between them a lot, but just like a square is a rectangle, but a rectangle may not be a square - management and leadership are a little bit like that... Where you can be a manager and also a leader, and you can also be a leader and also a manager, but sometimes they're really distinct.
+
+I think a lot about facilitators in meetings. Someone who's a facilitator of a meeting - they may not be managing anybody, they may not be managing the content, but they're probably leading the meeting in a way, even if they don't have any interest in what's happening. That's a pretty clear distinction to me. Every company is different in how they define it. In a lot of companies, staff folks - meaning staff engineers, staff designers, whomever - are leaders, but they don't have distinct management responsibilities. But again, it's all kind of messy...
+
+When I talk about management, I talk about it in terms of "Are you supporting other people as they grow?" That actually applies to people who are not managers too, which is why I like to use the definition of management instead, that way I can say "Okay, these skills may apply to you, whether or not you have technical HR responsibilities in your role."
+
+**Jerod Santo:** Gotcha.
+
+**Adam Stacoviak:** Do you think a leader is more on like the inspiration side of things than the doing of the details/tasks thing? Because management seems to be task-based(ish), and then leadership seems to be more visionary, that kind of way, where you're sort of inspiring greatness in others, or to go or believe in the direction. It kind of requires a bit of salespersonship to sort of say "This is our mission, whether you like it or not." And maybe they don't deliver that way. Maybe it's more like "This is our mission because... And this is why I believe in the mission."
+
+**Lara Hogan:** Right.
+
+**Adam Stacoviak:** So that seems to be more leadership. Maybe that's mainly how I break it down - it seems more like inspire vs. tasks, and do, and organization.
+
+**Jerod Santo:** Versus require.
+
+**Adam Stacoviak:** What did you say?
+
+**Jerod Santo:** Oh, I said inspire vs. require, just because it rhymed. Sorry.
+
+**Lara Hogan:** \[laughs\]
+
+**Adam Stacoviak:** So that was actually what I was gonna say, but I thought it was a little too cliche... Inspire, require...
+
+**Lara Hogan:** \[laughs\] That's hilarious...
+
+**Jerod Santo:** I'm here for you...
+
+**Lara Hogan:** Mind-melt. So actually, I've worked with a lot of managers that do that, and that's actually a part of their job. They can't get everybody in a row in the same direction unless they do that, too. And leaders often have a lot of tactical responsibilities. So I actually more think of it as a spectrum between being empowering and being directive. And a leader and a manager can end up anywhere on this spectrum, and we all have a default on that spectrum.
+
+My default is actually on the empowerment end. In the absence of context, I will ask everybody a million open coaching questions, which - Jerod, I know you wanna talk about coaching...
+
+**Jerod Santo:** Yeah.
+
+**Lara Hogan:** But that doesn't always work. Sometimes you need to be directive. Sometimes you need to mandate a desk move. Sometimes you need to tell someone what their job is. Sometimes you need to give feedback that's not coaching, but instead it's like "I need you to do this thing right now." That's directive. Again, some people default to the directive end of the spectrum. They're like "Okay, here's what I need. Here's how it's gonna happen. Here's how we're gonna move forward." But they don't leave lots of room for empowerment. They don't ask lots of questions. They don't find sponsorship opportunities. They don't ask people how they wanna grow.
+
+\[24:11\] So again, based on the circumstance, based on the situation, we all have a default, but that default won't be useful to us in all circumstances. Strong leaders and strong managers need to know when to switch up their approach based on the given circumstances. So I would argue that managers and leaders both have to do these things, and company-specific responsibilities really just change.
+
+**Adam Stacoviak:** I like that you called it a spectrum, honestly, because I almost feel like you can take a test and see where you land on that...
+
+**Lara Hogan:** Yeah...
+
+**Adam Stacoviak:** You can even evaluate the thing you're doing currently, and like "Which direction is this?" Is it sort of center in the spectrum, or...?
+
+**Lara Hogan:** Yeah, in one of my workshops we go through that. I give a bunch of example scenarios... I actually have people draw the spectrum and mark on the line where they would end up based on the scenario, like "You've found out that one of your direct reports is being named to your other teammates in a meeting. Do you respond with empowerment or direction?" "A senior leader is talking over you and interrupting you. Empowerment or direction?" "One of your direct reports comes to you with a competing job offer two times their current salary. Do you default to empowering or being directive?" Again, it's really interesting to see... And in the workshop we have everybody hold up their line, so we can see where people land...
+
+**Jerod Santo:** At the same time?
+
+**Lara Hogan:** Yeah. Some people are really close together in where they landed, some people are all over the place. And again, I think it's just a way to say, we have to adapt based on the circumstance. \[laughs\]
+
+**Adam Stacoviak:** Yeah.
+
+**Jerod Santo:** That sounds fun.
+
+**Adam Stacoviak:** That's interesting.
+
+**Lara Hogan:** We have fun.
+
+**Adam Stacoviak:** Well, I think it's kind of theory, to some degree, where do you land...
+
+**Lara Hogan:** Yes.
+
+**Adam Stacoviak:** So you have to understand the mechanics behind things and the frameworks behind things to maybe even determine that...
+
+**Lara Hogan:** Oh, yeah.
+
+**Adam Stacoviak:** And I guess you could somewhat guess too, if you weren't really skilled in the theory and the details behind things... But it really does give you an idea of like "Okay, this is where I'm defaulting to", as you've said a few times. "Maybe that's not a good place to default to, but it's okay. It's the truth." We can all sort of grow from the truth currently. Don't lie to yourself, basically.
+
+**Lara Hogan:** Right, exactly. Let's be honest.
+
+**Adam Stacoviak:** If you're defaulting and you're erring on the side of like "Well, that's not actually optimistic or that's not a good way to be", then you've gotta change. But at least you know where you're at.
+
+**Lara Hogan:** Yes. You've gotta be honest about it. And also be honest about where that default is most successful. One of the circumstances - what's the environment in which your default way of operating works best? That way you can start to think about "Okay, now when are the times when I might need to switch it up? And how can I experiment with that to know when it's time for me to switch it up? How can I know when to do that, and that really effectively, too?" \[laughs\]
+
+Break: \[26:32\]
+
+**Adam Stacoviak:** So Lara, we mentioned your excerpt from your book, Resilient Management, on A List Apart. I love that. And that's kind of where I saw this framework... I guess - full disclosure, I haven't read your book in completion. I have listened to some of it, to give myself a little pat on the back, but... You laid out this framework quite well...
+
+**Jerod Santo:** \[laughs\] Doing her a favor... "I listened to a little bit of it..."
+
+**Adam Stacoviak:** I listened to it a little bit... \[laughs\] I wanna listen to all of it. Actually, I think I'd rather listen to it than read it, now that I've heard it, because you've got a great voice; I think you speak very well through it, and I'm actually much more of an Audible listener than a reader. I do like to read, I said this yesterday on some outro of JS Party, or something like that... I do read, but I prefer to listen.
+
+**Jerod Santo:** \[laughs\]
+
+**Adam Stacoviak:** So with that aside, you mentioned how leaders, managers - they end up wearing these different hats, four different hats distinctly: mentoring, coaching, sponsoring and delivering feedback. Kind of go in to the details of what each of those mean, and we can sort of go from there.
+
+**Lara Hogan:** \[29:59\] Absolutely. So mentorship is all about sharing your perspective, sharing your advice, sharing what you've see work and not work, suggesting pitfalls that this person can avoid... It's all about sharing your own knowledge and handing it over to this person. So mentoring is really useful when someone's blocked and just needs a little bit of help getting unblocked, and it's also really useful for when someone's being onboarded to a new role, a new company, whatever. Those are really the only two use cases where mentoring is useful and powerful. The other three skills of coaching, sponsoring and feedback are the ones that I try to emphasize that anybody who's supporting other people should really focus on and hone their skills in.
+
+So coaching is all about asking lots of open questions and helping this person reflect, connect their own dots, develop their own brain wrinkles... So it has nothing to do with your own knowledge, which is very hard for many of us that have been taught that our knowledge is the value that we provide.
+
+So open coaching questions, championing someone, reflecting back what you're hearing them say - these are all tools that we can use to help someone kind of sit back and say "Alright, am I actually looking at this from all different directions? Have I truly identified the problem statement at hand, that's the real one that I wanna tackle? What's hard about this? What's surprising about this?"
+
+Sitting in that messy space unlocks so much more than we could have ever unearthed before, and that's where the power comes in - the ability to kind of look at something from all different angles, not just the rote, surface-level ones, and figure out what we need to do specifically to move forward.
+
+Sponsorship is feeling on the hook to get someone to the next level by putting their name in the ring, giving them visible opportunities to do valuable work, again, connected to business goals ideally. And then feedback - we all know feedback. Feedback is giving someone a steer, you know, as specific and as actionable as possible, to how they could do something even better or why what they're doing right now is really working. I could go on all day about those last three skills, why they're powerful, and how to do them.
+
+**Jerod Santo:** For sure.
+
+**Adam Stacoviak:** As you were saying that though, about coaching, I kind of thought, since I've got some experience with therapy, and working with psychologists, and co-hosting a podcast with a psychologist - they seem to really camp out on coaching.
+
+**Lara Hogan:** Oh, yeah.
+
+**Adam Stacoviak:** Because the thing I grokked most from a therapist and somebody receiving therapy is they wanna help them find the path by asking questions and let that person determine what's the right move.
+
+**Lara Hogan:** Bingo.
+
+**Adam Stacoviak:** Because it's not about them telling the patient what the right move is, or the determination, or whatever it might be, it's about helping them discover where they need to go. Because let's face it, if you come to your own choices, you have so much more power. When you're an advocate in your own choice, rather than given a choice, it's not really a choice. So you can feel far more comfortable in that next step(s), because you came to that conclusion yourself.
+
+**Lara Hogan:** Absolutely. And you're unique, your journey forward. You might be pulling on information from other people to help you shape that journey forward, but it's gonna be unique to you. No one else can give you the answers that you need. So that's why I personally find coaching really valuable, is everybody's so unique I can't assume that my experience is gonna map to theirs, or my solution could be useful to them. What's way more powerful that I can provide these days are open coaching questions. And open coaching questions - you know, they're not closed, they can't be answered in yes or no or a number, and they're not leading, like "What if you tried this mentorship, this advice?" That's giving a solution.
+
+So in coaching mode, which is really, really hard to stay in - in coaching mode you're asking genuinely curious, open, broad questions, and I find that the best ones start with the word "What", like "What's important about this? What's hard about this?"
+
+My number one favorite open question to use is "What are you optimizing for?" Because in any different situation, everyone's optimizing for something different. But we never think about it. Whenever I ask someone -- let's say we're talking about a conflict that they're having, and they can't figure out how to move forward. I'm like, "Alright, so in this situation, with this relationship, what is it that you're optimizing for?" Usually, they'll sit back in their chair and be like "Huh..." And that's what I'm looking for as a coach. \[laughs\] I want that moment of \[unintelligible 00:34:06.28\] like "I'm stumped for a second. I need to actually sit down and process this." That's where the magic of coaching happens.
+
+**Adam Stacoviak:** \[34:14\] Yeah.
+
+**Jerod Santo:** I laugh, because this is Adam's favorite statement, isn't it? "What are you optimizing for?" He says it all the time.
+
+**Lara Hogan:** Is it?
+
+**Adam Stacoviak:** It is. I think you have to know where you're going. You can't take the steps forward with any assurance if you don't know where you're trying to go. Otherwise you're just sort of haphazardly moving along your path and you think you know where you're trying to go, but if you haven't really considered "What am I optimizing for? What am I really trying to do in this situation?", your list becomes so much more concise. You can go from 8 to 4 really easily, or maybe 8 to 1, because you're like "These things are not important, so they don't matter. At least today."
+
+I think a lot of that comes from me reading -- gosh, what is that book called? What is it called, man?
+
+**Jerod Santo:** It's called --
+
+**Adam Stacoviak:** Essentialism. Essentialism. A lot of that book is really about getting rid of the trivial many in the vital few. You have to whittle down to the things that really matter, and you can't get to the vital few unless you get away with the trivial many... And you can't do that unless you know what you're optimizing for.
+
+**Lara Hogan:** I couldn't agree more.
+
+**Adam Stacoviak:** What's essential.
+
+**Lara Hogan:** Well, and when I'm working with coaching clients and I ask them the "What are you optimizing for?" question and they give me two answers, I force them to turn it into an "even over" statement. Like, "I'm optimizing for speed, even over quality."
+
+**Jerod Santo:** Prioritize those two.
+
+**Lara Hogan:** You're right. You've gotta get down to just one.
+
+**Jerod Santo:** Yeah.
+
+**Jerod Santo:** You can time-box it, like you said. Like, "In this season, I am optimizing for blah." Because you know it's gonna change, and that's cool.
+
+**Jerod Santo:** Another question that you have on your list which I ask guests all the time on the show which I really like is "What does success look like?" And that's a hard question to answer. I think we get -- multiple people kind of stop and have to think for a minute... And that's really what you're trying to get them to do as a coach, is to think for yourself and then go along for that ride.
+
+**Lara Hogan:** Yeah, totally. And we spend so much time talking about failures, it's hard to remember we also need to think about what success looks like, too.
+
+**Adam Stacoviak:** Yeah.
+
+**Lara Hogan:** One of my favorite questions also on that list that I stole from Paloma Medina, who I mentioned earlier - is "What is the worst possible outcome?" and then "What's the most likely outcome?"
+
+**Jerod Santo:** Yeah. Because we tend to focus on what's the worst thing that could happen, right? Which is the first part of the question.
+
+**Lara Hogan:** Right.
+
+**Jerod Santo:** Because it's worth knowing, it's worth thinking about that. But then also putting that in context of like "Now, how likely is it that that one happens?" And then usually, it's like, the reason it's the worst is it's literally the outlier of the statistical things that could happen. It doesn't mean don't think about it, but let's not give it so much weight in our minds. What's more likely to happen?
+
+**Lara Hogan:** Yeah, we can honor it. We can honor it by naming it. And then we can focus on the more likely outcome. \[laughs\]
+
+**Jerod Santo:** Yeah, exactly. So as I read through your different terms here, I guess the coaching was the one that has tripped me up. I don't wanna get too much into a semantic discussion about words, but coaching - I just bring a lot of context to coaching mostly in my life of sports, where I've been a coach and I've also been an athlete on a team... And it seems like what I think of coaching as prior to reading this - really actually what a coach does on a sporting team is actually all these things... Because they are instructive, they are giving you what you ought to do, but then they're also asking you questions and helping you grow, and they're also putting you in the captain spot, or "Hey, you're pitching today." They're sponsoring people.
+
+So I think when I saw coaching, I was like "Is this really what coaching is?" But I wonder if that's just like the difference between business, or maybe adult-to-adult coaching, versus sports... And maybe there just aren't lines that we can draw between the two.
+
+**Lara Hogan:** \[37:47\] Totally. Yeah, I think that that's one of the challenges, is that the term "coach", in specific contexts, absolutely means a blend of all those three things. In sports... I think I might even call that out in the book specifically - if you have any sports background at all, you might think of coach as probably actually more of a mentor. Most people who I talk to, when they think about coaching, or being coached on a sports team, they talk about mentorship. But you're right, it's not just mentorship, it's not just giving advice.
+
+**Jerod Santo:** A lot of feedback. You give a lot of feedback.
+
+**Lara Hogan:** Exactly. Absolutely. And as you said, sponsorship too, right?
+
+**Jerod Santo:** Yeah.
+
+**Lara Hogan:** So - totally. When it comes to the term "coaching", I think that's distinct from the role. And you're right, the role in a business context is like a leadership coach, or a life coach.
+
+**Jerod Santo:** Life coach, yeah.
+
+**Lara Hogan:** That's the kind of role that's distinct from, let's say, a sports coach.
+
+**Adam Stacoviak:** Yeah.
+
+**Lara Hogan:** One hundred percent.
+
+**Jerod Santo:** So dive more into delivering feedback... Because I think this is something that we all do, and we all don't always do well. You mentioned actionable...
+
+**Lara Hogan:** Specific...
+
+**Jerod Santo:** Specific... Is there any other heuristics of like "Here's how to give good feedback versus not so good"?
+
+**Lara Hogan:** Oh, yeah. I mean, I can give an additional whole episode just about feedback... \[laughter\] Originally, the chapter in the book had a whole different, separate feedback chapter, and then I started to realize it actually relates to mentoring, coaching and sponsoring, because we can do those three things of mentoring, coaching and sponsoring when we're giving feedback, too.
+
+So when I think about good feedback, a lot of the way that I think about it derives from SBI (Situation, Behavior, Impact), that framework. Situation - we get people to talk about the facts. What's my observation of the behavior I'm giving feedback on? Just the facts means not my assumptions, not my judgments... I'm trying to make sure I'm keeping my own opinions at the first part of the feedback, because what we wanna avoid is someone receiving a feedback getting amygdala-hijacked. So we wanna avoid their fight or flight response kicking in. And you all know, same is true for me - if I'm getting feedback, if I even smell feedback is coming, my whole body freezes up. My amygdala, the part of my brain that's responsible for trying to keep me safe and be on the lookout for threats - it senses a threat that's headed our way.
+
+**Jerod Santo:** Yeah.
+
+**Lara Hogan:** A lot of how I think about feedback has to do with trying to keep that amygdala chilled out and keep our prefrontal cortex, the rational, logical, practical part of our brain online.
+
+So just the facts means let's make sure that this person's amygdala is like "Yeah, that's true. That's all happening. What you described is real, and I can attest that that is a real thing." So start with the observation.
+
+Next though we can talk about the impact that this person's behavior has had. Traditionally, we've been taught to frame the impact, the behavior based on what we as the feedback giver care about. Let's say someone's writing really terse emails to me... I'm mad that all the time they write 4-5 words, it's a waste of my time, I always have to respond, and get more information... I feel like they're mad at me all the time. Those are all my perspectives and why I wanna give the feedback. I challenge everybody who's giving me feedback to instead of saying why you think that this feedback is important for the person to listen to, why did they care about it? What did they care about and how does this feedback relate to that thing?
+
+In the example of the person writing really terse emails, they probably care about getting this product done on time. So I might say "Hey listen, when you send me emails of this length, I need to respond and ask for more information, which adds time to the process, which means that the thing you wanna have done takes three times longer than it normally would." So you see what I'm doing - I'm distinguishing between \[unintelligible 00:41:11.15\]
+
+**Adam Stacoviak:** Yeah.
+
+**Lara Hogan:** It's really hard to put ourselves in their shoes, but again, we wanna keep their amygdala chilled out and their PFC online. So by describing it in terms of what they care about is really, really useful.
+
+And then the final thing I like to do is close with an open coaching question. Again, we've been taught to make a request, like "Therefore, could you please send longer emails?" But if you have hit the nail on the head with the facts, they're on board, and you've also hit the nail on the head with what they care about, they're already thinking about what they wanna do next. You don't need to tell them. That's mentorship mode. It's just saying "Therefore, could you please..." And usually, we'll come up with as good of an idea as they will. So instead of making this a one-way feedback dump, ask an open question, which turns into a two-way dialogue, where you're brainstorming together about what we could do going forward, what this person needs... Again, "What are you optimizing for? When you write emails, what are you optimizing for?"
+
+**Adam Stacoviak:** \[42:03\] Absolutely, yeah.
+
+**Lara Hogan:** Do any of the open coaching questions and see what that does for your feedback conversations, and hopefully you are lowering the chances of an amygdala hijack, of a fight-or-flight response happening in that feedback conversation.
+
+**Adam Stacoviak:** Yeah, that could be the bomb going off, like "Oh, come on now." And the next thing you know you're fighting, there's a conflict instead of a resolution, which - you really want the resolution...
+
+**Lara Hogan:** They've shut down...
+
+**Adam Stacoviak:** Yeah, they've shut down... There's no more communication, there's no more connection...
+
+**Lara Hogan:** No one wins.
+
+**Adam Stacoviak:** Yeah, exactly. I like that, that you lead with what you assume though what they're optimizing for. You're assuming because of certain observations - the length of the email, their characteristics, their behavior in meetings - that they care about time.
+
+**Lara Hogan:** Right.
+
+**Adam Stacoviak:** So you're sort of capitalizing on that and saying "If you really care about time, here's how we can--"
+
+**Jerod Santo:** ...spend less of it.
+
+**Adam Stacoviak:** "...here's the challenge we have together. Here's how it impacts you and me."
+
+**Jerod Santo:** Yeah, you're finding common ground as well.
+
+**Lara Hogan:** Absolutely. And you might get it wrong. You might get the impact part wrong. You may totally mistake what they actually care about...
+
+**Adam Stacoviak:** Yeah.
+
+**Lara Hogan:** ...in which case you're gonna have to try again, but only after you do some more digging. And if you do this prework upfront, you can spend time asking questions about what they're optimizing for, or what's most on their mind, what are they focused on the most right now, or what's worrying them, what's motivating them... Any of those kinds of questions can help you get the kind of data that you need before you give the feedback. And sometimes I like to actually open feedback with the impact that they care about first, like "Hey, I know you can't -- we talked so much about this. I know it's really important that we ship this by Tuesday, so I've got some feedback I'd like to give to make sure we could ship this by Tuesday." So you can see how by kicking it off with what they care about, their brain is like "Huh. \[unintelligible 00:43:39.21\] might get a little bit tense... But you know me, you hear me, you see me. Maybe I'll be open to hearing this feedback, because it seems like we maybe care about the same thing."
+
+**Adam Stacoviak:** And you're also kind of identifying the reward, which is kind of leading into the habit loop even...
+
+**Lara Hogan:** Totally.
+
+**Adam Stacoviak:** Because the habit is the email, and the habit is the length or lack of length of the email... So the reward really is like "Okay, I get back to work." But maybe you shift the reward and say "Okay, you get back to work, but we delay shipping because X."
+
+**Lara Hogan:** "But wait..." \[laughs\]
+
+**Adam Stacoviak:** So you're sort of like identifying their reward, or I suppose the team's reward. There's a lot of psychology really in this process.
+
+**Lara Hogan:** Oh, yeah.
+
+**Adam Stacoviak:** That's something I'm learning more and more... Especially as we look at your chapter titles, "Setting clear expectations. Communicate effectively" and then "Building resiliency." That's a lot of things that happens from a psychologist to a patient in therapy; it's very much a lot of this. When did you begin to study the brain, I suppose, to know the prefrontal cortex and the amygdala? You must have done some digging into research; which books did you read, how did you get curious?
+
+**Lara Hogan:** Totally.
+
+**Adam Stacoviak:** Where did you dig?
+
+**Lara Hogan:** I can tie all this, again, back to that same person Paloma Medina at Etsy, who has done a lot of research on organizational psychology, but also just like the neuroscience behind why humans behave the way that they do.
+
+Since she's worked at Etsy, she's done a number of incredible things, including starting a shop about productivity tools for our brains and the neuroscience behind productivity at work, psychological safety - all of these things. She's just incredible. So a lot of what I initially learned was from her. And honestly, like you all know, engineers - we just get so hooked on the data and the science behind things; it's an easy way in to get people to start thinking about "Okay, what are all the ways that we humans are behaving the way that we are?"
+
+One of the things that she taught me very early on was this framework, this acronym she calls BICEPS. And BICEPS stands for the six core needs that humans have at work. And again, this is all neuroscience, but also anthropology, and a lot of the social sciences trying to study why do humans behave the way that they do, and what do we need -- what do our amygdalas try to secure and ensure for us to keep us safe at work? And it's things like a sense of belonging, a sense of community. If we feel othered or left behind, our amygdala considers that a threat. A sense of improvement and progress, a sense of choice, fairness, predictability, all of these things in the BICEPS core needs - it's all neuroscience-backed. Thanks to Paloma I have learned so much more about all of those surprising ways that humans behave and deal with each other at work.
+
+**Adam Stacoviak:** \[46:12\] Yeah. How about any particular books that you've read recently?
+
+**Lara Hogan:** Absolutely.
+
+**Adam Stacoviak:** I can mention for me Atomic Habits has been so on my prefrontal cortex... I'm thinking about it a lot; I'm always thinking about habits, and habit loops, and just how the 1% improvement can change things. I'm curious what you're reading.
+
+**Lara Hogan:** A go-to is always "Switch: How to change things when change is hard" by Chip and Dan Heath. It's all about "How do I influence change in any context if I don't have power or authority or control over the situation itself?" They include a bunch of studies, a bunch of anecdotes, real case studies about how -- you all must be familiar with the "Don't mess with Texas" slogan, and where that came from?
+
+**Adam Stacoviak:** Yeah, I know the slogan. I live there.
+
+**Lara Hogan:** Yeah.
+
+**Jerod Santo:** Don't mess with Texas.
+
+**Lara Hogan:** Do you know where it came from?
+
+**Adam Stacoviak:** No.
+
+**Lara Hogan:** I also didn't know. I learned it from Switch. I'm gonna butcher this story, so apologies to everybody involved... Go read the book to get the real story. But at the end of the day, there was too much littering. There was a bunch of littering happening in Texas, and so the government tried to brainstorm, "Okay, how do we get littering to decrease?" And they made this ad with this cowboy picking up litter off the road and putting it in a trash bin and saying "Don't mess with Texas." And that's where it was born out of. And you know what - littering decreased significantly after this ad campaign kicked off.
+
+**Adam Stacoviak:** Wow.
+
+**Lara Hogan:** So again, it's all influence. It's all a lot of psychology that you mention, it's a lot of neuroscience-backed stuff... It's fascinating. How do we influence change to happen, positive influence where we don't have complete control on the situation?
+
+**Adam Stacoviak:** Yeah. Well, I think that may have worked - I'm gonna assume a lot, because I haven't read the book - because Texas seems to be very prideful about their land. The size of things, everything's bigger in Texas, things like that... And I think that probably -- and a cowboy is like a significant...
+
+**Lara Hogan:** He's a model.
+
+**Jerod Santo:** Yeah. The ethos.
+
+**Adam Stacoviak:** ...a model for a lot of people in this space. The cowboy hat, we have the rodeo here every single year, it's massive... I'm not from Texas originally, so I can't claim these things as native for me. I came here \[unintelligible 00:48:13.29\] which is a common phrase to a transplant to Texas... But I'm from the North-East, but long story short, I didn't grow up with cowboys and I didn't grow up with the things that I see here in Texas all the time, but I can see that's probably why it worked well...
+
+**Lara Hogan:** Oh, yeah.
+
+**Adam Stacoviak:** ...because they took this common model and the pride of the land and the respect for the space and the desire for it to be the best it can be and attached that to something that really matters, which is "Stop littering", right?
+
+**Lara Hogan:** Yeah. And that comes back to the BICEPS core needs list. The B stands for Belonging. We are a part of the community. We identify with this role model, we wanna support each other as the community of Texas... And Significance is the S. So in the hierarchy, we have a lot of pride in our land, in our state. Again, these two core needs - the book goes into this in so much detail, but those two core needs, again, our amygdala is working really hard to secure and to ensure for us... They're present, and it's a huge way to get people to listen, to get people to care... People have to be motivated to do something.
+
+Break: \[49:12\]
+
+**Jerod Santo:** So Lara, a lot of our audience are engineers, or were engineers and now are managers and leaders... And it's always a difficult thing deciding if that's a move for you. Once you've decided that it is, making the transition, like what does success look like in that case? You've done the transition a while ago, and now you're succeeding and teaching other people how to be managers and leaders... So advice for us and for our audience about how do you make a decision like that, and once you do, how do you navigate that successfully?
+
+**Lara Hogan:** I love this question. I talk to and work with a lot of people that have made the transition multiple times. One of my former colleagues, Dan, has written a bunch about how do you choose, how do you make a decision, how do you know when to go back and forth, when is it right... There's just so much good stuff out there. The way that I like to think about it is -- the same question we've been talking about, the "What are you optimizing for?" question. Get really clear on that, for yourself, because that's a precursor to knowing whether or not you can actually achieve that thing you're optimizing for in a different role.
+
+If you're optimizing for gaining more power, for example, and you're thinking about becoming a manager, you may not actually get more power. It will be really important for you to do some research, talk to some other managers to ask about what they have influence on, what they have authority over etc. because it's often a bad surprise.
+
+If you're optimizing for making a new team charter, a new vision, giving the team purpose - maybe that's a possibility at your company. If you're optimizing for a change of pace, \[unintelligible 00:53:59.20\] and that could also breed additional questions.
+
+\[54:09\] So getting clear about what you're optimizing for will allow you to figure out whether or not you can have success in the other role, in either direction. People who skip this step have asked themselves "What am I optimizing for when I'm thinking about changing roles?" often change roles and then hate it... Because they find out that they can't achieve the thing that they were optimizing for, because they were working under a bunch of assumptions, or that's just not how it works, so the role doesn't have as much authority or power as they thought it would.
+
+When I was experimenting with leadership and management and thinking about that, I just really wanted to support the people that I was working with day-to-day... I was working at a really small startup, and we were basically UI/UX frontend site performance developers, and we didn't have a manager that knew any of that stuff... Our manager really only knew Perl. And as much as we could spend all day talking about Perl, I really wanna be able to provide a little bit more structure, a little bit more career progression for folks that were experimenting at the time with CSS 2.0. "Oh, my goodness..." You know, thinking about those sorts of things...
+
+So for me it was about providing team environment where we could self-identify as this frontend/UX group... And so for that reason, becoming a manager worked. I was able to achieve the thing I was optimizing for with that role change.
+
+There have certainly been times when I've thought about other kinds of role changes and I wanted to figure out "Could I be successful in that new role based on what I was optimizing for?" and the answer has always been different. So yeah, I would definitely suggest thinking to yourself and getting really clear with yourself about what you're optimizing for before making a switch.
+
+**Adam Stacoviak:** Yeah. Something that I learned while doing the show Brain Science which I've mentioned before, with Dr. Mireille Reece, was -- a thing she had said was "Try it on." So any new big decision, anything really in particular, so in this case this transition - try the thing on. How could you sort of get a feedback loop quickly. I guess step one would be "What are you optimizing for?" Okay, sure, I do wanna think about this decision or consider it. How can I first take a step to try it on. How can I try this decision on temporarily? How would my drive to work change? And maybe in these times there's no drive to work. How would my interactions with my counterparts that I really care about, like in your case - I wanna step up and help in these ways, because I care about this team and what we're trying to do... How would just my relationship change as a result of this?
+
+**Lara Hogan:** Totally.
+
+**Adam Stacoviak:** How does my relationship change at home, beyond just simply working -- sometimes we make work decisions or professional decisions and kind of forget that we have other constraints that don't map to that opportunity. And like you had said, if you don't consider what you're optimizing for.
+
+In our case, for us running this business, Jerod and I - we have certain personal constraints that help us navigate choices we make in the business. Things we just definitely probably will never do... And I have a hard time saying absolutes; Jerod, you probably know me to do that, so that's why I say "probably never", but maybe not... But the point is that we have certain things that give us structure and constraints to say "Okay, when we make that transition or we do that kind of thing, how can we try it on, or know what we're optimizing for to do that?" That's such a key phrase. "What are you optimizing for?" is such a key phrase in so many places. Really, it is.
+
+**Lara Hogan:** Yeah, absolutely. You go to the grocery store, and try to decide between two different apples... What are you optimizing for? \[laughter\] \[unintelligible 00:57:24.12\] Yeah, absolutely.
+
+**Jerod Santo:** Price, freshness, bulk...
+
+**Lara Hogan:** \[laughs\] But back to that point of trying it on - Camille Fournier's book "The Manager's Path" is just so brilliant in so many different ways, and one of the ways is that her first few chapters are all about before becoming a manager, what are the different ways you might try on some of those responsibilities.
+
+**Adam Stacoviak:** Very cool.
+
+**Lara Hogan:** So in what way can I be a mentor, in what way can I be a tech lead...? Trying it on before you actually go in. So if folks are trying to figure out how they could try it on, definitely check out Camille Fournier's book, "The Manager's Path." It's great.
+
+**Adam Stacoviak:** \[57:58\] We're a fan of Camille, and she also wrote the intro to your book, the foreword for your book...
+
+**Lara Hogan:** She did.
+
+**Adam Stacoviak:** And then she also obviously wrote her own book, The Manager's Path, and I think as part of her story too, a blog post you had written while she was writing that book was so influential to her, she had to kind of incorporate some of your ideas... So I definitely wanna give a shout-out to Camille. She's awesome.
+
+**Lara Hogan:** Hey, Camille! \[laughs\]
+
+**Jerod Santo:** Do you find -- and I'm asking you to generalize here... Do you find that engineers make good managers?
+
+**Lara Hogan:** That is an impossible question to answer. I like it.
+
+**Adam Stacoviak:** \[laughs\]
+
+**Jerod Santo:** Give it a shot.
+
+**Lara Hogan:** People of any ilk make great managers and terrible managers. I find that one of the hardest parts about transitioning to management from any individual contributor worker work is the lack of measurable progress. So in engineering you're shipping things, you've got code reviews, you see things launch. You actually have observable progress forward. In management a lot of your work is invisible, or there's a long lead time between when you do a thing or say a thing, and you see the results of that thing, if ever. You might do a thing or say a thing and think it's incredible, and then nothing ever happens.
+
+One of the ways that I coach managers struggling with this is just start to think about celebrating tiny, tiny wins that might have to be confidential with a supportive group that you trust to keep those wins confidential. Someone can be like "Alright, that rework - it's done. I can't tell if it's good yet, but it's done. Let's celebrate." Finding some way to basically measure and observe progress in yourself and in your work as a manager is challenging, certainly.
+
+**Adam Stacoviak:** Yeah, I like celebrating the wins, too. We say that a lot around here. Even one time I think we were in a call - I can't remember how far back, but Jerod was just sort of like down a little bit and he's like "I just need a win. I just need to ship something and get it out there and just know it's done." He's very task-oriented, and if there's a list in front of him that's not done, he's fretting about it. He wants that list to be done.
+
+**Lara Hogan:** I feel that.
+
+**Adam Stacoviak:** And that's a different context, but celebrating the wins is something that I like to say a lot too, for us even, because I think too often you just forget to -- like, if you waited to celebrate, it's almost like process over goals. If you only celebrate the goal and not the process - that's kind of like the thing, like that's a process. Getting the work done was part of the process.
+
+**Lara Hogan:** Yeah.
+
+**Adam Stacoviak:** It may have been the goal too, but if you don't celebrate the little parts of those processes happening, getting done, and you wait till the goal, the goal is always moving. Almost always moving. It's always sort of changing, and you sort of camp out in this process land and just celebrate those wins.
+
+**Lara Hogan:** Yeah. I wrote an article for Increment on this exact topic. For anybody that has invisible work, how do you start to identify where you can measure or even just mark those wins? I borrowed a lot of research about microtasks and how our brain gets that sweet, sweet dopamine hit from the act of checking something off of a list. And that's what we crave, that dopamine.
+
+**Adam Stacoviak:** Yeah, it is. It's all for the dopamine. Dopamine is a -- I'm sure everybody knows what it is, but it's basically this chemical in your brain that you love to get, that is a reward factor. I'm paraphrasing terribly, of course, but it's this thing in your brain that occurs when you do things that you like.
+
+**Lara Hogan:** Yeah, yeah.
+
+**Adam Stacoviak:** It's feedback to say "That was good. I'd like to do more of it, basically." So to paraphrase --
+
+**Lara Hogan:** Your brain even releases it in anticipation of checking something off of your list. It's great, it's awesome.
+
+**Adam Stacoviak:** Yeah.
+
+**Jerod Santo:** And then what happens if you don't get the thing checked off? Does it go back?
+
+**Lara Hogan:** It's a bummer. It teaches your brain that maybe we don't need to do this... \[laughs\]
+
+**Jerod Santo:** Yeah. Maybe this person's a failure. So is that article called "Your brain on progress"? Is that what it's called?
+
+**Lara Hogan:** It is, indeed.
+
+**Jerod Santo:** Alright, we'll put that one in the show notes, along with all the other things.
+
+**Lara Hogan:** Thank you.
+
+**Adam Stacoviak:** I wanna read that too, because I haven't read that one yet, and I'm all about progress. What about stories? I'm sure someone like you has tons of stories... We didn't well-define, I think, some of your past history; you mentioned somewhat where you've been before - engineering director at Etsy, VP of engineering at Kickstarter, you mentioned a couple... I'm not sure if Etsy was that startup -- I think you were there a while back...
+
+**Lara Hogan:** \[01:02:02.17\] No, I was at a DNS company before Etsy.
+
+**Adam Stacoviak:** Okay.
+
+**Lara Hogan:** So at Etsy, actually -- when I think about stories, a lot of them are from Etsy, because it was such a formative time in my career, just being at that kind of organization, under the leadership that I was under... You just learn so much about being a human around other humans that care about humans. A lot of the work that I do today can be directly tied back to that time. And when I think about war stories, horror stories, I think about the relationships that I formed there that I've carried on today... Like, what did I survive, what did we all survive there? And I think about a time when I actually had to fire someone for the first time, and I wasn't getting the support that I needed from my manager at the time. He was checked out, he was unavailable, he just couldn't help, and I didn't know what I was doing. I didn't know if I should fire this person, I didn't know if I was making the right decision, what's the process, what do I need to make sure I do, how do I make sure that I'm supporting this person...?
+
+I ended up basically hanging out with a bunch of other managers in this program called \[unintelligible 01:02:54.29\] that Paloma had created, where - it was a confidential, kind of like support group for managers across the company. It was a cross-functional group, really small groups. I think there were maybe eight people in my group. We were all kind of baby managers, and I was sharing with them confidentially the questions I had, and how challenging it was and how stressed out I was, and just how sad I was about it. And even though we were all baby managers, they all had different experiences before of going through something like this, within different organizations even.
+
+So one person offered to roleplay the difficult conversations with me, so I could get practice and really hone what I needed to say. One person helped me really get a good gut-check on how to make this decision, make sure it was good for everybody involved. One person had experience \[unintelligible 01:03:39.21\] so she gave me advice on that process and connecting me with someone to talk to.
+
+And by their powers combined, they got me through. Each of their different experiences, different skillsets were so, so, so incredibly helpful and powerful to help me survive. They took me out for drinks afterwards, so I could just kind of cry a little bit, you know...?
+
+**Adam Stacoviak:** \[laughs\] Yeah.
+
+**Lara Hogan:** And I realized this power in this group was so important, because your manager can't be your everything. Your manager has one particular set of skills; one particular set of places that they can sponsor you for, or types of feedback they can give. Maybe they're even bad at feedback, maybe they're bad at mentoring, maybe they're bad at coaching. It could be everything. You need a group, you need a collection of people that I like to jokingly refer to as a Manager Voltron. Not just one manager, but who's in your Manager Voltron?
+
+I don't know if listeners are familiar with the concept of a Voltron. I don't know if you wanna give a definition... \[laughs\]
+
+**Jerod Santo:** Go for it, Adam.
+
+**Adam Stacoviak:** What's the definition, Jerod? Are you a Voltron fan?
+
+**Jerod Santo:** I'm just checking out Voltron for the first time. I'm trying to grok this...
+
+**Lara Hogan:** \[laughs\] Amazing.
+
+**Jerod Santo:** Full disclosure...
+
+**Lara Hogan:** What does Wikipedia say?
+
+**Jerod Santo:** Voltron - of course, everybody knows - is an animated television series franchise that features a team of five space explorers who pilot a giant super-robot known as Voltron.
+
+**Lara Hogan:** Voltron! Yeah, bingo. If you were familiar with Captain Planet growing up - very similar vibes.
+
+**Jerod Santo:** Oh, yeah.
+
+**Lara Hogan:** By our powers combined, Earth, Wind, Fire, Air, Heart, or whatever...
+
+**Jerod Santo:** Exactly. Captain Planet.
+
+**Lara Hogan:** Yeah, exactly.
+
+**Adam Stacoviak:** \[unintelligible 01:05:06.26\]
+
+**Lara Hogan:** Yeah, absolutely. Same vibe. And you need all of them. You need all of them, a supportive, powerful being to support you going forward. So I like this idea, as we go forward, who are the people in your network that can provide the different kinds of support that you need? Who can provide good mentorship, good coaching, good sponsoring, good feedback? A good eye for the politics at your organization. A good person to shadow at company meetings. A person with a completely different leadership style than you are.
+
+Actually, I made a little worksheet, a little bingo card for everybody to kind of brainstorm who is already in their Voltron for this variety of different skills, and where are the gaps, where might you want to find someone to fill that void for you as you grow and learn, that way you can collect a supportive group of people that can continue to level you up as you grow together.
+
+**Adam Stacoviak:** \[01:05:58.13\] Yeah. You know, listening to all this, it seems like there's really just such a wild framework to be a manager. I never really understood that it was like -- maybe I thought you were born with it; you were born with the ability to manage, the ability to lead... But it seems like it's such a learned skill. Like most things, you just sort of discover more the things that you can do well in life tend to be things you can really learn, not just something like "Oh, that person's good at it because they were born with the skills." It's probably somewhat true they've got some assurances and maybe a certain personality type. I tend to be very forward-thinking, I tend to be very self-assured; I have self-assurance, so when I score on like -- what's that, the Strength Finder I believe it was...? And I almost wonder if you have something like that, where you can sort of take a test, like "Am I fit to be a manager?"
+
+You mentioned the spectrum before, if you have anything like that. I'm curious, but -- you know, that's something that for me is, you know, this is something you can learn; you can go to a course like you might have, or read a book like you've written, or Camille's book, or other books that you can suggest... But this is something that you can study. There's theory behind it, there's framework behind it, and so just because you don't have those skills today doesn't mean you can't bury your head in a book, or take a course and learn this kind of thing, because it seems like there's a lot of do's and don'ts and framework to it that are waiting to be found, basically.
+
+**Lara Hogan:** Yeah, I would argue that anybody can be a manager. It's actually important that we have a diverse set of approaches to management and to leadership. It's important that there's not just one archetype. We of course all have an archetype of what a manager looks like, or what a leader looks like, but that's not gonna serve all populations, all organizations, all companies and all stages.
+
+I've always approached this from "Here's a set of skills, here's a set of tools that you can have in your toolbox that you can learn how to apply based on who you are, based on the organization you operate in, based on what your team needs." And that context is gonna keep on evolving. So you as a manager, you as a leader need to keep on evolving your toolbox going forward, because again, it's not one-size-fits-all, it's definitely not one side of like do's and don'ts, it's "Oh no, here's a new one I haven't seen before. How do I adapt? Who can I learn from to figure out how to approach this one in a way that works for my style, and also works for the environment?"
+
+**Adam Stacoviak:** Yeah. I mentioned before, I thought I would throw this at you - you weren't much of a fan, or you haven't read much of Seth Godin, is that right?
+
+**Lara Hogan:** I have not read much of Seth Godin, no.
+
+**Adam Stacoviak:** Are you familiar with the book Linchpin, at least?
+
+**Lara Hogan:** No.
+
+**Adam Stacoviak:** Not at all. Okay, good. I'll give you the paraphrase of the book, and I'm curious what your thoughts are on this... Linchpin is essentially that you are indispensable; that you have to be so good that they can't ignore you, essentially. So I sort of bucked at that. I really thought that's what I had to be as a leader. I was a product manager, I felt like I had to be the linchpin, I had to be indispensable, I couldn't be replaced... And then I thought "Well, that's just super-arrogant. That's super self-centered. Aren't we all truly indispensable? Aren't we all truly replaceable at some point?" And I thought -- you know, this question I wanna share with you or ask you is "Cog versus linchpin." Should someone aspire to be a linchpin, or should someone in the leadership aspire to be a cog? Because a cog sort of fits in and helps everybody else move; they're very helper, very servant potentially even, whereas a linchpin is like "I am the only one here that can do this", at least to some degree. So I'm curious what you think about cog versus linchpin.
+
+**Lara Hogan:** Based on your description, I think I could not be further diametrically opposed to the idea of linchpin, just because, you know, our job supporting other people as they grow - it needs to continue to evolve and change and we cannot be the only one. Personally, again, based on the short description. I cannot imagine a world in which having a bunch of linchpins is a healthy way for -- what's everybody else then? \[laughs\]
+
+And when I think cog, too - certainly, servant leadership has its pros and cons. There's definitely times when we need to be a servant leader, and other times we need to be more directive, maybe, than that... Again, pitfalls to every approach, which is why there's definitely not one one-size-fits-all approach to leadership.
+
+\[01:10:04.11\] But I'm all about understanding what the needs are, figuring out how you might be able to support them or understanding when you can't, and finding out "Where else can we get the support that I or my teammates or my organization needs to move forward?"
+
+**Adam Stacoviak:** I personally land on this idea that I need to be a cog. I really studied hard and I was like "You know what--" I read the book and I thought -- the book was actually very influential to me, and helped me gain confidence. But over time, I learned that that's not really what I need to be. And I'm gonna share something I wrote on the subject just to see if it resonates with you. And this is me sort of coming to this conclusion, essentially, so this is my words. "I'm a very sharp, highly-specific, purposely-purposeful cog that's part of a much bigger, much more grand machine. I play a very specific role, I highly need \[unintelligible 01:10:49.13\] so others can do the same. I serve the unit, the team, and its mission. Not myself."
+
+So as a leader, I felt like that's what I needed to be. I needed to serve that mission, and less of myself. Not so much completely not myself, but I need to think about my team's motives, my mission, the unit etc. rather than just simply saying "Adam, it's all about you."
+
+**Lara Hogan:** I love that. And many organizations need exactly that. They need someone who's gonna come in and understand the dynamics, understand what's needed, and support it in that way. Other organizations need some version of a linchpin at some moment in time, and I think that's the important part about understanding the context and also understanding who we are, so we can make sure that even if we can't fit the mold of what an organization needs, we can help make sure that that cog gets fit sometime.
+
+**Adam Stacoviak:** Yeah. Well, Lara, it was awesome talking to you. It's years in the making; like I'd mentioned at the top of the show, I'd found your post in 2019. Don't ask us why it took us so long to invite you... \[laughter\] Maybe we were -- I don't know. I don't know what it was.
+
+**Lara Hogan:** Now is the time.
+
+**Jerod Santo:** Now is the time.
+
+**Adam Stacoviak:** So we reached out via email, you said yes, here you are... It was awesome. Is there anything we didn't ask you? Anything we didn't bring up that you'd love to share on shows like this, when you do podcasts and you speak to audiences? Is there anything left unsaid, essentially?
+
+**Lara Hogan:** Honestly, I'm just so excited to talk about all of these things with you. I think I could go on for hours and hours. So much is left unsaid for another episode. \[laughs\]
+
+**Adam Stacoviak:** Yeah. Alright, Lara, well thank you so much for your time today. It's been awesome talking to you.
+
+**Lara Hogan:** Thank you both.
+
+**Adam Stacoviak:** We appreciate you.

Make sales not features (Interview)_transcript.txt

@@ -0,0 +1,579 @@
+**Jerod Santo:** Today I'm joined by Kendall Miller, a bubbly extrovert. So I have expectations, Kendall.
+
+**Kendall Miller:** Extrovert as a service, not just a bubbly extrovert. Don't act like I don't sell it, Jerod.
+
+**Jerod Santo:** Okay... How do you price such services?
+
+**Kendall Miller:** \[laughs\] Well, that's complicated, because what bubbly extrovert means at every company is a little different... But it's surprisingly valuable in a world of technology nerds to have - I don't want to say an ounce of personality. There's a lot of people with an ounce of personality...
+
+**Jerod Santo:** Sure.
+
+**Kendall Miller:** ...but to bring a truckload of personality is a lot. So yes.
+
+**Jerod Santo:** So I imagine you walk into a meeting and everything changes. There's like before Kendall and after Kendall. Is that fair?
+
+**Kendall Miller:** When people notice it is when I take vacation and I come back, and they're like "Oh, thank God, because at least somebody is going to talk before the meeting starts. Somebody's going to interrupt occasionally, and put in something bubbly and entertaining." Yeah.
+
+**Jerod Santo:** Yeah. As a service to others. It's for everybody else.
+
+**Kendall Miller:** It's really -- you know, I'm here to be a blessing to the world. We're surrounded by software engineers and product people, let's just be honest... \[laughs\] I love them, I don't want to Bash them. My entire world and my entire career has been in tech, basically. Well, most of my career has been a tech, but...
+
+**Jerod Santo:** Well, you wear many hats. One thing I read about you that I'm sure you wrote about yourself is that you help builders of technical products turn them into successful technical businesses. And so I thought we'd start there... What's the difference? How do you help? Why do we need help? etc.
+
+**Kendall Miller:** Well... Oh, man. So I work with a lot of early-stage founders, almost always technical founders. I'm not the technical founder. I have a technical background, but it's been 20 years since I've written meaningful code. I've never been paid to write code, and that's probably a positive thing. I find recursion much too interesting... Which as an aside, I do believe that recursion is our way out of the coming AI apocalypse. We've just got to get really creative with recursion and they'll just never be able to take over.
+
+**Jerod Santo:** They'll get stuck. Okay.
+
+**Kendall Miller:** Yeah. It seems really obvious to me.
+
+**Jerod Santo:** I'm with you.
+
+**Kendall Miller:** Anyway, so --
+
+**Jerod Santo:** I thought we just pull the plug, but that's my move... Is just pull the plug. But yours is way more diabolical.
+
+**Kendall Miller:** The plug is getting more and more complicated...
+
+**Jerod Santo:** This is true.
+
+**Kendall Miller:** We've created these power sources that are portable, unfortunately.
+
+**Jerod Santo:** This is true.
+
+**Kendall Miller:** Yeah. Anyway, so I'm often working with a technical founder, and technical founders have some particular quirks, in general, and most of them seem to be kind of common across the industry... And we can get into some specifics there, but the first and most important thing that a technical founder does is they almost always think they can build their way to a successful company. And that happens occasionally, but it is definitely the exception. Because the best products do not win. And I'm going to s\*\*t on a very large company here, but... Microsoft exists, and we know that the best products do not win, right? You win because you learn how to sell. Does having a good product help? Absolutely. But there are terrible products out there where the founder has figured out how to sell it, and they have wildly successful businesses. And there is an incredibly large graveyard full of the world's best products that nobody ever used, or nobody ever sold even. Sometimes they can get millions of users and they can't figure out how to sell it. So the difference between a technical product and a technical business? A technical product is a hobby, until you have people who are paying you for it.
+
+**Jerod Santo:** Right. And so you think sales or positioning is what it takes in order to take one to the other. Because I think you're right, but you're also crushing my soul. I just want the best thing to just win... \[laughs\]
+
+**Kendall Miller:** I mean, I could literally -- one of the companies that I'm working with right now, the founder built Nomad. And Nomad over Kubernetes - every company I've ever spoken to that used Nomad and used Kubernetes said Nomad is by far the better product. Like, night and day the better product. And I used to run a Kubernetes consulting shop, and helped, I think, the largest user of Nomad migrate to Kubernetes. And the reason they moved, they said "Nomad's better in every way. We hate that we have to do this. But when we get online and we google for "Has anybody ever run into XYZ edge case?", there's nobody out there who's ever seen it before. And we're the biggest company, running it at the biggest scale, and we're sick of that. We just want to go someplace that everybody knows."
+
+\[08:12\] And it's kind of the same reason everybody still uses WordPress. It's not that it's the best CMS, it's just that it has incredible market share. If you Google any problem, you're going to find a plugin that somebody has already written, that already solves your problem, because it's been around for forever and it's been wildly successful.
+
+So yeah, it's not just positioning though. You can have fantastic positioning. It doesn't mean you're talking to the right people, it doesn't mean you're pavement pounding... And a lot of founders are like, they'll build something, they'll put it out there, and they wait. "Oh, nobody found it. Well, I could do some advertising, write a few more blog posts... Or I bet if I built this other feature, then they're gonna love it." I mean, it's just so much easier to build. And I know this; I run a couple of companies myself, and I would always rather just keep building, because it's more fun than go sell. It's hard to sell, especially early on, because you're like "Hey, I know you've never heard of us and we have no credibility, and you don't know anyone that can vouch for us, and also, our product's really good. We swear." Right? So... It's hard.
+
+**Jerod Santo:** So how do you help these technical people? I mean, they hire you and you help them, or is there general advice that could apply to people that aren't necessarily hiring your services, that they could walk away with and say "Okay, here's --" I mean, obviously, getting out there, you said pounding the pavement... I don't know what else you said, but I'm sure there are other steps in there that I could write down and do.
+
+**Kendall Miller:** \[laughs\]
+
+**Jerod Santo:** Like, what are some actionable steps?
+
+**Kendall Miller:** I mean -- so the first thing is... You've heard it said first time founders build, second time founders sell. There's a lot of different variations on that, but it's a common thing. The selling -- what makes a salesperson a salesperson is not that they've learned this trick. "Oh, if I say this in this way, then they're going to buy." There are a few tricks like that. I've learned a few that I've been surprised actually are very effective... But it's not like I'm going to turn my close rate from 5% to 90% because I learned to be a sleazy dirtbag sales guy. The truth of sales is number one, it's a numbers game. You knock on enough doors with the world's worst vacuum, eventually you knock on a door where somebody's vacuum just broke, and they have companies showing up in a half an hour, and they are stressed the heck out and they don't care how bad your vacuum is, they need it right now. So at the end of the day, it is just a numbers game. Having a great vacuum that does a great demo helps you sell more vacuums. But you've still got to knock on a million doors.
+
+What makes a salesperson a salesperson is they've knocked on every door in the neighborhood, they have no idea what to do next, and they wake up and keep going. And that's not in the DNA of the average technical co-founder or technical founder. And so it's a thing that they have to learn the hard way, and almost always by the hard way, without putting in air quotes, what that means is they've failed once. They built a company and they couldn't get enough people to use it, or they reached a couple million and they needed to reach a lot of million, because they were venture-backed. And then they realize in the second round that they have to go bigger, harder.
+
+So what do I do? Well, so you asked a couple of questions there. First of all, general advice - sell, sell, sell, sell, sell. Sell it before you build it. Also, build something much, much smaller than you think. Because a lot of companies go build a product this big. Huge. And then they find out after talking to everybody that the product should go this way, and 10% of it is relevant. And it's really hard to pivot when you've boiled the ocean already.
+
+\[11:50\] If you build a product, that really should be a feature, and you could build in a weekend, then when you go out and talk to the market and find out "Oh, actually, the interest is over in this direction" or "I need to build it that direction", it's not that stressful to pivot, it's not that hard. You can listen to the advice from customers and actually build a thing that they want.
+
+My favorite companies are the companies that when I talk to technical people, they're like "Why does that even exist? I can build that in a weekend." Sure. You could build that in a weekend. You'd build it poorly, it would be not feature rich, and you wouldn't be able to sell it, because you have no design sense and you don't know who to talk to. Right? But if you build it in a weekend and then you add on, you can add on in the right direction without being stressed about it because you already boiled the ocean.
+
+So that's the first thing - build a lot smaller than you think, sell it a lot sooner than you think. If you can't get somebody to pony up money for it really early on, it's not going to be easier when it's bloated and big and heavier, and you've put 10,000 times more of your opinions into it.
+
+The other thing that I see a lot of technical founders do is they don't do the math. You have to build a business around something that actually makes sense. One technical co-founder I worked with a long time ago built a B2B SaaS product that engineering leaders were going to buy, and he priced it as such that - I was talking with him saying, "Well, now hang on a second... You've sold to one of the world's biggest companies, and it's a $5 a month deal?" "Well, more engineers are going to take it up." Okay, but how many of these do we have to sell before you can earn a living? And this is a bootstrapped business, it wasn't even venture backed... And he's sitting there and doing the math, and like "Well, I guess I've got to sell about 10,000 more..."
+
+So either you can sell it for $500 a month, or $1,000 a month, or $10,000 a month, or whatever it's going to take you to actually get to the sustainability that you need within a timeframe that makes sense, or you need to not sell it. But selling it for $5, patting yourself on the back that you got a success story, and finding out that you could do this for 25 years and you might be making $2,000 a month... That doesn't work.
+
+**Jerod Santo:** Right.
+
+**Kendall Miller:** So there's a lot of things that people miss.
+
+**Jerod Santo:** How do you pair these two ideas? Idea number one is build something small and focused, and arguably because of those two things not all that valuable... And charge $500, a thousand dollars. Like, charge a lot, build a little. It takes a lot of confidence to do something like that. Like "Well, I've got this little thing that does one thing. It's going to cost you a lot." It seems like one or the other. But are you saying you can accomplish both?
+
+**Kendall Miller:** So, first of all, you can get people to buy something that you haven't built. And there's a difference in the conversation between "Is this interesting to you?" and "Would you pay for it?" There's a lot of people in tech who will say "Oh, that's super-interesting. I love it." But that's different from "I would happily pay for that." So there's a little bit of nuance there. But how do you jive those two things? That's the hard part about picking a niche. I mean, there's times where you need to go boil an ocean, and build, say, a cloud, before you're going to have adopters. But it's not like Amazon started with a cloud. It started with compute, and then found out all the other things.
+
+So when you're going to go compete with the world or build something new and interesting in the world -- I mean, there's a reason they call it an MVP. Viable is not "The smallest I'm capable of building" in minimum viable product. Viable is the smallest thing that I can get somebody to buy, is really how you should think about that. And buy at a reasonable price. But if somebody says "Hey, I really need XYZ problem solved", you say "Well, what's the first bit of that problem that I could solve for you today, and you'd sign on a dotted line because it would be valuable enough to move? And can I go build that in a weekend?" And a weekend is a little handwavy. Maybe it takes you a week. But if it takes you nine months, the chances are more likely than not you built the wrong thing.
+
+**Jerod Santo:** \[16:02\] Mm-hm. It seems to me obvious that there are unicorns out there, but most technical people do better if they're paired with somebody salesy, designy, extroverted, perhaps... Someone like yourself. Like, perhaps you and I could team up and build a good business, because I have builder chops and you have sales chops, and we jive etc. That combination is really hard to find. It might be easier to find a spouse at this point.
+
+**Kendall Miller:** \[laughs\]
+
+**Jerod Santo:** So -- I mean, seriously. So many people... Is there a Tinder for co-founders? Probably there is.
+
+**Kendall Miller:** There is, yeah.
+
+**Jerod Santo:** Is there?
+
+**Kendall Miller:** Yeah. There's a couple of systems out there for --
+
+**Jerod Santo:** I was like "Where do we meet?"
+
+**Kendall Miller:** I mean, even YCombinator has like a co-founder matching setup, I believe. And there's a couple of different places that do something similar to that. But I think the hard part is finding the bubbly extrovert who's technical enough to understand the product at a deep enough level to talk coherently about it. Now, for a lot of those technical co-founders, they never really understand it at a deep intellectual technical ability. They understand it in the same way that you learn a foreign language. If you're around it long enough, you can kind of sound like the other person, and you're just kind of BSing. But you start to sound like a native if you've spoken it long enough.
+
+**Jerod Santo:** Right.
+
+**Kendall Miller:** So it is difficult, but what I think is interesting is that there are VCs that back an awful lot of technical founders without that business person, and just assuming that they'll figure it out. This morning -- I have an old sales leader that I worked with (I'm trying to be vague here), who is working with a technical founder who the VCs told him "We will back you if you hire a sales background co-founder, because we don't believe that you can make this go on your own." So that is a thing that does happen.
+
+**Jerod Santo:** Yeah. And there's currently a lot of business people, or as I sometimes call them idea guys, who are drooling right now over the prospects of vibe-coding their way without a technical person to sales success.
+
+**Kendall Miller:** Sure.
+
+**Jerod Santo:** And I think both those routes are probably fraught.
+
+**Kendall Miller:** \[laughs\] To say the least. Yes, Jerod, they are both fraught.
+
+**Jerod Santo:** So how did you as a non-technical person -- or I don't know, what's your background? Were you a computer science major? Probably not. How'd you get in and then how'd you eventually learn the language?
+
+**Kendall Miller:** Well, so it is complicated. So my distant background is technical. I started writing code in fifth grade, and my first programming language was Awk, which - don't start with Awk, would be my encouragement.
+
+**Jerod Santo:** Nice.
+
+**Kendall Miller:** I went from Awk to BASIC...
+
+**Jerod Santo:** Don't finish with Awk either, yeah.
+
+**Kendall Miller:** Don't finish with Awk either. And I worked up, but I spent the most of my time with C++. I entered college as a sophomore, almost entirely with computer science credits that I had taken in college classes and APs in high school... And with the assumption that I was going to study computer science. But I made a decision late in high school "I think I like people better, and computers completely consume me, and so I will fail at choosing between the two." So I do have the tech technical background. I worked for a startup in 2004 in Denver that built Twitter about five years too early... And then I left the field completely, I got an English literature degree, two masters in the history of religion, I moved to China and worked for a nonprofit for 10 years... So a wildly different background happened for a while, but when I moved back to the States about a little over a decade ago, I was the first hire at a small startup... Not realizing I was the first hire, because I didn't know to ask, "Do you have money, and do you have people that work for you?" I know now, but anyways.
+
+**Jerod Santo:** What did you ask them? Like "Are you hiring?" "Yes." "Okay."
+
+**Kendall Miller:** Yeah. He offered me more money than anybody else offered, and I took the job...
+
+**Jerod Santo:** \[20:07\] There you go.
+
+**Kendall Miller:** Because nobody knew what to do with me. They're looking at my background in China going "What the heck is this?" So I did sales for them, we succeeded at sales, I hired a sales team, I hired a marketing team, I took over operations... I actually ended up taking over engineering when the CTO made an engineer cry... And then eventually took over the company and we sold the company. So it's that experience where I did literally everything at the company over the course of about eight years. That put me in a position where I could go out -- then I started meeting with other technical founders, I ended up on the board of another company in Denver, that has grown, been successful, had a big exit to PE about a year and a half ago... Ended up just advising a number of different places. And I say advising, it sounds like they're paying me in equity or cash money, and most of the time it was just me helping somebody out. "Hey, I've been there before. Let's talk about this." "Hey, I've got somebody to refer your way." "Hey, talk to this person, talk to that person."
+
+**Jerod Santo:** Pure altruism, or what?
+
+**Kendall Miller:** So is it pure altruism? I mean, it makes it sound like it's ever that clear of a thought... No, I think networking has network effects, right? When I moved back to the States from China and logged into Slack at this new company and found out that I was the only employee, I panicked.
+
+**Jerod Santo:** \[laughs\] "Where is everyone...?"
+
+**Kendall Miller:** And so I was like "Well, I'd better go network, because I'm going to be out of a job in no time." And so I went out and just networked like crazy. Who can I meet? I'm going to be genuinely interested in anyone I can find. And finding out that that led to hires... You know, the first sales hire that I brought in was that, the first engineer that I brought in was that... That led to all kinds of hires, that led to sales, that led to eventual business opportunity... I mean, a guy I hired nine years ago is the guy that I'm in business now with, that -- I have an opportunity to talk about this. I make a liquor called Friday Deployment Spirits. This is our Generative A Rye.
+
+**Jerod Santo:** Oh, my goodness...
+
+**Kendall Miller:** It's a rye whiskey. We also make a gin called Force Push Gin, focused on the tech world. Anyways, it's a guy that I hired eight years ago, because I met him at a meetup, I needed a person who did exactly what he did... We worked together great for a little while, and then years later we ended up starting a business together. So networking has network effects. So when I'm out meeting people and I'm genuinely interested in getting to know them - which I am, because I find people fascinating - I find, "Hey, you need a job and I know somebody who's hiring for you. And if I refer you to that person", and I don't take... A lot of people have asked me, because I refer a lot of people for jobs. So a lot of people have asked me, "Why don't you take a recruiter fee?" I would probably make a lot more money if I did. But as soon as I monetize the network directly, the network loses all of its value.
+
+So because I've referred a person over here, that director of engineering now owes me a favor. And so now I take a job at this new company, and I think that this product might be relevant to that director of engineering. And I don't mean owes me like -- you know, it's not quid pro quo. But when I call him and say "Hey, would you take a look at this?", he's going to say yes, because I helped him hire the last five people he's hired.
+
+Actually -- I mean, a good friend of mine in San Francisco, who got fired from a very big company... I don't actually think he got fired. I think he left. I don't know. I don't know what happened. Very large company. And he's one of these guys that's too big of a big cheese. Like VP at very large organization. And nobody in my network is the CEO of Microsoft, right? Like, I know a lot of people, but I don't know those people. So I only know one very large cloud company's execs, and I wrote to them and said "Hey, this person came available on the market. He's more expensive than almost any network that I know, so you're the only place I can refer them to." Well, they just hired him, I found out yesterday.
+
+\[24:08\] So now there's a C-level executive at one of the largest cloud companies in the world that doesn't -- I'm not saying owes me one, but will definitely pick up the phone when I call. And so those kinds of things just add up, and so what ends up happening is somebody pings me, has a question about "Hey, I'm building XYZ. Here's why it works. Here's why it doesn't. Can I get some feedback?" I look at it... 70% to 80% of the time I just s\*\*t all over their ideas. Sorry, I'm not supposed to curse here, because you're going to bleep them out. I say bad things all about -- you know, I shut them down. I tell them it's a terrible product.
+
+**Jerod Santo:** \[unintelligible 00:24:41.01\] Be yourself.
+
+**Kendall Miller:** Well, myself is -- I worked for a religious organization in China for 10 years, Jerod, and one of the things I decided when I moved back was that I was going to curse more, and I've been wildly more successful than I anticipated.
+
+**Jerod Santo:** Okay.
+
+**Kendall Miller:** But yeah, I mean, I ended up \*bleep\* all over their ideas... But about 30% of the time or 20% of the time I'm like "Wow, this is actually really interesting, and here's three or four companies I think might need that right now." And I go make those intros, those turn into meaningful conversations for those companies, and that company comes back and says "Can we hire you? You did that really fast and really easy." And my goal is not to make sales. I don't want to be a sales guy. But I can go in and get those first few product feedback conversations. I can help you build a sales team. I can help you build a marketing team. I can help you understand what you're doing wrong.
+
+But even then -- I mean, honestly, a huge percentage of what these guys are doing wrong is they don't know that you've got to show up on time to an interview. It's the little things. You've got to communicate.
+
+**Jerod Santo:** Basics.
+
+**Kendall Miller:** Yeah. Because you've been an engineer in a hole, where the thing that was valued was that you outputted code, and you were really good at that. And that's different than what it takes to run a company and to get people rah-rah-ed behind you and get people excited about the vision for where you're going, and give people the freedom to go build on their own. And I'm talking for a long time here, as I tend to do...
+
+**Jerod Santo:** No, that's fair. I'm having lots of thoughts. One of them was - have you seen the episode of Friends where Joey tells Phoebe that she cannot commit a selfless deed? And so the entire episode she goes about trying to have a selfless act...
+
+**Kendall Miller:** Yes...
+
+**Jerod Santo:** Because I asked about "Is it pure altruism?", which is a stupid question, because there's no such thing as pure altruism.
+
+**Kendall Miller:** Right.
+
+**Jerod Santo:** Because it's like, no, but you try to do a selfless act, and somehow it comes back to you anyways.
+
+**Kendall Miller:** Yes.
+
+**Jerod Santo:** So I was thinking of that.
+
+**Kendall Miller:** Isn't it where she ends up carrying a baby on behalf of somebody, but realizes she's going to feel good about it?
+
+**Jerod Santo:** Yeah. I mean, she felt good at the end of the day... It's like "No...!" and she was very angry. I loved that particular setup.
+
+**Kendall Miller:** Yeah...
+
+**Jerod Santo:** The other one is you say you don't want to be a sales guy, but I'm trying to figure out what kind of guy you want to be. Do you like zero to one and then move on? Are you just trying to help out where you can? What part of it excites you?
+
+**Kendall Miller:** So pseudo sales is always going to be a part of my job, right? I'm always going to be talking to people. It's always going to lead to sales conversations. There's a lot of times I end up at a new company and somebody in my network pings me and says "Hey, that's kind of interesting. Can I talk to you about it?" and it turns into a sales conversation. I don't like being the sales guy, because then I have to hassle my network, check in, follow up... Usually, I want to hire somebody else that's going to do those kinds of things, so it doesn't have to be Kendall hassling somebody. I don't want to hassle. I don't mind if a salesperson who works with me or for me is hassling the hell out of one of my friends; that's their job. But it doesn't, if it's affecting my relationship with that person, that's a problem, if that makes sense.
+
+So what do I want to do? I mean, that's the messiest part. That's the hardest part for me every time, Jerod, is because what excites me is just solving business problems. I like getting in, finding out that there's things that are a mess, and just going and helping solve it. To give you an example, in the last two days I've written job descriptions, I've come up with prospecting lists, I've interviewed people for marketing roles, for operating roles, for sales roles... I'm interviewing for a CEO right now...
+
+\[28:10\] I touch every single part of the business, because I'm comfortable in almost all of it. I'm not going to go review a PR. That's the one spot that I'm probably not going to be useful... Although at a previous company that I was running I would about every year go issue a PR, just to scare all the engineers... You know, there's always something small, like a typo or something in documentation...
+
+**Jerod Santo:** So you have your spirits company, but I'm thinking why not apply your talents at something that you're building for yourself, or with somebody, versus helping other people? One of the things you do is building this spirits company?
+
+**Kendall Miller:** Yeah, so there's a bunch of things I do. So what I get paid for directly - I work with a cloud company called Civo out of the UK. I work with an AI company called Tensorlake, out of San Francisco. And then I'm on the board of a company called FusionAuth. Think an Auth0 competitor. Nut then I run a couple of things myself. So there's the spirits company I have a business partner with, we make a gin and a whiskey, and we're looking at some other interesting things that we may do next... I had a really crazy opportunity come up yesterday that it's too early to talk about, but it's exciting. I run a network of CTOs globally called CTO Lunches. So we have about 1600, 1700 CTOs worldwide. We put on lunches all over the world. That's nice, because that is a way that builds the network... Honestly, what I like about it is I get to fly places on the company dime, eat lunch on the company dime, and now I'm in the liquor business, so liquor's a company expense... And there's all these different things. But CTO Lunches is another business that I own and run -- I co-own. But I do think that CTO lunches has the capacity to get to -- if everybody fired me tomorrow, I think within six months I could make enough money off of it to make it my full-time gig. But I like touching other things enough that I'm not walking away.
+
+**Jerod Santo:** Yeah. So how does that one make money?
+
+**Kendall Miller:** We'll come back to that. There's one last thing that I do also, is I do have an advisory group called Grow Big Advisors, where I've paired together with a handful of friends to do this sort of startup advising. They have slightly different skillsets than me. One's a sales leader, one's an engineering leader, one's got the marketing background. So it's different things, but in exchange for a small bit of equity, we provide this kind of just advisory, rather than hands-on. So that's a little different. So those are the things that I'm building for me.
+
+**Jerod Santo:** So you have a lot of stuff going on, and you're not just helping other people with their businesses.
+
+**Kendall Miller:** I stick my fingers in a lot of pies, yes.
+
+**Break**: \[30:52\]
+
+**Jerod Santo:** So why the spirits pie, and...?
+
+**Kendall Miller:** \[laughs\] So we were -- I had just come back into town after some time away, and I was meeting with four old friends, and we were sitting around at a brewery, talking about "Why can't I buy a bottle of whiskey called "I did SOC2 and all I got was this bottle of whiskey"? So the list of names was immediately very long. And Rishi, my business partner and I stuck around after the other two guys left, and Rishi goes "Well, Kendall, I, I think this is maybe a good idea." And I said, "Well, Rishi, I think this is a good idea." And he goes, "Well, I'm not messing around." I said, "Well, I'm not messing around." He said, "Well, I kind of think we should do it." I said, "Well, I kind of think we should do it." And the next day I was on the phone with a distillery saying "Hey, how do I buy a whole bunch of liquor if I want to do this?" And we ended up in a partnership, and it's worked out.
+
+**Jerod Santo:** So I've noticed this as a trend, which is why I think it's a) probably a good business, and b) a relatively easy one to get into, is every celebrity has their tequila, or their whiskey, or their gin named after them, or whatever it is. It's like branded alcohol for this celebrity, and --
+
+**Kendall Miller:** Aviation Gin. I'm Ryan Reynolds. Thank you, Jerod. No, I didn't want to make the connection, but since you did... We can also talk about my football team in a little bit, but...
+
+**Jerod Santo:** There you go.
+
+**Kendall Miller:** Anyways. Sorry, keep going.
+
+**Jerod Santo:** No, so I just -- I've noticed that and I'm like "Well, this must be lucrative, and relatively easy", because you're basically just branding, aren't you? You're not opening a distillery. And you just said you went and bought a bunch of whiskey, right?
+
+**Kendall Miller:** Yeah. Well, so the whiskey and the gin are made for us, it's not white-labeled something else. So a lot of places will go -- man, I don't know if you want to get me talking about liquor, but we can talk about liquor.
+
+**Jerod Santo:** How deep you can go on this...?
+
+**Kendall Miller:** So the vast majority of distilleries are buying a base liquor from Indiana, because that's where a huge percentage of the distilleries are, for \#reasons. The alcohol laws in America will blow your mind, that are left over from prohibition. So anyways, our distillery makes it from grain, so it's not a base spirit that's just distilled again nine more times, like Tito's does and then brags about it. "Nine times distilled..." How bad was what you started with that? You had to distill it nine times...? And it's distilled once, because it's a fantastic product. Anyways, and it's made for us. It's a recipe for us. So it's different, it's contract distilling, rather than just white-labeling something else. But yes, I did not open a distillery. I do not want to open a distillery.
+
+**Jerod Santo:** That's the hard part.
+
+**Kendall Miller:** I just handle it once it's in the bottle. Well, I have to buy the bottle, and distribution's a gigantic pain in the butt because of American liquor laws.
+
+**Jerod Santo:** So I'm thinking then -- I mean, of course, I'm target market, because I'm a nerd, and Generative A Rye lands perfectly on me... But there's not very many -- like, how many...?
+
+**Kendall Miller:** It's also really good whiskey, Jerod. Sorry, keep going. \[laughs\]
+
+**Jerod Santo:** Well, that's fine. That's all well and good. But the naming is what I care about. You're targeting alcohol at technical people, basically. Isn't that like a small cross-section of the world? Are you limited in your --
+
+**Kendall Miller:** Well, that's why one bottle is $1,400.
+
+**Jerod Santo:** Oh, okay.
+
+**Kendall Miller:** No, I'm kidding. It's --
+
+**Jerod Santo:** You just sell it to your CTO Lunches. That's your entire market.
+
+**Kendall Miller:** The whiskey is $125, and the gin is $95. So it's not cheap. It is a luxury item.
+
+**Jerod Santo:** \[36:10\] That is expensive.
+
+**Kendall Miller:** It is a very, very good item. So let me be clear about a couple of things. One, I would have totally done this exact same thing with a terrible distillery, and a terrible end product, because I think the marketing has legs. We have a fantastic product, that I can't believe how good the product is. I actually think the gin is the best gin I've ever had, and I am a gin snob. And I don't feel like I can say that about my own gin, but it is so very good. And I can say it because I didn't make it. I had some input into the recipe, but that's all. And the whiskey is also very, very good. So it is a luxury product, and it costs a lot of money, and I'm selling it to a group of people who have a lot of disposable income, like to make jokes like this... My thought is not that Jerod will become a daily drinker of Generative A Rye; my thought is "Sometime in the next year Jerod is going to want to give a gift to somebody he knows in tech, and he's going to ship a bottle of Generative A Rye." Or director of engineering at XYZ company is going to buy his team all a bottle of whiskey for Christmas, and they're going to buy Generative A Rye. My goal is that.
+
+I don't think anybody's going to buy a $95 bottle of gin, even though it's a very pretty bottle of gin, and it says Force Push right there on the label, and it's -- it's Friday Deployment Spirits, Jerod. It's funny.
+
+**Jerod Santo:** It is funny.
+
+**Kendall Miller:** I still don't think it's going to be anybody's daily drinker.
+
+**Jerod Santo:** Gotcha. Well, I just wanted to tee up your sales pitch and see how you'd do this.
+
+**Kendall Miller:** How did I do? Can you give me a rating?
+
+**Jerod Santo:** I would drink it. I mean, I don't know if I would -- $1,400 bucks is... That's a call order.
+
+**Kendall Miller:** \[laughs\]
+
+**Jerod Santo:** I'm no CTO of anything though.
+
+**Kendall Miller:** It's not 1,400 bucks, I want to say again. 125 bucks versus 95 bucks.
+
+**Jerod Santo:** Oh, I'm sorry, why'd you say it was 1,400?
+
+**Kendall Miller:** No, I started at 1,400 bucks to screw with you, so that the 125 bucks that it actually costs would stress you out less.
+
+**Jerod Santo:** Oh man, you even price-anchored me.
+
+**Kendall Miller:** See, talk about sales tactic. Yeah, 100%.
+
+**Jerod Santo:** So sell me some of your other -- I won't say use car salesman tactics, but just... What are the winners where you can just say something and it'll work? You said you've got a couple of those.
+
+**Kendall Miller:** Yeah, so a couple of things that I think are really interesting... If you're selling to technical people, don't capitalize their name when you send the email.
+
+**Jerod Santo:** Seriously?
+
+**Kendall Miller:** Make some spelling mistakes. Yeah, otherwise it looks like it's a mail merge.
+
+**Jerod Santo:** I see.
+
+**Kendall Miller:** If you can just do the littlest thing to make it look like this was actually written by a human... And if you can write it by a human, it's even better. But you misspell somebody's name - do you know how much more likely they are to respond to you? Because they're like "Oh, this, this asshat didn't just plug it into a CRM and send me a mail merge. They're actually reaching out to me."
+
+**Jerod Santo:** Don't they also think that you don't pay attention to the details?
+
+**Kendall Miller:** 100%. And sometimes they respond with like a "Dude, come on. You could do better", but they responded.
+
+**Jerod Santo:** Okay. Fair.
+
+**Kendall Miller:** So first of all, there's little things like that, even if it's just lowercase the name, make a spelling mistake in your message so it doesn't look like it was sent by a machine. Nobody wants the machine outreach. That's number one.
+
+**Jerod Santo:** Alright.
+
+**Kendall Miller:** Number two, one of the most interesting things is every company needs to know -- where do I start? Founders always underprice their product. I do this myself. You weren't willing to pay 1,400 bucks, but 125 seemed cheap to you once I got back to it. I probably could have sold the bottle for 200 bucks and Jerod might've bought it. I don't know, I'm probably handwavy. But founders always underprice their product. And one of the most important things for fixing that is finding out what are people's budgets. So how much am I saving you - those kinds of conversations, but they're really hard to have. So there's this tactic called the bucketing method. And so what I say is "Jerod, I've got a nice bottle of whiskey here I want to put you into. It's a 2025, you're going to like the way that it drives and feels."
+
+**Jerod Santo:** Hm. That was a good year.
+
+**Kendall Miller:** \[40:07\] It was good year... And how much have you got to spend on this, Jerod? Are you hoping to spend $5 on a bottle of whiskey, are you hoping to spend a hundred dollars on a bottle of whiskey, or do you have $5,000 to spend on a bottle of whiskey? And there's something about that last number being so big it's outrageous, that almost everyone, almost every time, will go "Well, it's not $5,000. I was thinking more like 50." Okay, great. Now I've got a ballpark for what was in your head. So this method - you name a number, you name a number twice as big, 10 times as big... That last number has to be so outrageous that it makes them get mad. And then they'll almost always come back with what was in their head.
+
+**Jerod Santo:** Interesting.
+
+**Kendall Miller:** So that's called the bucketing method. That's an interesting thing. And then there's one last one, which is just in general in sales... If you can ask a question like "If this existed, and it was perfect for you to meet your needs, something like that, what would it look like?" And now, you're going to sell the product to me. And if I can sit still and shut up and let you just sell to me, the less I say in a sales call, the better you're going to feel about what I'm selling. Even though I haven't told you anything about it.
+
+**Jerod Santo:** Hm. How do you mean I'm going to sell it to you? Because I'm going to describe what it would be, and then you're going to say "That's what it is"? Or what do you mean?
+
+**Kendall Miller:** Yeah, I mean, it depends on what -- you can't, say "Oh, well, that's exactly what it is", but you can go back with... How do I say this? Here's a concrete example that's be relevant to everyone in your audience, even if they're not founders. Interviewing. When you're interviewing, you're trying to sell yourself. Right?
+
+**Jerod Santo:** Yeah.
+
+**Kendall Miller:** The worst possible thing you can do is go into an interview and say "Tell me what you need", and they list 50 criteria. And then you try to go through those 50 criteria and say "Here's why I meet every single one of those criteria." You will fail, because you don't. You're never exactly what they have in mind. Never ever, ever.
+
+**Jerod Santo:** Sure.
+
+**Kendall Miller:** So if you can say "Hey, in your perfect world, what are the problems that I would solve?" Not "Who exactly am I?" "I'm going to show up on day one. I'm going to wear this color. I'm going to talk to this person and do this thing", but "What is the problem I'm going to solve?" Then they start selling to you the problem that you're going to solve for them. And at the end of the conversation you say "Well, I might not be exactly what you have in mind, but here's how I'm going to solve your problem anyways." And keep it nice and short and sweet, and you convince them that what you've just said meets their criteria. Now they have just sold you to themselves. And that sounds super-sketchy; it's really not that complicated. There's one book on this that I encourage people to read. It's called Spin Selling. And spin is not like "I'm going to spin you into--", it's an acronym for something. But the one useful thing they talk about is this bit, of get the other person to be pitching you instead.
+
+**Jerod Santo:** Yeah. Interesting.
+
+**Kendall Miller:** It's a small thing, and it's easy to lose track of, but... I mean, literally, interviews is like the one place where particularly technical people get caught up. Like "Yes, I'm exactly that. I've written Python for 17 years, exactly like you asked, and I've known Kubernetes since Wozniak was born. Please hire me."
+
+**Jerod Santo:** Well, you mentioned Kubernetes. I know you've dabbled in the cloud, or you're kind of hanging out in like the infrastructure section of technology... Is that really where you -- besides the alcohol section, what other aisles do you hang out in?
+
+**Kendall Miller:** \[laughs\]
+
+**Jerod Santo:** Are you in the infra aisle, in the cloud? Where do you hang?
+
+**Kendall Miller:** So yeah, I'm currently working with this company Civo, which is interesting because Civo is a cloud company... I mean, think an AWS or at Google; they're Kubernetes-first. What makes them different is a couple of things. One, it's UK-based, which right now happens to be really good for business, for \#IDontKnow reasons... But the thing that's also interesting is they will ship you the whole software stack. So it'd be like if you used AWS for a few weeks and you're like "Hey guys, this is great. Can we have it?" And then AWS just shipped you AWS. So Civo will do that. You can buy the software stack.
+
+\[44:20\] They'll also ship you a hardware appliance where you can throw it in a rack. So what's interesting to me about that is the 10 years that I was in China, I basically missed the data center world... Because you know, in 2004, at that startup I mentioned, that built Twitter a little too early, we had some server rooms in the back that we would shut down a machine at a time and install the newest version, and then shut down the next one, and install the newer version... Or we'd put up a website that said "Down for scheduled maintenance." Remember those?
+
+**Jerod Santo:** Mm-hm.
+
+**Kendall Miller:** And then when I came back, it was cloud, and blue/green deployments. And I missed the whole data center world. So Civo is kind of interesting to me because it's my first time touching anything data center-related. Not that the customers necessarily are all doing that. Most of them are probably just using our cloud. But I find the data center bit really interesting, especially as people are starting to flee the cloud.
+
+So that's one bit. It does tend to be infrastructure. About a year and a half ago, almost two years ago, I went and raised money to start an AI startup. So I did plan to do an AI startup... My co-founder -- the VC said "We're ready to write the check" and my co-founder called me and said "So I think my wife's going to divorce me if I do this." So we shut the whole thing down.
+
+**Jerod Santo:** Really?
+
+**Kendall Miller:** Still good friends. But that was in infrastructure, because I do think there's some interesting infrastructure problems around AI in particular that are still unsolved. The reason infrastructure is interesting is because, you know, a single SaaS product with single SaaS vertical - you either hit the market perfectly or you don't. With anything infrastructure, you get to ride the waves, because what you're selling is the pickaxes. And it's not even pickaxes, it's blue jeans. So it doesn't really matter if it's gold rush, or we're cutting down a bunch of trees and clearing the forest. Everybody needs blue jeans.
+
+So that's the fun thing about infrastructure, is you're solving the problems that everybody needs to go solve their problem. It's an easier bet. So I find that more interesting.
+
+**Jerod Santo:** Right. And there really aren't workarounds. I mean, you're going to need it or you're not going to need anything. Infrastructure is going to be there, whether it's rented, bought, cloud, on-prem... It has to be there in some form, and so like you said, it is kind of more like the blue jeans, where there's not going to be really a -- there might be a down market, but never a no market.
+
+**Kendall Miller:** Well, okay, and then the thing that's relevant here that I think is interesting is part of what was fun about the Kubernetes boom 10 years ago when I was first involved in it - it's like there's this new kind of workload, a container. And it's not really that different, but it turns out the way that we orchestrate a container is slightly different. And the tools that ended up winning in that orchestration were slightly different. And there was a whole new paradigm in declarative infrastructure, rather than having to orchestrate everything like we did in the days of Puppet or Ansible... It was kind of exciting, because it was a new kind of workload, and all the paradigms were going to change.
+
+And KubeCon in like 2015, 2016, 2017, whatever... 2015 might've been too early. Maybe I started in 2016. But early on, it was a whole bunch of nerds sitting around in a room feeling like we're changing the internet, and this is fun. Right? And I feel like that's the only place that that's also happening, is in AI. And AI is -- you know, inference servers, the way that we train them, the way that we run them, the way that we audit them, the way security looks in AI... It's a new kind of workload, that requires a different kind of interaction. It's not that different; it's just a workload at the end of the day, and you box it up in some kind of wrapper at the end of the day... But there's interesting problems, because the things that go in it don't come out in the same way. And the things that come out don't come from the same places. So auditing that, lifecycling that etc. is new, it's interesting, and the people who are working hard to solve those problems I think are having a lot of fun.
+
+**Jerod Santo:** \[48:17\] Yeah. It's definitely different enough that I think there's -- you know, there's an MLOps community, which is basically was DevOps, or whatever you want to call it, prior... Like, whatever operations looked like.
+
+**Kendall Miller:** Well, which was platform engineering prior, which was Linux sysadmin before, which was...
+
+**Jerod Santo:** Yeah, this is one of the things... There's lots of titles in the cloud.
+
+**Kendall Miller:** If you go back far enough, it was a lot of guys writing a lot of Awk, just saying.
+
+**Jerod Santo:** Oh, good point.
+
+**Kendall Miller:** Yeah. Anyways, keep going.
+
+**Jerod Santo:** And they still might be the ones -- the old graybeards still using their Awk where it makes sense.
+
+**Kendall Miller:** Still using their Awk, yeah.
+
+**Jerod Santo:** Well, I was just going to say, there are differences. You mentioned security. Can you enumerate perhaps what makes it different? Why would there be a subculture that comes out of this? It's just compute and data, right? Like, what makes it different enough?
+
+**Kendall Miller:** Yeah, it should just be compute and data. Okay, so a concrete example - a friend of mine... Just this morning in the CTO Lunches community, a friend of mine works at a very large, very regulated industry, and he's a VP at this big company. And he's arguing that their company can't use any AI products, and it's super-frustrating. I think just today Shopify tweeted something like "It should be a reflex to use AI. And that is the absolute standard for all of our engineers at this point." So this friend who's a VP is like "Well, how do I get to that when I'm in a company where I can't use any of it?"
+
+So there are some solutions that have come out that address it from his perspective. I'm going to start there and I'm going to come back to the infrastructure bit. But from his perspective, the things are "Who's putting what, where?" And when you have a company like his, that's big and super-well-regulated, and everything's blocked, you know what people do? They pull out their phone and they type something into ChatGPT. So now they're putting data that they shouldn't be putting into the big public models that are probably -- they've got the free version, so the big public models are training off them, so that's the worst possible outcome.
+
+So there's a company, an old coworker of mine's, one of the early engineers, that's called SurePath AI. I don't know if you've heard of this, but SurePath AI sits in between almost like a proxy for a big company like this, so that every single thing that goes to, say, ChatGPT, they can stop and say "Oh, that has PII in it." And they redact it, so ChatGPT doesn't even get it. So when you say "Hey, ChatGPT, who is Jerod from the Changelog?", it's going to say "Hey, you said, \[unintelligible 00:50:48.04\] name from the Changelog. Did you mean to give me a name? It seems more like a placeholder." Like, that's going to be the actual response from ChatGPT, because they're going to intercept that.
+
+**Jerod Santo:** They're pulling it out.
+
+**Kendall Miller:** Yeah. So there's those kinds of cases. Now, there gets to be more interesting edge cases there, like, well, I want HR to be able to use AI and train on enough of our data internally that they can ask the AI "Give me a list of all the salespeople" and it can do it, because that's going to increase their productivity. But I don't need finance to have access to that, and I definitely don't want HR to have access to what finance has access to. So there ends up being fine-grained security there, even if you're self-hosting those models. So that's one thing.
+
+The next thing is when you're talking about even just self-hosting a model without a product like SurePath - and this is not an endorsement for SurePath. The guy that works there is a good friend of mine, I think very highly of him, and I hope they're making a kick-ass product. I've never used it, so I don't know, and I can't vouch for it. But the folks that are building something internal have to look at it, and how do you audit what the model's been trained on? How do you audit what the model's learned? How do you audit who's had access to that model? How do you audit if a malicious actor got in and started asking questions to that model? It's not that different from just looking at audit logs of a traditional server of any kind... But there ends up being some unique challenges there, and especially when you lifecycle those servers, those inference servers, and you have a new model on it, or you put in some new training data. "Who had access to the new training data? When did that training data go in? Now what did they have access to?" The problems actually get pretty interesting. And the rabbit hole goes deeper and deeper and deeper and deeper and deeper.
+
+**Jerod Santo:** \[52:39\] It seems like they snowball, and kind of fractal off even, to now you're multiplying problems against each other.
+
+**Kendall Miller:** Well, and when you're using AI to detect what you shouldn't be doing, or you should -- and that's the other thing that AI is really good at, is AI can watch you type a name in and go "Oh, that's a name." "Oh, you shouldn't' be saying that", so where does that model live? Who's hosting that one? Who's auditing that one? Yeah. I mean, it's --
+
+**Jerod Santo:** Who watches the watchers...
+
+**Kendall Miller:** This is a whole new world, and the people who are missing out on the AI gold rush - because a lot of us are Luddites over in the corners, going "Oh, this is one step farther than I want to go. OpenStack was enough." "We tempted fate when we trained the computers."
+
+**Jerod Santo:** All I need is a command prompt and Awk. That's all I need.
+
+**Kendall Miller:** See, you get me, Jerod. Now, I need to be better at RegEx if Awk is going to be all I'm going to do, because there's a -- anyways.
+
+**Jerod Santo:** So say we all.
+
+**Kendall Miller:** So say we all. But this is where all the fun is, and there is a whole bunch of things changing. It's also the only companies with money right now.
+
+**Jerod Santo:** Right. Yeah, I was expecting you would say that eventually. It's like "Well, this is where the money is, so why not be involved in the exciting thing where the money is?" It just kind of makes sense.
+
+**Kendall Miller:** Well, I mean, Civo has an AI offering that is actually pretty compelling. It's a private AI offering so that you can be comfortable with your data. And then this other company in San Francisco that I'm working with, Tensorlake, that does understanding of unstructured documents... It's actually really fascinating for me. I'm learning a lot. I'm excited to be in the AI world, because that is where things are moving... And I'm surprised how much I'm learning about it. I end up having conversations with ChatGPT every day. Yesterday I was asking a million questions about MCP servers, or MCP -- it's a protocol, or whatever.
+
+**Jerod Santo:** Yeah, Model Context Protocol.
+
+**Kendall Miller:** Yeah, thank you. Like, how does this work? Why does it work? What are the limitations? How does it interact with the browser? How does it interact with AWS? Is it going to write this kind of code for me, or do this kind of interaction for me? The rabbit hole always goes deeper.
+
+**Jerod Santo:** It does. It does. And I was just going to ask how much you're using these tools in your own work, because you have so many disparate tasks, it sounds like, lined up for any given day... Do you tell the AI "Please spell Jerod's name wrong for me as I send this email out", or how do you...? How much are you using these tools in your work?
+
+**Kendall Miller:** \[laughs\] Well, what's interesting is -- so for internal things, I actually feel like ChatGPT has just recently... And I mostly use ChatGPT. I dabbled with Claude and a couple of the other things... At Civo, I use Civo's models, which are Llamas that we're hosting. But ChatGPT just in the last few months feels like it's crossed over from "Hey, I need an outline on XYZ problem. What should I think about that I haven't maybe thought about?" And it would give me one or two decent ideas, and the rest of it would kind of be fluff. And now it's so good at it that I'm regularly thinking "I'd be stupid not to use this." And so a lot of times I'll type up a few things, "Hey, here's the 10 things I'm thinking about as it relates to this problem. What am I missing?" And it will structure a few of the things that I put together into the same category and then give me new categories.. I have it set up to be extremely, obscenely brief, which is funny, because I'm being obscenely verbose on this podcast... But in life I value brevity a lot. And I have it set up to always prompt me with "Here's what you didn't ask, that you probably should have asked." So at the end of every single response, it says "You should ask this."
+
+**Jerod Santo:** \[56:30\] Okay...
+
+**Kendall Miller:** And it surprises me how often it catches things that I didn't think. So to give you a concrete example - and the short answer is I'm using it all the time.
+
+**Jerod Santo:** Yeah. It sounds like it.
+
+**Kendall Miller:** It's just completely gone through the roof. But just in the last month I kept trying to use it and trying to use it, and I couldn't find times and spaces where it made sense... And now it's become a second nature for me. But I had done a write-up on a bunch of things that needed to happen in this one organization, at this one company. And I fed it in and said "Make this better." And at the bottom there was this one section where I said "Here's some of the problems that I'm seeing." And ChatGPT said "If you want to sound like a senior executive, don't express the problems. Talk about how you're working to address the problem. Don't just bring them up." And I was like "Ah, this is embarrassing. I should 100% know this, should 100% be interacting this way." And it slapped me on my hand before I went and took that to the person I was turning it into, and that was for me the eye-opening "I'm stupid every time I don't do this", because it's so good at this.
+
+**Jerod Santo:** Yeah. It's like an outbound filter, more than anything else. It's just like, before I go outbound...
+
+**Kendall Miller:** Yeah, I mean, it comes up with categories I don't think about, it gives me feedback on my writing that I didn't think about... It tells me to be more brief because I'm overly verbose in my writing... I mean, even today I asked it to put together a couple of lists for me. Hey, I need a list of this kind of company, doing this kind of thing. And I want 50 of them. It takes 20 seconds. It's wild.
+
+**Jerod Santo:** It is. It really is. Now, does it name your liquors for you, or do you do those yourself?
+
+**Kendall Miller:** Thus far we have a very long spreadsheet of potential names. I will say, the Generative A Rye is amusing, because I popped into the -- so the CTO Lunches, we have the free group that's 1,600 CTOs that all meet for lunches all over the world. And then we have a paid group; that's a small community that hangs out at a Slack. And right before we launched the whiskey, I went to the group and said "Hey, I have a couple of ideas for the whiskey name, and here's what I think it's going to be." And one of the people in CTO Lunches said, "Well, what about Generative A Rye?" And that ended up being what we use.
+
+**Jerod Santo:** Now, did they get a commission on that, or...?
+
+**Kendall Miller:** He's asked for a commission. I think I gave him a bottle.
+
+**Jerod Santo:** \[laughs\] That's the least you could give.
+
+**Kendall Miller:** I promised him a bottle. So if I haven't given him a bottle, you know who you are, Topher... I owe you a bottle.
+
+**Jerod Santo:** There you go. Give him that $1,400, that special edition...
+
+**Kendall Miller:** That special edition.
+
+**Jerod Santo:** \[unintelligible 00:59:07.04\]
+
+**Kendall Miller:** I handwrote the number on all of them. I'm holding bottle 22 right here.
+
+**Jerod Santo:** Oh, that's cool. Gosh, you're really selling me them.
+
+**Kendall Miller:** They're all small batch.
+
+**Jerod Santo:** I'm wanting one of those. Dang it. Now you price-anchor me. I'm thinking $125, affordable. \[laughs\]
+
+**Kendall Miller:** Feels like no big deal. Yeah. Knock on the door. Would you like to buy a $50,000 vacuum? No? Well, great, because I've got a hundred-dollar vacuum.
+
+**Jerod Santo:** So can you -- these are probably a private spreadsheet, but can you humor us? Can you workshop a few options out there? Maybe get a reaction...
+
+**Kendall Miller:** Oh my gosh, let me see if I can. I think that I might not be willing to share any of these.
+
+**Jerod Santo:** Okay. Take a moment and review... Maybe give me your bottom half. Don't give me any of the real good ones. Give me what ChatGPT would spit out.
+
+**Kendall Miller:** \[59:54\] I don't know if I can even pull it up real quick... Let's see.
+
+**Jerod Santo:** Because that's what I've found specifically with creative work, is that ChatGPT is so mid on everything that's supposed to be funny, or ironic, or tongue in cheek... It's like, no, dude... Like, the cheeseball factor is at 10, and they can't quite do anything funny. But boring prose is fine.
+
+**Kendall Miller:** Okay, so I opened it up. There is a number in here that I'm happy to share.
+
+**Jerod Santo:** Alright, let's do it.
+
+**Kendall Miller:** So we thought about InGINeering for a gin...
+
+**Jerod Santo:** Say that again? I missed it. I missed it entirely.
+
+**Kendall Miller:** InGINeering, but the GIN is all capitalized.
+
+**Jerod Santo:** Gotcha. Gotcha.
+
+**Kendall Miller:** One thing that I definitely hope that we do is I want to have a vodka that we sell, and it's two different labels. It's the exact same fucking vodka, and you can only buy them in a box set, and one is called tabs and one is called spaces. Because it's vodka, and they're both empty, and I just... I love this idea.
+
+**Jerod Santo:** I like that one.
+
+**Kendall Miller:** I think we will get there eventually. We have Incident Response Emergency Kit as a name... We have "I did SOC-2 and all I got was the stupid bottle of whiskey"...
+
+**Jerod Santo:** I like that one.
+
+**Kendall Miller:** One of my favorites is actually "I work in computers", because I think for the people who aren't in tech but want to buy a gift for their person who does work in tech - I mean, most of them don't know why Force Push is funny, and they're like "That's --"
+
+**Jerod Santo:** Right. It's almost too inside --
+
+**Kendall Miller:** Yeah, maybe it's inappropriate. They don't know what to buy. So the "I work in computers" as the gift to give to the tech person I think is really funny.
+
+**Jerod Santo:** \[laughs\] I do like that one.
+
+**Kendall Miller:** Yeah.
+
+**Jerod Santo:** That might sell.
+
+**Kendall Miller:** So we've got a lot more... Yeah, I'm saving the good ones.
+
+**Jerod Santo:** Yeah. I would, too. Why waste them on a silly podcast?
+
+**Kendall Miller:** I mean, it started with -- "A Friday Deployment" was one of the names of the liquors, and we later decided that was going to be the name of the company. Friday Deployment Spirits. Because you want to get in that Friday Deployment Spirit. Either you don't want to deploy on Friday, or you have to deploy on Friday and you're going to need to drink...
+
+**Jerod Santo:** That's right. Might as well be some gin.
+
+**Kendall Miller:** That's exactly right. What else do you want to talk about, Jerod? Anything else that I haven't covered?
+
+**Jerod Santo:** I'm just wondering how you like gin personally. I think gin is disgusting, and it tastes like a pine tree, so...
+
+**Kendall Miller:** Oh man, that's exactly right. That's what you're missing.
+
+**Jerod Santo:** \[laughs\] You like to drink pine tree. Well, you are from Colorado. There's a lot of \[unintelligible 01:02:25.25\]
+
+**Kendall Miller:** There you go. Well, when I was a kid, my dad would drink gin occasionally, and I would smell it and I was like "That smells disgusting." He'd be like "Have a sip." And I'd take a little sip and I'd be like "That's disgusting. That's disgusting." And I remember growing up and being like "Is that the one that tastes like pine needles? Ewgh, no."
+
+**Jerod Santo:** Exactly.
+
+**Kendall Miller:** And then I became an adult, and I had some pine needles, and I was like "This is amazing...!"
+
+**Jerod Santo:** \[laughs\] What changed?
+
+**Kendall Miller:** I don't know.
+
+**Jerod Santo:** It's still the same taste. Yeah, it's weird.
+
+**Kendall Miller:** For what it's worth, Jerod, our gin has some juniper in it, and you can tell, but it also has a lot more going on. So if you're not a big juniper guy and you want to try a G&T that you might be really excited about...
+
+**Jerod Santo:** Alright. Well, I'll tell you what - if you send me a bottle, I'll give it a try.
+
+**Kendall Miller:** If I send you a bottle, will you do an entire hour and a half episode where you just sit there and drink it and then give a review of it that gets increasingly unhinged as you get farther and farther through the bottle? Because then I will send you two bottles.
+
+**Jerod Santo:** \[01:03:21.28\] I can get you halfway there. Well, I don't like the taste of gin, so I can't promise I'm going to drink it for 90 minutes, you know? But I will try it on the show, and maybe send one to Adam, we could have a drink, and we could maybe do a review. Probably not 90 minutes, but... Yeah, I could do something with it.
+
+**Kendall Miller:** I believe you can stretch it to 45 minutes. I will 100% send you a bottle of liquor if you do a review on the Changelog.
+
+**Jerod Santo:** Alright. Thus closes my sales pitch... \[laughter\]
+
+**Kendall Miller:** So wait - we came into this hoping to have a bubbly conversation, you got a little bit of that, and you're going to end with a bottle... You want the whiskey, too. I mean, let's be honest.
+
+**Jerod Santo:** Oh, I actually do like whiskey, whereas gin I don't like.
+
+**Kendall Miller:** Do you like rye whiskey?
+
+**Jerod Santo:** Probably. I'm not a connoisseur. But what's the difference? What's rye's specific...?
+
+**Kendall Miller:** Rye is more peppery. It's a kind of different -- I mean, I guarantee you've had rye whiskeys if you're a whiskey drinker, and you might just not... Unless you try a rye next to a wheat, you might not notice the big differences.
+
+**Jerod Santo:** Fair. Yeah, I don't have an advanced palate. I do know that I prefer bourbon, and that's --
+
+**Kendall Miller:** I don't have a refined palate either, and I make liquor for a living, so it's cool. Sorry, keep going. What were you saying?
+
+**Jerod Santo:** No, I was just going to say, I generally like bourbon, which I think has more of a sweety -- you know, more sweet than peppery. But...
+
+**Kendall Miller:** Yeah. Yeah. The corn gets sweet, but also a little bitter... A wheat one is like really smooth, and sweet, too. But a wheat makes bourbon feel downright offensive. If you drink wheat whiskey for a week and you tried bourbon again, you'd be like "Whoa...!" Because it's got so much stronger of a kick.
+
+**Jerod Santo:** Gotcha. So it's way smoother.
+
+**Kendall Miller:** Yeah.
+
+**Jerod Santo:** And a rye is going to be in the middle?
+
+**Kendall Miller:** I mean, a rye - it depends on what the other grains are in it. The rye tends to be more like a -- I say pepper in part because it's a little bit of a spice that's sort of added onto whatever the other flavors are that are going on. So you can have a very sweet one with a rye kick, you can have a bourbon with a rye kick...
+
+**Jerod Santo:** Gotcha. Well, I'm excited. I'm excited for these --
+
+**Kendall Miller:** I'm excited if you're excited.
+
+**Jerod Santo:** I wish Adam was here, because he's actually more of the whiskey guy than I am... He's got all the -- all the stuff you just said, he would have been saying the same thing, assuming it's true...
+
+**Kendall Miller:** Just make him feel guilty. Guilt covers a multitude of wrongs.
+
+**Jerod Santo:** That's right. Alright. Well, Kendall, awesome meeting you. Great speaking with you. Anything --
+
+**Kendall Miller:** Thanks for having me.
+
+**Jerod Santo:** Anything you were wishing I'd ask you, and I never did, and you're like "Gosh, this guy's a dope. I've been waiting the whole time"?
+
+**Kendall Miller:** I mean, I think the obvious question that you didn't ask is "How can a man be so friendly, and look so good, and be so humble?" And I just don't have an answer to that, but...
+
+**Jerod Santo:** Some of the world's mysteries will just have to be left as such.
+
+**Kendall Miller:** \[laughs\] Thanks for having me. This was fun.
+
+**Jerod Santo:** Yup. Thank you.

NATS and the CNCF kerfuffle (Interview)_transcript.txt

@@ -0,0 +1,467 @@
+**Jerod Santo:** Today, we're joined by Derek Collison from NATS and Synadia. Derek, you were on one of my favorite episodes of Go Time years ago. "Challenges of distributed messaging systems." Yes, Go Time is no longer being produced, but the feed is still out there, the episodes are still out there... Please, go listen to that. I learned so much from you about these things. So welcome to the Changelog.
+
+**Derek Collison:** Ah, I appreciate you guys having me on. Thank you.
+
+**Adam Stacoviak:** Literally years in the making. Jerod, do you remember going to KubeCon a while back?
+
+**Jerod Santo:** I do.
+
+**Adam Stacoviak:** The Circus?
+
+**Jerod Santo:** That's right. The circus in -- was it Chicago?
+
+**Adam Stacoviak:** A couple of years back. We met the Synadia folks and the NATS folks... I think it was not quite early days for Synadia, but like it was early-ish for announcements, maybe even marketing. Derek, you can kind of speak to this, but... Yeah, do you remember that, Circus, and meeting those folks?
+
+**Jerod Santo:** I do.
+
+**Adam Stacoviak:** Yeah.
+
+**Derek Collison:** I don't know if I was there. What year was it? Because Synadia was formed at the end of 2017, on the heels of Apcera. But I don't remember... I don't think I was at the one in Chicago.
+
+**Adam Stacoviak:** Many years after that. 2023?
+
+**Derek Collison:** Okay. So not too far back, yeah.
+
+**Adam Stacoviak:** Yeah. Tail end of 2023.
+
+**Jerod Santo:** Yeah, you weren't there, but we had a few friends there that were listeners as well, and we came to know each other, so... It's been nice knowing you all. I've known about NATS for a very long time. When did NATS start? I mean, it's been around forever now, right?
+
+**Derek Collison:** Yeah. NATS was kicked off as part of a system that I was building at the time at VMware called Cloud Foundry, which was my spin on Heroku for the enterprise back in 2010, I guess. And at the time, I had come off of time at Google, I spent about eight years there, but prior to that was at a company called TIBCO, which before that was called Technicron. And we had been designing lots of messaging systems that then were used to scale up distributed systems for customers and clients. So I just had a natural way of building systems, like command and control, telemetry, eventing, using these types of constructs. But TIBCO was closed source, and so we couldn't use that.
+
+So we reached out to Alexis Richardson, who was running a company that was kind of in charge of RabbitMQ, which at the time was kind of the major open source player from a messaging construct. And RabbitMQ is still part of, I think, VMware/Broadcom now, part of Cloud Foundry or whatever that has turned into be. But at the time, it was more of an enterprise type of a technology. What I mean by that is that if you asked it to do something, it would bend over backwards trying to do it for you, even at the detriment of everyone else. And we had some challenges there, which led to the creation of NATS, which was kind of like a dial tone, and think of it like the electricity in your house, or your condo, or whatever. I can plug in a really bad blender and I might blow out something in my own place, but hopefully it doesn't take out the electric grid for the whole neighborhood. And so NATS was born on a very simple principle of "Just do a couple of things very, very well." And yeah, it was the proverbial weekend project. So I left on a Friday, and Monday we plugged it in.
+
+\[08:35\] And believe it or not, applications that were written for that initial version, the protocol that speaks between the clients and servers still run today. So you can just fire up something from the Cloud Foundry days and just run it against a modern NATS server and it just works. So we've tried very, very hard - and we're not perfect by any stretch, but we tried really, really hard to make sure we were always backward-compatible with the protocol. Because we knew the protocol, even though it was designed very, very quickly, in only a couple of days, that was kind of the really big thing from a technology standpoint there.
+
+**Adam Stacoviak:** I'm not a distributed systems person, but NATS feels a lot to me like Tailscale, but at a different layer. Is that how you would describe it to somebody who doesn't build distributed systems, or have sort of the depth and knowledge you might have?
+
+**Derek Collison:** Yeah, I'm a huge fan of Tailscale, and the folks that were either at Google, or involved in the Go programming language that are over there. But it's an overlay on top of the constructs that we still know, which is IP-based addressing, almost everything is point-to-point... I don't know about Tailscale, but most cloud providers write trunk UDP broadcasts, or multicasts, and so it's a point-to-point, location-dependent type of a system that now has surfaced up over a very secure overlay with WireGuard.
+
+What NATS was trying to do was kind of change the notion of what we called intelligent connectivity. And specifically, when folks that are listening say "Well, what does that mean?" It's really fairly simple from an abstract standpoint, which is everything is location-independent. So one might say, "Oh, well, things are kind of location-independent today", but I would argue that there's a lot of unnatural apps going on below the covers to get that to appear that way. Load balancers, GSLB, AnyCast \[unintelligible 00:10:27.08\] So the biggest first one is location independence, and the second one is instead of one-to-one request reply as the dominant pattern - HTTP, that's all you can kind of do, sans SSE and WebSockets - we are end-to-end and both push and pull. And the push becomes very, very interesting for certain use cases, where instead of keep asking "Hey, have you updated your temperature? Have you updated your temperature?", just tell me whenever you update your temperature. And for scalability and distributed systems, just those primitives from the connective layer become pretty powerful at scale.
+
+**Jerod Santo:** So NATS really took off, right? I mean, huge adoption. Can you tell that story?
+
+**Derek Collison:** Yeah, so - I mean, NATS was built specifically to power those subsystems in Cloud Foundry. And I had no desire to do anything more than that. So it was, for at least probably the first 18 months of its lifetime, specifically geared toward just supporting what we needed in Cloud Foundry; from command and control, telemetry, eventing, things like that. When we started the company Apcera, which was trying to do a redo of Cloud Foundry on some of the things I felt I didn't spend enough time and energy on - security, governance, federation, all these different types of things - we rewrote NATS in Go. NATS was originally written in Ruby. That's what Cloud Foundry was written in. I still love the language... Not great to deploy things into production that are Ruby-based, although I know DHH has done an amazing job at trying to simplify that. But back in the day there was a lot of moving pieces to deploy the same...
+
+**Jerod Santo:** Gnarly.
+
+**Derek Collison:** \[12:03\] And we were starting to hit some performance issues, with very, very large systems needing more performance there. So we had -- at least at the time, we were looking at either Node.js (JavaScript), or Go, which was very brand new. I think when I started playing with it, it was 0.52. And I felt that the speed advantage, with the concurrency, the IO mechanisms that they had, but mostly the fact that you could build pretty much a static binary, meaning shipping to production was just move one binary, was the direction that we took. And it was a great direction. I mean, I think Go has grown up and there's people that like it and people maybe that don't like it, which is totally fair... But we've gotten a tremendous amount of mileage out of it. And at the time, looking back, it was the absolute right decision.
+
+Now, going forward, when we're looking at massive, massive extreme scales that some of our customers are pushing us on, we're starting to bump into some things in Go that we need to address. And whenever I'm in a language and we start to do unnatural acts to get around certain things within the language, that's when we kind of say "Alright, let's take a step back and - are we at another inflection point?" And we're not there right now, but with the introduction of JetStream, which is the data persistence mechanism that NATS uses, that was added on about four, four and a half, five years ago, I guess, we're starting to see a lot of friction with the way Go does disk IO and things like that. And so we're trying to figure out what makes the most sense as customers are saying "Hey, remember when I said a million was good, then I said 10 million was good? Now I need a billion, you know what I mean? Could you do that in the next two years, type of stuff?" And those are great problems to have, because they're hard and they're fun to work on... But it challenges you to think outside the box sometimes.
+
+**Jerod Santo:** Yeah. What are you looking at?
+
+**Derek Collison:** You know, I don't do it anymore... I'm sorry to say that I've dropped the ball. But I used to learn a new language every year, and I would just write a NATS client for it to learn it. And so I learned Rust, and there's certain things in Rust that were really, really attractive. But to be frank, I just didn't enjoy programming in it, when you were doing a low level library. If you were consuming other things and building like an upper level app, that was kind of a composite of lots of different things together, especially the packages or crates were really, really well done. It felt really, really nice and clean. But I was at a low-level library, trying to do some things... And we got it, and it's one of our most popular clients.
+
+Again, I'm older probably than the majority of the demographic that's listening to the podcast, but I like languages that I can hold most of it in my head. I've never been a big IDE person. I still program in Emacs. I do use AI, and I ask Claude or ChatGPT questions and things like that, but I don't have that Copilot, Windsurfer code mentality yet... But I do like Zig quite a bit. But Zig right now -- you know, Andrew was trying to take on some really gnarly problems, which I have the utmost respect for, especially the colorless async... And that kind of tied in a little bit to event loop structures, and they kind of got pulled out. We need some sort of direction there before we finalize a Zig client, but I did write one. I have one that kind of uses threads and does some interesting things.
+
+And then Joran from TigerBeetle, I've known him for a while... Full disclosure, I'm a minor investor in their company, and I love All Things Distributed; I think that was their conference. And I've really been watching them, and Bun, and of course, Mitchell with Ghostty on what's possible. But each one of those event loops and io\_uring interactions is all kind of a little bit snowflake, not exactly... But I think Andrew is now thinking through a way to uplift the abstraction of IO, that could tie into an event loop as something similar to allocators in the language, where you have to pass it into a package or to a constructor, and say "Hey, if you need to use IO, please use this essentially interface." And so I'm really interested to see where that lands. But I like Zig. Believe it or not, I like Mojo, even though I'm not a Python person.
+
+**Jerod Santo:** \[16:33\] Really?
+
+**Derek Collison:** I just am a huge fan of Chris Lattner's and what he's trying to do there. I think it is impressive.
+
+**Jerod Santo:** Sure.
+
+**Derek Collison:** But Zig's probably number one right now in my mind. But again, we need a little bit more direction... Not from the server perspective, but if we wanted to do a client on "How do we interrupt between all the different event loops?" ...whether they reincarnate the one that they have in the language, but then Mitchell's \[unintelligible 00:16:55.14\] I think, and then of course Bun has one that's, I think, a pull from TigerBeetle's... And TigerBeetle has one... And so that'll be good when we get there.
+
+**Jerod Santo:** Well, a completely different problem space, but you like Chris Lattner... The Ladybird folks are probably picking up Swift; from C++ to Swift. Have you considered Swift?
+
+**Derek Collison:** I haven't. We built a Swift client. I wasn't involved in it. Again, I dropped the ball on my learn a new language every year. I actually did program a little bit Swift... I think the interesting part about where Zig is is that we can control direct IO, meaning we could go through io\_uring on Linux, if it's there, applicable, or whatever... And it allows us also to control more precise layouts of the data structures, and of course memory.
+
+So anyone who's been programming in Go for a long time and has a tremendous amount of investment in a big codebase, you start to see things pop up. It's like "Oh, avoid JSON at all costs, because the built-in JSON marshallers \[unintelligible 00:17:59.17\] lots of memory." The network IO and the disk IO subsystems in Go are very different. And the disk IO subsystems are pretty much a native thread that kicks back into your goroutine that says "Okay, it completed. The write completed", or whatever. And I think what we need is a lot more fine control over those things.
+
+So what I've been looking at is the adoption or the ability to introduce io\_uring in a meaningful state. And I think Zig's done a pretty good job there. Maybe not within the language per se, but with folks that use it. Tiger Beetle and Bun come to mind. And so that one's kind of a big one. The data structure layout, the garbage collection and taking control over disk IO... Which - JetStream, we've got folks that are coming in and say "Oh yeah, I need to do a million stored messages per second, replicated three ways, and then an async mirror over in a different region and a different cloud provider" type stuff. So we can abuse the disk, the storage subsystems pretty well, and we might be getting close to kind of everything we can accomplish with Go... I'm not saying that that's definitive, but it's starting to feel that way, as we're starting to do - not unnatural acts, but we're starting to do a lot of things inside of the server to get around how Go actually marries concurrency with disk IO.
+
+**Jerod Santo:** Hm. So when you said some of your customers need like a billion - is that like message passes per second? Or what does that mean, a billion? A billion what?
+
+**Derek Collison:** \[19:38\] Some need a lot of messages per second... NATS has always been pretty performant. Last time I checked, just raw speed, a single server could do about 20 million per second messages sustained. All of our clients can't do that. They have more of the logic, but a raw just, you know, shove it in there... But when we say billions of things, it's usually stored messages, and then subject cardinality. So in event sourcing and event streaming, every single message has a unique subject. And what we've found was is that people are starting to treat the stream of messages... And we have materialized views, like key-value and object store on top of them, but they're starting to treat them like a subject-based addressing lookup thing, meaning that they use subjects all the time to get messages out, or filter messages, or do some other operation.
+
+And so you could imagine trying to figure out an index that looks like a traditional data structure that anyone who's done a comp side, 200, 300-level class would understand... But overlay subject semantics with wildcards and partial wildcards is a little bit tricky, if that makes sense.
+
+**Jerod Santo:** Yeah... Like a giant hash table in the sky, that also has some very gnarly semantics that you can use.
+
+**Derek Collison:** Yeah... And to be honest with you, it's called the \[unintelligible 00:20:57.00\] and it's an adaptive radix tree, with path compression and expansion, but then it has an overlay where we can say "Hey, you just stored a billion subjects in here, hopefully in an efficient way. Pick out and call me back on everything that matches foo.\*", you know what I mean? In other words -- or a wildcard. And so we have that, but that lives in memory. And so if you have 10 million messages, but only a thousand subjects, you're fine. You'll never even see this. But we're starting to see people where they've got billions upon billions of messages, and everyone has a unique subject. Next thing you know your server's taking up 30-gig just for the indexing on the subject-based addressing layer.
+
+**Jerod Santo:** Right...
+
+**Derek Collison:** And so that's one of the big challenges, is that the team is looking into solving for 2.12, the next version of the server.
+
+**Jerod Santo:** Well, it's cool that you still have your head down in the technical trenches of this project, because it seems like, from what I've seen from you all and your name signed on things recently, it's all been higher level business, and maybe technical legalese, not technical technicals... But it's cool that you're still close to the metal of this project, and not just the business guy now.
+
+**Derek Collison:** Yeah, I mean, I definitely wear multiple hats, obviously... And as of late, I've been more concerned about upper level discussions around Synadia, and NATS, and the CNCF, obviously, which we can talk about, but also the business side of Synadia... But I identify with builders. I feel that working this hard, you want to get to a place where you get to do the stuff you like to do, at least part of the time...
+
+**Jerod Santo:** Right.
+
+**Derek Collison:** And I still love doing that. Am I as good as other folks in the company? Probably not anymore, but I still love to reason about things, think hard about things, tackle really, really hard problems... That's kind of my relaxing motion in terms of work, if that makes sense.
+
+**Adam Stacoviak:** So if NATS was roughly around 2010-ish era, and you really didn't plan to do anything with it beyond its original case, at what point did you say "Well, you know what? We should be a company. We should create Synadia." And that was 2017, based on what you said before. How did that -- how did you go from there in 2010 to there in 2017?
+
+**Derek Collison:** Yeah, it's an interesting story. So one of the things that folks might not know is that when we were doing Cloud Foundry, VMware had never released anything that was open source. And my pitch to Paul Maritz at the time, who was the CEO who was running VMware - I said "I really think we'll have more bang for the buck here if this is released as open source."
+
+And so when I did NATS, I said "I'm not going to keep this closed", so I actually just opened it as MIT. And then we did -- I can't remember what license we put on Cloud Foundry, but it was open source, which was a really big deal. When we went to Apcera, we did the rewrite in Go, and we started to contemplate whether or not this could be something bigger than just an OSS project. I mean, it was powering all of the Apcera stuff, but similar to Cloud Foundry: Telemetry, command and control, and eventing.
+
+\[24:08\] And so as we looked at Apcera -- and Apcera was doing a commercial version of Cloud Foundry plus plus, to be very, very simplistic about it. But as soon as Kubernetes came out, certain things Kubernetes did that people liked, certain things that the Apcera Continuum platform did, they liked... So it wasn't a technical battle per se. It was "Where is the consumer bias going?" And Kubernetes, because no individual stakeholder needed to make direct money off of it, drove the consumer bias to zero. And so our ACV started declining. And so I called an emergency board meeting and said "Hey, I see the writing on the wall, and all I can tell you for the 30 plus million dollars left in the bank over the next four years is that we can tread water, but I don't think it'll change the outcome." And then I think pulling it into Ericsson and doing something constructive with some of the internal projects that I know they were working on made better sense.
+
+At that point in time, there were certain people that were going to leave, certain people that were going to stay with Ericsson, and then there was NATS, the team that was working on NATS in the middle. And at the time, Ericsson said, "Yeah, we don't necessarily have any need for NATS. It's open source, we can use it... But if you want to do something with it, we can." And a bunch of folks internal prior to that decision-making process had already kind of approached me saying "Hey, we might be able to build a business out of this", under the assumptions that distributed systems will continue to have more moving pieces, but more importantly, that these pieces are going to start to be stretched out. So stretched out across regions, across multiple cloud providers, but more importantly out to this - at the time - new thing called edge. And the bet that we made was that edge would dominate interaction models and would dwarf cloud within probably a decade... And that cloud's not going anywhere - don't misread that - but that cloud would become the new mainframe, but all the interactions would be at the edge. And the rules are very, very different. So as many rules changed as we went from data centers into cloud, and we had the Heroku guys, Adam, and the -- was it the 12 things...?
+
+**Jerod Santo:** 12-factor app.
+
+**Derek Collison:** 12-factor apps, and then of course, the cloud-native architectures, and of course then the CNCF... All of those things - that step function from data centers to that I believe will be even bigger stepping from cloud into edge, where it's not batteries included. You can't bring all the stuff that the cloud is offering, and you have to kind of - not think differently about the what, meaning you're still drawing the same triangles and squares on a whiteboard as you're architecting, but the how would be radically different. And so that's when we made the leap to create Synadia, at the end of 2017, based around a foundation that modern distributed systems will look very different than they do today, and they will be pushed very hard in terms of dynamic agility. They can change all the time.
+
+Again, I probably am further along career-wise than some folks on the call, but it used to be that when a system was set up like this in the '90s, early '90s, no one touched it. I mean, I can walk into any financial institution on Wall Street, and I could go into their server and point exactly where our stuff was running. And it never deviated from that. And that's very different than what I think a lot of the audience probably thinks now, where it's like "I don't know, it's somewhere in AWS East 2", you know what I mean? "But I have no clue what machine it's running on."
+
+And so our thought process and our priors were that that type of step change is going to accelerate with edge, and that the current tools of doing everything with like HTTP or HTTP derivatives just simply won't be able to keep up, in our opinion.
+
+**Jerod Santo:** How has that bet played out?
+
+**Derek Collison:** \[27:56\] Um, very, very well. What I mean by that is that in the '90s, with TIBCO, I was on the engineering side. I wasn't in an executive role, or had any type of position around pricing, or anything like that. But just from an engineering perspective, we cared deeply about interacting with these customers, and such like that. When we watched what was happening with edge and how fast it was moving and how different the rules were, at TIBCO we had to educate the market. And I remember how painful that was. A lot of the sales cycles were 80% education. We need to explain to you why we think you are going to have a problem in the near future, that we think we might only be able to solve.
+
+In this world, which was good for Synadia, people were coming to us going "Hey, I tried to pull Kubernetes, and JVM-like stuff, like Kafka, and put it into an edge thing, and it just -- it didn't work for me. And I need to think differently about it. Can you help us?" And so when we started Synadia, we wanted to start a company based on "If you're running a production and it's critical to what you're doing", which luckily our stuff is critical to every customer that we have, meaning there's no coffee breaks. If it goes down, they will light my phone up... That that was the opportunity we were looking for in terms of starting a company. And we didn't have to educate the market. They were already in pain, to start the company, and that they would want a commercial agreement with us to wring our neck if something went wrong. And I think that's a really good way to start a company. It's not the best way to scale it, so you need a backup plan as you evolve... But to start it, I think that's a really good place for folks to start, especially if it's based around OSS.
+
+**Break**: \[29:41\]
+
+**Jerod Santo:** So OSS... Should we open the can of worms around OSS?
+
+**Derek Collison:** Sure.
+
+**Jerod Santo:** Most recently there's been a - I don't know what you call it; a skirmish, a disagreement...
+
+**Adam Stacoviak:** A kerfuffle.
+
+**Jerod Santo:** ...a kerfuffle between CNCF and yourself, at Synadia, specifically around NATS and its involvement in CNCF, and a potential relicense... Of course, we've been here many times in the world of open source and business and licensing... It's something that everyone's trying to figure out, and some people have or have not, and then other people think they have, and then they have not... And then things change, and people are offended... And so there's been a lot of - I don't know, maybe drama is just the best word, around the business and the project. Maybe start with the CNCF and NATS' relationship to it prior to the most recent blog posts and press releases.
+
+**Derek Collison:** Yeah, so what the audience might not know is that the CNCF was kicked off at a large meeting at the Switch Data Center in Las Vegas, of which I attended. And I was actually part of the founding governing board. So a lot of people might think "Oh, Derek's role with the NATS project and NATS being a project in the CNCF - that's kind of the totality of it." But that's not the case. And even early on, we were looking at what value CNCF could bring to the ecosystem from a consumption perspective, the users, to the people or the companies that want to utilize that, and then of course, the projects themselves... And then for me, it was like the companies that are driving said projects. In the case of something like Kubernetes, where it's all of the big boys, so to speak, there's not a lot that the CNCF needs to do there. It's kind of taken on a life of its own and it's marching down a path.
+
+What we started to realize is that as time goes by - because this was, I think, over a decade ago - that both the CNCF and projects can evolve. And where we were at was looking at the question "Are we the best fit for CNCF?" And no fault of CNCF's per se, but it just felt pretty clear to us that certain projects were at this tier, and other projects were way down here, and we kind of were in the latter category.
+
+In no way were we attempting to relicense the whole codebase. I mean, one, you can't do it, because once it's released, it's AP2, it's there. That will never change. But the CNCF in early discussions -- because we said "Hey, would you be open to letting the project leave if the landing spot made sense?" Meaning we could do a joint statement where it's like "No, it's landing in this new foundation, made up of lots of big users and customers or production uses of the NATS ecosystem. Commitment to AP2, X, Y, Z." And during those early conversations, which started in February, we were asked "Hey, are you considering a license change?" And we're very well aware that we can fork and do a Synadia version enterprise server, and X, Y, Z, and so we said "We're considering, just for the server, that we might do like a BSL." And at the time, the BSL - for us, it wasn't translated well out into the media... But the reason that we were considering it was we felt that it was the best show of commitment to both OSS and to our customers. And I'll explain what I mean by that. The BSL has usage clauses that you have to define when you release the software. And one of them is there's a period of time after which it converts back to whatever language you pick, which - we would pick AP2. And it doesn't matter what Synadia does, the Linux Foundation does, CNCF does, that is a legal contract that once we release, let's say, a server with the BSL license, after two years it's converted to AP2. Even if we don't actually update the copyrights, or update the license, or whatever. And so we felt that from an OSS ecosystem.
+
+And again, we've gotten lots of additional information, lots of additional feedback. Most of it constructive, some of it not. And we've been listening. But our initial thought process as we were going down this path was "Hey, this signals really strongly that we're not going to hold anything back. It will always kind of revert back to generic drug prices." Some of the good parts of what the patent system originally was trying to do early on. I'm not claiming that the patent system is necessarily good these days. It's a necessary evil, right?
+
+\[36:16\] In addition, if you would imagine a two-year BSL window, and let's say Jerod wanted to be a customer of Synadia - and customers are worried about vendor lock-in, price gouging, all those normal things - you could say, "I want to do a two-year contract, so I have predictable pricing", and at the end of that, that all stuff that I'm paying for converts to AP2, and I can now use it for free. Or I could say "Oh, I like the next version, too. I want to use that and \[unintelligible 00:36:39.10\] and do that.
+
+So I'm not saying that it transpired the way we wanted it to, but I think people's thought process that we're just evil and we're just trying to price gouge and be greedy and stuff like that - I don't feel that that necessarily is correct, at least as we were discussing internally. We care deeply about our commitment to open source. But we are starting to see a very disturbing trend where customers that everyone would recognize - they're in the Fortune 50 - that are using NATS to power production-level services or functions or products, not only had never reached out for any type of commercial agreement with us, but actually have policies that say "If you're in the CNCF and you're incubating and graduating, we're not paying for it, period."
+
+**Jerod Santo:** Really?
+
+**Derek Collison:** And to me, I thought that that was at least worth the discussion. And it might not be a discussion between myself or Synadia or the CNCF, but the broader ecosystem, I think, needs to look at "Do we want open source projects that drive value, that can be critical, that are driven by a small group of individuals or a company that's trying to make a business out of said value?" And Synadia has lots of different revenue buckets, so I don't want listeners to think that this is our only drive... But I was at least concerned enough to say "Are we signaling to the ecosystem, the broader ecosystem outside of us, that we only want projects that look like Kubernetes? We don't want any projects that are being driven by a single company." And I thought that was worth the discussion. For sure.
+
+**Adam Stacoviak:** How much time before the blog post that I think broke the news, so to speak, from the CNCF directly - how much behind the scenes conversation was going on to sort of illustrate your points and your concerns?
+
+**Derek Collison:** Quite a bit.
+
+**Adam Stacoviak:** Was there a meeting? Okay.
+
+**Derek Collison:** It had started in February.
+
+**Adam Stacoviak:** So was this was blog post a surprise?
+
+**Derek Collison:** Yeah. So we had gotten to a point where we were trying to assert the fact that we wanted them to consider letting us leave, and they were stuck on some of the trademark stuff. And so for the listeners who haven't been in this, when you're doing negotiations behind the scenes, you posture up and you do all kinds of stuff, and then you figure out when you can't move the stakes anymore, and then you make a decision. Right? And our decision was always going to be, "Hey, if they dig in, even if our legal teams say that we could potentially win this, we have to consider the damage to the ecosystem and whether or not it's worth it." And we had already probably decided that if they were just going to dig their heels in and say "Nope, you absolutely can't do that. If you want to go down that, it'll have to be a legal recourse", that we were going to say "Okay, we'll figure out something else." But they -- I can't speak on their behalf, and I'm not trying to, and again, I'm extremely encouraged by my talks with the Linux Foundation and some of the ongoing stuff with the CNCF, and that mutual statement -- again, everybody's had a hand in that pie, so it kind of gets diluted down...
+
+\[39:54\] But I can tell you from my individual discussions with folks, especially at the Linux Foundation, I'm very encouraged with them saying "Okay, let's see if we can do better here", which is all I really kind of wanted at the time. But I did feel that we could be a good steward, and if we could say "Hey, here's what we're committing to", and that is commensurate with kind of what the CNCF is doing around certain projects, I thought it was worth the discussion to have.
+
+Now, a lot of people don't agree with that, and have very emotional responses to that... And although I don't necessarily agree with it, I understand that people will fight very hard for consumption being zero... But production is never zero. Someone has to pay for it. I mean, unless it's just something -- and whether it's time, or money, or both... And so we were concerned enough to say "Hey, let's have a discussion about that."
+
+**Adam Stacoviak:** Mm-hm. The main concern, I think, if I understood correctly, around even making this a concern for you was the fact that Synadia was having challenges securing certain contracts because of its incubation status of NATS, which Synadia is built on top of as the commercialized version of it... Is that the summary of your initial concern?
+
+**Derek Collison:** No... I mean, we didn't graduate, and that's a public record, and it was mostly around diversity of the contributors to the server. So GitHub has all the stats so you can look at, all the people that have contributed to it. And I didn't agree that that was super-important for the success of our project long term. So I'm not saying the CNCF is wrong and I'm right, I'm just saying we had a differing opinion of what successful projects look like. And there's some that I've driven in my career that now have been around for 30 plus years, two of them have been around for 20 plus years, so I kind of get what it is, but it doesn't have to only look like Kubernetes, in my opinion.
+
+And so that was kind of the first one where I was like "Hm... The CNCF is evolving where they bring in lots of different projects, and that's a daunting task of them having to do all that", and hats off to them on that. But as you watched how things move through the channels, and got towards graduation, there was some friction on "Does the project look like what the CNCF wants it to?" And again, totally their right to do that. And then also, I was noticing little perturbations in companies doing strange things to try to drive the business model where they were the main contributor. Their company was the main contributor driving said project; you know, outside of NATS. And we were watching that pretty closely. We don't get too many people worrying about the CNCF or our graduation status. Every once in a while we do... It's like "Hey, why haven't you graduated?" and we just say "Oh, because the CNCF's graduation criteria looks like this. This is how kind of the server looks like." And it usually is that quick of a conversation; like 15-20 seconds max, with customers, every once in a while. I think I've gotten two of those in the last six months, and it was just a "Hey, why haven't you graduated?" type stuff. So it's not that per se; it was as the CNCF evolves, are all projects good fits? And is the only answer "You have to fork it away and then we'll kill it"? And I don't feel like that should be the only answer. How about that?
+
+**Adam Stacoviak:** Right. There's so much brand equity in the NATS name. Obviously, you founded it, you created it.
+
+**Derek Collison:** Yeah, I created the NATS name driving home from VMware, on a Friday, I remember. And everyone laughs, and the marketing came up with a funny term, but... You know, I remember just shaking my head, driving home, going "Geez, I don't want to have to create another TIBCO server." I was like "Not another TIBCO server..." I just kept saying that to myself driving home, but I said "I feel like I have to." Well, at the end, Not Another TIBCO Server is exactly how it got named.
+
+**Adam Stacoviak:** Yeah, yeah.
+
+**Jerod Santo:** \[43:58\] So when you first joined, or when you first started incubating - is that the entry-level term there? ...with NATS, were you aware of the CNCF's desire? Or were there requirements like "This project will move towards a non-single corporate backer model", or whatever language would fit? Did you know coming into it that they wanted you to kind of change the format or the org structure of NATS?
+
+**Derek Collison:** I don't think that was discussed at that level early on. So we were at the Switch kickoff meeting, we saw some of the early projects, they wanted us to be one of the earliest projects... I was hesitant, so I said "I want to be part of this, but I want to wait and see." And Google was doing similar things early on there. Like "We just want to kind of see how this is going to kind of pan out." But I think it was more from a sustainability, "How many people are using it? What type of impact does it have, both within the CNCF, but then externally?" were more of what I remember very early on of "How do we classify a graduated project?"
+
+Now, as the TOC - that's the Technical Oversight Committee - keeps changing and new people come into that, they started changing the criteria. And so instead of kind of a gut instinct on a vote type stuff, they started formalizing some of the things that they felt needed to gate between sandbox, to incubating, to graduate. There might be some more levels in there, I'm not sure. And so at the time, though, we weren't aware of that, and it was only kind of really solidified when we actually went through the graduation process. And again, that's all online, the back and forth between us and the TOC on trying to graduate.
+
+**Jerod Santo:** So was your failure to graduate then a shock to you, or did you see it coming? I did not see it coming. It took us a while - and again, it's all public record - to figure out what's going on here... And it was specifically about diversity of code contributions to core pieces of the project. Now, again, CNCF - I think they've done an amazing job with all of the different pieces that are flying at them... But if you look back historically, you could see certain projects that were like "Hm, how did that one graduate?" You know what I mean? That was a lot more similar to these types of projects than others. And I think they've slowly cleaned that up and kind of solidified that process and those criteria... It was just I didn't agree with the criteria. And again, I'm not saying I'm right and they're wrong by any stretch, but I did fundamentally disagree that not all projects --
+
+**Jerod Santo:** You must think you're right then, right? Like, if you disagree fundamentally, then you've got to think you're right. Well, in certain projects. Could Kubernetes have been driven only by Google and gotten to where it was? No, but it has a very different role of kind of greasing the skids between cloud provider mobility, and that was kind of some of the original \[unintelligible 00:46:56.27\] and the folks that built it... At least from me talking with Diane back in the day, of why they were doing it. But I don't think saying that that project fits means that a project like NATS doesn't fit.
+
+**Jerod Santo:** Right. But they disagree with that.
+
+**Derek Collison:** Correct.
+
+**Jerod Santo:** And they're the bosses of the CNCF, so they get the final say on that. Now, when your conversations began back in February, this is when you realized you were not going to graduate, or you'd already kind of known it for a while? Like, you had decided "Well, we're not going to be a good fit anymore." Right?
+
+**Derek Collison:** Yeah, that was -- the graduation I think was in 2020, as well as the trademark fight.
+
+**Jerod Santo:** Okay, so a long time ago.
+
+**Derek Collison:** So this was more of a "Hey, we just think we're not a good fit anymore, and let's figure out if if there's a way where both the CNCF and Synadia can message to the ecosystem that this is okay, ut as long as there's a landing spot that holds true to what the CNCF wants, commitment to open source, things like that, would that be possible?" And obviously, the answer is no, but I still think it was worth a conversation. And that's where we went.
+
+**Jerod Santo:** \[48:01\] Right. So in February, you wanted to kind of slip out the back Jack, it wasn't going to work out... And those conversations began. And then they asked you if you were thinking about a relicense and you said "Yes, we're thinking about BSL." And then there was probably some more conversations between there, but then they decided to publish the blog post... Or what am I missing in the interim, between --
+
+**Derek Collison:** So I think we were very clear in our commitment to AP2 and OSS. And that it would be a fork that we put BSL work into, that then would - again, in our opinion, would signal a commitment to OSS. Because no matter what we did or said after that two years, which was the number we had picked internally as we were going through that process, it would convert over. And so I was saying -- the licensing stuff is kind of separate. We're talking specifically about the open source nature of NATS, and whether or not it's a good fit, and if it can land somewhere else with both the blessing from the CNCF and the commitment from Synadia, that allows them to feel comfortable doing that.
+
+They then started getting to the point where we were starting to figure out when the stakes would move or not move... Again, and it's totally fine, negotiating types of stuff... And they put a hard stake in the ground, "Nothing will ever leave the CNCF. We'll kill it instead."
+
+**Jerod Santo:** NATS cannot leave the CNCF.
+
+**Derek Collison:** Nothing can leave. Ever.
+
+**Jerod Santo:** Nothing, ever, including NATS.
+
+**Derek Collison:** Correct.
+
+**Adam Stacoviak:** And that was my question too, was like "Has it ever happened before?" Which, I'm appreciating the details behind the scenes here.
+
+**Derek Collison:** No, it hasn't.
+
+**Jerod Santo:** Okay, so it's like Hotel California... Is that the one? You can check in, but you can't check out.
+
+**Derek Collison:** And again, I think that that's totally fine that that's their position. I still felt it warranted a conversation, because that doesn't feel right to me. So if you look at like Grafana - it had an interesting thing with Cortex, and they were going to start driving more stuff, but it wasn't necessarily compatible with the CNCF, but they were still committed to the open source nature of what they were trying to do. But it wasn't compatible. And so the CNCF said "Nothing ever leaves. So you can fork it and then abandon this one." And that's what they had to do. And again, I'm not saying they're right or wrong, but when they purport that's a success story, I talk to other people and it doesn't feel as much of a success story. Mimir started, and that's what almost everyone on the Grafana side uses, and I think Cortex is mostly supported by AWS, so that they can run it and not have to pay anything to Grafana. Again, the consumption of zero is a sense of entitlement that I also think is kind of interesting, but production side is never zero. Someone's got to pay.
+
+**Adam Stacoviak:** Yeah, it's a hard position to be in whenever -- and almost, knowing this, I think maybe this even exposes a different scenario where maybe you were not fully aware... Like Jerod was asking, were you aware of X, Y, and Z when you first joined the CNCF or you incubated that whole process? And I think over time, you get more and more privy to certain details that are truths about what it truly means to join the CNCF. And I'm not saying anything negative or bad, but not being able to leave, like coming in the door and not being able to leave with the thing you think about over a weekend and you name is challenging... Because if you don't pass the bar of incubation and graduate, and you're not a -- I don't know what it means to not graduate. Does that mean you're like not?
+
+**Jerod Santo:** What's it mean to graduate even? What does that mean?
+
+**Adam Stacoviak:** Yeah. I don't get all of it. Does it mean you're not valuable? Does that -- I don't understand the graduation statuses either. And I'm sure there's a lot of nuance, too.
+
+**Derek Collison:** \[51:44\] Yeah, there is. And even after our TOC process around graduation where we did not graduate, there was a slight change that said "Incubating or graduated projects are considered mature and production-ready." But within the grand scheme of things, someone from the outside, a user who's saying "I'm just looking for some software that might solve my problem", they see graduated as a higher level than incubating. And so they're like "Why haven't you graduated? Because you've been here for a while. You started at incubating back in 2018", I think is when we officially joined, if I recall correctly. I might be off on that one.
+
+And so yeah, I wanted a conversation... We had spent -- and I'm not saying that CNCF has not also worked to promote NATS, maybe not as much as Kubernetes and things like that at a KubeCon or whatever, but they did do a part there. But we lived and breathed this 7/24/365 for almost 15 years now. We've put everything can. And not just myself, it's just all the folks that were at Apcera, and then of course now at Synadia. And not saying that what we thought was necessarily correct per se, but I find it interesting when people go "Oh, you really don't care about the project", you know what I mean? Or "You're just trying to do a money grab and gouge people for money at this time." That's a little unfortunate. I get it. And again, I know this whole topic is very emotional, and I can appreciate that... But that's about as far from the truth as you could possibly get from what we were really trying to do.
+
+And again, if you looked at some of these companies -- and we talked to other customers of ours, big customers, and said, "Oh, these people that are kind of on the same level of you in the Fortune 50, they refused to pay." And when that kind of came up, they actually kind of changed their tone and said "Ooh, so we could actually get access to stuff that we're helping fund by having a commercial relationship with Synadia", for a period of time, two years before... On one conversation they called them the freeloaders. And he's like "It's not Bob and Vinny's pizza shop out of Venice. It's people on the Fortune 50 that have the means to pay. And this isn't a little internal developer tool. It's a production-level, critical service." They actually went "Oh yeah, we actually kind of like that idea."
+
+And so from a partner and ecosystem standpoint, those conversations - we had lots of them because we lost control of the narrative. The CNCF wanted to all of a sudden move things into the court of public opinion. That's totally their right. We didn't have a heads up though, so my phone started blowing up when that happened... But we kept business as usual, we kept doing commits, PRs... And we actually did a release the very next day, a full-blown server release, which is a really big deal. It takes a lot of effort for us to kind of do all of those.
+
+So we've been trying to signal to the right folks, looking at the right pieces of data that, "Nope, there's a disagreement here. We're going to try to work it out, but we're still committed to what we've been doing all along."
+
+**Derek Collison:** Yeah. You mentioned some of the emotional responses online, and I've read a few of them... One of which, which I'll summarize, because they use more flowery language than I'll allow myself... But maybe you have a chance to answer this particular outsider perspective, which if you're just watching from the sidelines - I can't remember if we'd hit Record yet or not; we started talking about armchair quarterbacks, but... You're just sitting in your armchair, quarterbacking. And what you see is Synadia brings NATS into the CNCF as an incubating project, it gets all the benefits -- even though it sounds like you were getting all the benefits that you maybe thought you would get, because Kubernetes maybe sucks a lot of the air out of the room... But gets all the benefits, the conferences, whatever; the clout that you get of being a CNCF project... And then hangs out there for a while, decides they want to relicense and take their ball and go home. And that makes people upset.
+
+**Derek Collison:** \[55:57\] Oh, yeah.
+
+**Jerod Santo:** What do you say to that sentiment? Why is that wrong, if it is?
+
+**Derek Collison:** The reaction is not wrong. What I have an issue with is the conflation of facts within the reporting. And so it's easy to draw that conclusion. So I don't get upset at people being very upset at us, and more so directed directly at me. The question is, is how much energy do we want to put into trying to make sure the facts are correct? The whole licensing and leaving the CNCF were conflated when they never should have been, because they were never conflated in my initial conversations with --
+
+**Jerod Santo:** Unrelated conversations.
+
+**Derek Collison:** Yeah. They just asked me and said, "Hey-- because this is a problem. We're seeing this and we don't know what to do about it yet, but we wanted to know." And I'm like "Yeah." And I said, we're starting to see said three-letter companies use our stuff and get a lot of value and contribute nothing back. And I'll be very transparent... Yes, contributions comes in a lot of ways, but for a company like Synadia, a commercial agreement and funds is the only one we really care about. Yes, we care about issues, and PRs, and all that other stuff, but just being very, very transparent, they don't keep the lights on. And again, Synadia has plenty of money and we have lots of revenue opportunities, and we're driving lots of those... But at the time, I was like "This to me feels wrong. It just doesn't feel correct, that the consumer bias is being driven to zero for companies that it makes sense for them." Because if you sell a piece of hardware or a service, it's kind of nice to have a guarantee that all the software you're going to use to augment that will always be free, and keeps your margins where they are. I just think as an ecosystem, as we're evolving, things are changing - the CNCF has changed a lot - that we still need to think through some of these and figure out "What do we really want as a global ecosystem around this?" And charity is not a business model. So the whole notion of "Oh, you can do donations", that's not a business model.
+
+And so what this lesson has taught me is that the biggest challenge, obviously, is in a company that develops OSS software, that wants to make a business off of that software, specifically as software; not necessarily as a service. Service changes the consumer bias. So if it's something physical or somebody's running something for you, your natural bias is "Oh, I have to pay for that." But if I give you a piece of software that we spent over 30-some million dollars developing over the last decade or so, you start to see this notion of entitlement. "Oh, I deserve this for free, and I don't have to give anything back." And we even have some people that just like bark orders at us, like "You need to do this in the system." And we're like "Okay, that's interesting. You're not a customer. We don't even know if you use it." And everyone has these problems. All the OSS projects have these problems. But for what we were looking at, I was like "That doesn't feel right to me." And I wanted to reach out and open a dialogue and have conversations that would hopefully spur some constructive dialogue. But just for the record, we never were abandoning open source. We were never going to blanket-relicense everything. That's just simply conflation of facts in the media, and then people reacting to them, and then that being amplified quite a bit over the social media channels.
+
+So that's the -- I have no problem with people being emotional about it, but the way it was presented and the fact that we not only lost control of the narrative, but then all of these articles never reached out to us or me at any point in time and said "Hey, what's your side of the story?", you know what I mean? And I was like "That's interesting..." I even have one now where someone pointed me to a YouTube video where the guy - just F-bombing me and trashing me left and right...
+
+**Adam Stacoviak:** Oh, my gosh...
+
+**Derek Collison:** \[59:52\] And I'm like "I don't know who this person is. I've never met them. They never talked to us." And I asked the whole team, I said "Has anyone interacted with this person?" They're like "Nope." They're just a YouTuber that --
+
+**Jerod Santo:** That's how you get the views, you know...
+
+**Derek Collison:** Yeah. And I get that, but at the same time, I kind of shake my head, I'm like "Okay..."
+
+**Jerod Santo:** Not helpful.
+
+**Derek Collison:** No, it's not. And again, if as an ecosystem we say to ourselves "We don't want open source projects that drive mission-critical functionality that are driven by a single company" - I don't agree with that, but if as a global ecosystem, we say that... But I want people to think really hard about what that really, really means.
+
+**Break**: \[01:00:32.15\]
+
+**Adam Stacoviak:** If you had your way, like if this went without a fight and you were able to do what you were trying to do, what were you really trying to do with the project? If things went your way, this fight didn't happen, and you could leave the hotel, so to speak, what would you have done? What was the goal?
+
+**Derek Collison:** So earlier on we were talking, I think, about - I don't know if we were being recorded or not - programming languages. So what I was looking at originally was if we created a foundation that had a Synadia representative, but also had representatives of all the major users of NATS, and customers of Synadia, in a 5013(C) or something like that, that had a commitment to AP2, open source, XYZ, that's probably what we would have done. But it was simply just for the betterment of NATS. And everyone that sat on the board, similar to kind of like the TOC - although the TOC is more of "Hey, we'll get involved if you need to move from sandbox, to incubated, to graduation, or if something's off the rails. Otherwise, we let projects manage themselves", this would be more of bringing in a lot of our top customers; the Walmarts of the world, things like that, to have a seat at this and be able to help direct where the project would go.
+
+**Adam Stacoviak:** And right now in the CNCF that's just not a full-on reality. That's more limiting, more governed by other folks...
+
+**Jerod Santo:** Yeah, couldn't you do that inside the CNCF?
+
+**Derek Collison:** We could, meaning... I mean, right now, we steward the project --
+
+**Jerod Santo:** But then you'd graduate, wouldn't you? I mean, you did that, and then they graduate you.
+
+**Derek Collison:** No. The graduation criteria that we failed - and I don't know if it's still the same; I have to check with that - was the code contributions in the server were not diversified. Too many were coming from Synadia.
+
+**Jerod Santo:** Okay, so it's not like governance, it's actually code contributions.
+
+**Derek Collison:** That was what tripped us up in terms of trying to graduate in 2020.
+
+**Jerod Santo:** That seems to be off. That doesn't make sense... I mean, it's more measurable, but if you established even a TOC of your own - or TSC, a technical steering committee - comprised of Synadia folk plus your largest customers... Right?
+
+**Derek Collison:** Yeah.
+
+**Jerod Santo:** ...and they didn't write the code... Like, who cares who writes the code if they're helping vote on the direction of NATS? I would think that would be a diverse open source project.
+
+**Derek Collison:** \[01:04:07.03\] You know, I would tend to agree, but again, I'm not trying to say "Oh, the CNCF's wrong", or try to speak on their behalf. I just am --
+
+**Jerod Santo:** And they aren't here to represent themselves, so...
+
+**Derek Collison:** Right, right. So I don't want to do that.
+
+**Jerod Santo:** ...we also have one side of the story. We're not journalists, we're just podcasters; we're here to talk. But worth saying, they're not here to defend their side of any of this, so...
+
+**Derek Collison:** Yeah. But what happened to us, with us not controlling the narrative, I want to be very sensitive to the fact that they do have differing opinions... And again, I'm extremely encouraged by our talks and how they're like "Okay, yeah, let's see if we can do better here", from a graduation flow potentially, but also -- I don't know if it's this notion of... I don't know the answer, and I don't think the CNCF knows the answer either, but for example Modular, I don't know if you saw, they just did a license change as well. They're getting more and more permissive, because they want broader adoption. And it's AP2, which is great.
+
+**Jerod Santo:** Who's this?
+
+**Derek Collison:** Modular. They do Mojo; that's Chris Lattner's...
+
+**Jerod Santo:** Okay. Modular's Mojo, okay.
+
+**Derek Collison:** Yeah, yeah. But what the interesting thing was is that they say "If you use it in production, you just have to let us use your logo in the use case." And I was like "That would be good. I would like that." Because we have so many companies -- I mean, if you go to NATS.io and look at who uses us, those people opt in. They send us and say "Hey, please put our logo on there." But there's so many more that use us and don't want that, or actually say "No, you can't use our logo whatsoever on that." So even something like that to kind of go "Oh, we can see Acme company is actually using that. And we know Acme company. That's really cool." That would help a company like Zenedia. That's that trust factor.
+
+**Jerod Santo:** Sure.
+
+**Derek Collison:** And so I shot that over to the folks at the Linux Foundation and just was like "Hey, would you guys ever consider something like this?" I don't think they will, but I'm trying to do my best and operate in good faith and saying "Okay, if we want to make this better, I'm going to tell you things from my perspective that could make it better for a project that looks like NATS, and companies like Synadia that are driving that, and we'll see how that goes." I don't expect anything materially to change radically, but at least they're open to the conversations now, and are starting to be receptive to "Okay, we see that there's some challenges here that we might be able to improve."
+
+**Jerod Santo:** So Modular, setting that up - effectively Apache 2 with an additional constraint on it. Isn't that then technically like a proprietary license now? Like, they've customized a license, and now it's not exactly Apache 2 anymore. It's like Apache 2 with this other thing, which is very similar to like what Meta did with Llama, which is like it's pretty much open source, except for this one... Now, their constraint was much more strict than "We want to put your logo on our website." It was like "If you have 10 billion users, you can't use it", or something like that. But...
+
+**Derek Collison:** I mean, you raise a good point. I'm not a lawyer and don't pretend to be, but I think in that notion, the license strictly governs what you can and can't do with the code, meaning the code itself. So for example, even within the CNCF the code is AP2, but the CNCF has other regulations around how you use the trademark, and how you can use the logo, and how you can do derivative products, and use the naming.
+
+**Jerod Santo:** Sure.
+
+**Derek Collison:** So I think what modular is doing is fun. It's outside of that pure "What can I do with the actual code?" But what they're saying is is that if you're using it in production, then you're giving us the right to use your logo. I think those are separate. But I understand it feels like it could be sticky, because the difference between a trademark versus a copyright versus a code license are all nuanced in how they kind of interplay, for sure.
+
+**Jerod Santo:** Right. Which reminds me of something else that I read - and this was just, again, armchair commentary, so I'm not even sure if this is true or not. But when you become a CNCF project, you typically will actually turn over IP to the CNCF, right? Like, the actual intellectual property becomes part of the trademark...
+
+**Derek Collison:** \[01:08:12.03\] Yeah. Trademark, logos, or whatever.
+
+**Jerod Santo:** \[unintelligible 01:08:11.01\] but yeah. And with NATS, did that happen?
+
+**Derek Collison:** So what happened with the trademark was kind of interesting... And again, it wasn't necessarily stated. And I tried to correct some of the facts and I got pitchforked even worse. But essentially, when we did the original kickoff for the CNCF at the Switch - this is my recollection, was that you only legally transfer those upon graduation. And I remember Google kind of being a little nervous that if Kubernetes started directly as a graduated project, that they would have to figure out all of that stuff with all of their lawyers, and X, Y, Z. And so that was in my head. Now, to the CNCF's -- to benefit them, I think they quickly changed that once Kubernetes graduated, and they said "Nope, as soon as you're in, you essentially have to transfer. Or at least once you move past sandbox." And we joined as an incubating project in 2018. So the rules had shifted from what my recollection from the original kickoff meeting was.
+
+But besides that, the thing that, to be honest with you, frustrated me - you know what I mean? ...just being transparent. But again, I like the conversations that we're having with the Linux Foundation and CNCF... But as we were trying to graduate and we had just joined, all of a sudden major leagues baseball came and said "Hey, we don't like your trademark, because it conflicts with the Nationals baseball team out of Washington."
+
+**Jerod Santo:** Oh...
+
+**Derek Collison:** And so I was like "Well, that kind of is silly." And anyone who's done patents from your listeners know that there's a little bit of a game to that as well. The whole patent, and trademark, and all this other stuff... But on the trademark side, I was like "I can't imagine someone being able to argue correctly to confuse a baseball team with a piece of software."
+
+**Jerod Santo:** Yeah. Different domains... I think that's spurious.
+
+**Derek Collison:** Yeah. So I approached the CNCF and said "Hey, we got this opposition. We want to fight. I asked our legal team. They think it's going to cost around this much." And the CNCF at the time was not willing to take that fight on, and suggested that we just change the trademark, and that if we do that and make it easier on them, that they would give us a keynote slot so that we could announce the name change. But I \[unintelligible 01:10:35.04\] to the name. I cared deeply about the name, and the logo, and you know...
+
+**Adam Stacoviak:** For obvious reasons.
+
+**Derek Collison:** Yeah, because remember, this had been gone to -- it was named, and we kind of started to have logo for almost eight years before we officially joined the CNCF. So it had a probably longer lifespan than people realized prior to the CNCF. And so we fought it, on our own dime, as a small startup that was only a year old. We had just gotten seed funding... And we successfully defended it. And so upon completion of the successful defense, the trademark was awarded to Synadia, with the deal.
+
+And so at the time, when I was talking to the CNCF, they were like "Well, we own the trademark." I go, "Well, technically, you don't." Now, this will get listeners frustrated and upset, and I get that, and I'm totally empathetic to why you would get upset. And I'm not asking for empathy by any stretch, but I was very frustrated that something that they said they cared so deeply about, they threw away in one email and said "Oh no, you guys figure it out." And then to say "Oh, but - congratulations, you won. Now give it to us." That also rubbed me the wrong way. But technically, what the CNCF was saying is "No, the commitment was you were going to give it to us." And so I definitely understand their position as well. And we are committed to that. And part of our agreement, we're going to transfer everything over, the lawyers are working out the legal docs, and things like that. But that was kind of my frustration, to be transparent. And then they never asked for it after we successfully won that. And so again, the way things were conflated and released out was, I think, very targeted, in terms of trying this in the court and public opinion.
+
+**Adam Stacoviak:** \[01:12:31.12\] I've got two questions, but one I'm going to pause on... When anyone, a project - maybe this one in particular, but any project - decides to take that project, donate it to the CNCF, and do all the things you did; incubate etc. Is there ever a point where you sign a specific agreement that's clear? So I feel like - Jerod, you've heard me say this before... Clarity and expectation. Like, if you have clarity on an agreement... It's the whole point of an agreement. It's less about binding you, but more like "Here's the line we both stand on. This is what we both agree on." Was there any clarity, or is there a \[unintelligible 01:13:05.28\] scenario where you're signing, and it's clear, and "This is what will happen, and there's a series of events"? Did you ever sign something?
+
+**Derek Collison:** We signed something to get into KubeCon in 2018, which had a line that says "If you have a project in the CNCF, you promise" -- I'm making some up, but I think it's online. "You promise to abide by all the rules for CNCF projects." That link -- I still can't even get into that link today. And so I think there's an opportunity here to improve that process.
+
+And also, I would imagine the listeners are very well aware of those emails that say "Our terms have changed" or "Terms of service have changed", and you get all these emails, and most of us don't even read those... I never saw any of that. I'm not saying they weren't sent, but I never saw any of that to say "Oh, hey. What you think, Derek, from the kickoff meeting, we changed it. And now it's this." So I think there's an opportunity to make that a lot more clear. But most people who are coming from an opposition to me say that's always been the case. That should have been crystal clear to you that you donate the trademarks, and X, Y, Z. And I understand that, because they see the CNCF as it has evolved. They didn't see it when it was first getting kicked off.
+
+And so I think they're totally right in their assertion that I should have known better, but I still think there's an opportunity to make it a lot clearer. Like, even now, I can't remember if it's as soon as you get part of the CNCF, or once you move past sandbox. I think if sandbox, you can take it back, if you don't get past sandbox. I think. But I could be wrong there. And so even with all this going on, it's still not necessarily clear what all of the rules of engagement are... But I just see that as an opportunity to improve for everyone, if that makes sense.
+
+**Adam Stacoviak:** Yeah. Well in that case there, if the link is broken or there's some sort of language that's compressed, like there's a whole other document that's compressed in like four words, essentially, based on what you've just said... And so those are usually addendums to an agreement. You know this, because you've probably signed tons of agreements in your life. Those are usually addendums, and that's usually compacted into a single product. This is the agreement I'm signing, and if there's terms beyond the agreement I'm signing...
+
+**Derek Collison:** They're attached.
+
+**Adam Stacoviak:** ...they're attached as an addendum, to kind of -- to help everyone safeguard themselves from scenarios potentially like this. Okay, so my second question is this - it sounds like y'all have an agreement, the blog post is out there... It sounds like kumbaya, to some degree; maybe not. Is any of this conversation you're having with us right now going to get you in trouble, more so?
+
+**Derek Collison:** \[01:15:53.27\] I don't know...
+
+**Adam Stacoviak:** Because you're being very clear.
+
+**Derek Collison:** Yeah, I'm trying to be very clear.
+
+**Adam Stacoviak:** For good reasons, but you're also being very clear. And it could be very clearly wrong, clear. Not wrong in terms of like truth, but more like "Hey, come on now. Don't say that in public, on a podcast that gets listened to by a lot of developers in our world..."
+
+**Derek Collison:** Yeah. I hope not, because we're entering into this in good faith, and I've tried to make sure to interject that I'm extremely encouraged by the discussions that we have committed to transfer the trademarks over. We actually did finalize early this morning, which I'm happy to talk about our actual direction in terms of whether we're going to do a fork, or what commitment we're going to do to the server, X, Y, Z... And so we're operating in good faith, and I would imagine they are as well, and it's a trust but verify, you know what I mean, type situation. But I'm very encouraged by them being open and receptive to some of the challenges that companies like Synadia or projects like NATS might have.
+
+And hopefully, there's some meeting of the minds around paths forward that might level the playing ground, if that makes sense.
+
+**Adam Stacoviak:** Yeah.
+
+**Derek Collison:** But I would imagine, you know, there's going to be some people that are going to pitchfork again, and that's totally their right... But that's how things went down, and that's kind of where we are right now... And I still feel deeply about the NATS project and where it's going. I mean, one of the other things that listeners might not understand is there's different, in my opinion, personas to projects that are kind of on maintenance mode... Meaning "We'll fix bugs, and if there's other stuff, but the features are pretty much complete." NATS and the server are definitely not in that camp. We've got over two years' worth of plans, and still the majority of our R&D budget goes into stuff that sits in the NATS.io ecosystem. Part of that open source project that sits inside the CNCF, from all of the clients, obviously the server, a lot of tools, Kubernetes, Helm charts, Systemd, all kinds of things like that. We're still probably spending 70 plus percent of our R&D budget just on that. And that hasn't changed, and we don't plan on slowing that down. It was "Hey, is there an opportunity for us to recoup some of those costs for people that care the most about some of those crazy enterprise features, massive scale, security, whatever?" And that's kind of how we started to go down the path of "Hey, what if we did a commercial fork, but may get BSL?" Again, it didn't land well because of the narrative, but our thought internally was that to us signaled the biggest commitment to OSS and to our customers trying to avoid vendor lock-in, or price gouging. But it was in a vacuum. And so once we got additional information, we realized that we probably would have to reevaluate that thought process.
+
+**Jerod Santo:** So how would the commercial fork with BSL be different than open core? I'm not understanding that.
+
+**Derek Collison:** We would be on -- BSL also is source available. It's not an OSI-compliant license.
+
+**Jerod Santo:** So NATS would be -- the entire NATS project?
+
+**Derek Collison:** Just the server. We would put all of the effort into that fork, which meant that once it was released, then the two-year window started. And so then that would flow into the AP2 codebase that's covered by the CNCF. That was the original thought process. Now, we're not going to do that...
+
+**Jerod Santo:** What's wrong with open core, though? It seems like you could just accomplish the same goals without relicensing. You just have -- you know, "You want your billion transaction messages per second, then we've got an add-on for you", or whatever it is.
+
+**Derek Collison:** Yeah, you're absolutely right, and what we've heard from the ecosystem is we like that model because that's what's familiar to us. What I try to point out to people is that commercial forks - it's totally up to Synadia, and maybe that's totally fine, and people are saying it's fine, but again, I don't think folks think all the way through some of these things sometimes... It's at our discretion what we put in there, and it's also our discretion that we don't have any commitment to put it back into the open source community whatsoever.
+
+\[01:20:05.08\] So let's say there's something, Jerod, that you really, really wanted, and you're an avid user and stuff, and Synadia decides "Oh, we're only going to put it in the enterprise version." You're stuck. Now, of course, you could try to do your own version of it in the OSS side... But at the time - and again, this was in a vacuum, and \[unintelligible 01:20:20.02\] additional information, but I was arguing hard that the BSL would signal a better commitment to saying "Everything will become AP2 after two years. No matter what we put in there, it will become that way."
+
+**Jerod Santo:** Right.
+
+**Derek Collison:** And so that was kind of our original thinking. We knew that was going to be a fork. The question on whether or not we could use the NATS name, who it was part of where we lived in the CNCF, or if it came out, and who technically owned that trademark... But we already knew -- we knew that that was what we could do, it's just if it's commercial we have no guarantees to either our customers from vendor lock-in and price gouging, or the OSS ecosystem. And so at the time, I was like "Ooh, I feel this is a better signaling mechanism." But we needed to control the narrative, meaning once we actually did decide - and that's the other thing that folks might not know... We hadn't finalized our decision when the can of worms got opened up, and so everyone's like "You've got to tell us what you're doing." I'm like "We don't even know what we're doing yet." We were just still having internal discussions on that, because it was a totally separate subject from the relationship with the CNCF.
+
+**Jerod Santo:** Sure. Well, do you know what you're going to do now, or are you still trying to figure it out?
+
+**Derek Collison:** We've been meeting every day since the whole debacle started, and so... We already had gone down a path of doing composition \[unintelligible 01:21:43.04\] So what I mean by that is a lot of customers came to us and said "Hey, we want to take control over authn and authz." We have a pretty rich authn and authz system, but they go "It's not enough. We need something else."
+
+And you could imagine a world where we said "Oh, we're going to do a Synadia server and we'll put all that stuff in there, somehow." Instead - and this is already done. This code was done probably a couple of years back. We did an auth callout mechanism, meaning the AP2 server, you can say "Hey, you're no longer in charge. Something over there is in charge." And then we, Synadia, could build lots of IP around for OAuth for authn, you know, X, Y, Z. OPA integrations, things like that. And so we already were going down that path, so our decision now is we will not do a BSL stuff. We know we have the right or reserve the right to do a commercial fork, but I think what we're going to do is we're going to post - and we can do this in the show notes - of everything we've committed to go into the AP2 server, without any delay. The BSL would have had a delay on there. And that we're going to expand our commercial offerings through composition, meaning you're running the server, but now all of a sudden, it's got security appliance, which we've been working on, and some really interesting stuff... Think of it like a \[unintelligible 01:23:00.21\] for NATS, plus a whole bunch of other stuff. And so we're going to go down that path.
+
+**Adam Stacoviak:** Well, I can tell you, there's so much nuance in this. History, nuance, good intentions...
+
+**Jerod Santo:** Emotions...
+
+**Adam Stacoviak:** ...not a lot of bad intention necessarily, but maybe some confusion, some lack of clarity... OMG. Please, fix that stuff.
+
+**Derek Collison:** Well, when a subject matter is very emotional, it puts your brain into fight or flight, which means that you have to make decisions very quickly, on imperfect information. That's how we all survive. That's why we're all here, because we ran away from the lion, instead of going "Oh, look, a cat. Let's pet it" type stuff. So I'm not being critical of folks doing that, because I understand why they're doing that... But hopefully they'll understand that they definitely had imperfect information, and not even close to a complete story. And it's not their job to understand all of that, but at some point hopefully they go "Oh, there's a lot more to this than we might have realized." We're not saying we're right, or wrong. I'm not doing that at all. I'm just saying it's worth a dialogue and there's a lot more pieces to the pie that folks might not know.
+
+**Adam Stacoviak:** \[01:24:14.00\] And to the credit, I think, of the potential pitchfork folks, or the folks in the crowd, not in the arena, is that it seems to be almost a - not literally a daily occurrence, but pretty frequently these days, something happening around a license change or a rug pull. It's so much that it's become a thing we've said here, "Rug pull not cool." That is our phrase. "Rug pull, not cool" is totally not cool. And that's an additional one, Jerod. We could put that one on a T-shirt. I think they're used to, and we're used to hearing this, and you think immediately negativity. You immediately think "This is nefarious. There's something going on here." And so it's almost as if there's a level of PTSD in the community. And I think as purveyors of and participants of, enthusiasts of open source software, they have a right to that, in a way.
+
+**Derek Collison:** Sure.
+
+**Adam Stacoviak:** But also, there's so much nuance here that they have to let some of the details percolate. And it's not always clear. So...
+
+**Derek Collison:** And you're right.
+
+**Adam Stacoviak:** Which is why we do this podcast. This is why we have you on this show. Like I said, this is years in the making, because we were going to have you on the podcast anyways... We've been talking on LinkedIn about getting you on the pod... And then this happened. I'm like "Well, Derek, can you talk? Are you able to pod?" I mean, I don't know if you can even pod. And you're like "Yeah, I'm cool." And so we had you on the pod. And that's why we do this, is so we give that clarity.
+
+**Derek Collison:** Yeah. And I knew the rug pull stuff, I've been listening, and I know Joran was talking about it, and others have been talking about it... And so I'm totally aware. And what my hope is is that while we're focusing on license change and rug pulls, that we up-level to have a dialogue of "Why is this happening? What are we as an ecosystem failing at, and what can we improve upon?"
+
+I said earlier on in the podcast, charity is not a business model. Well, until you're a company that is self-sustainable, and you're not just running on VC dollars, that's still kind of a charity model. So in other words, right now, our customers are funding a lot of the development, but the VCs who have funded Synadia are funding a lot of the development that everyone feels entitled to for free, right? And so my hope, and my sincere hope is that we can up-level and have a dialogue about "Why do we think this is happening? And are there things that we could potentially improve from an ecosystem perspective, that could do that?" And again, I'm not a fan of this, but one might be "No, we don't want any projects that run critical infrastructure to be from just a bunch of folks, or a single company." I don't think that's right, but that could be one answer. Or "Hey, if you're a big, well-known Fortune 100 company and you're using something in critical production, give back." Kind of like your rug pull thing, mine is "No pay, not okay." \[laughs\]
+
+**Jerod Santo:** Haah! I like that one, too.
+
+**Derek Collison:** Because again, production side is never zero. Someone has to put in time and money to produce this... But on the consumption side we're getting to everyone expects it to be for free. And that eventually has to be resolved somehow. Now, for projects like Kubernetes this resolved, because the Googles and Microsofts and Amazons of the world are motivated to put stuff in and they have other revenue streams that can offset the cost of participating in the software. That's totally fine. I'm not saying that that's wrong. I'm just wanting the dialogue of saying "Is that the only model that we really think will survive?" And I don't want a world that looks like that myself.
+
+**Adam Stacoviak:** \[01:27:54.18\] Yeah. Well, good for you for sticking up, I suppose, for yourself, or trying some attempt and having this clarity... I know the narrative hasn't always been in your favor, but I appreciate you coming on the show, even if it gets you in some trouble. I sure hope it doesn't.
+
+**Derek Collison:** Well, I'm sure it will, but... \[laughter\]
+
+**Adam Stacoviak:** ...just coming on here and being clear... And honestly, I think - not that it's here at the end, but all along, anyone from the CNCF who represents sandboxing, incubation, graduation and the clarity there, and maybe some of the things we brought up here, which was the agreement you signed and the terms that you alluded to, and the lack of clarity there... I'm happy and we're happy to have anybody on this show. I think it's good for everyone to have some of that clarity. So if you want to have something on this show and talk about the details, what it takes to join the CNCF, what those are, we would be more than happy to have that conversation in detail.
+
+**Derek Collison:** Yeah. And I think that's great. I really do. And like I said, my hope for all of this is that we can have an open dialogue about "Are there things that we can prove upon?" Outside of Synadia, outside of the CNCF, but just as a broader spectrum. And I think there's some interesting questions that should be discussed on "How we we want to go about that?" and what role the CNCF and the Linux Foundation plays, what role maintainers play... Because one of the other things that we got a lot of pushback from is I feel deeply that the lifeblood of a project is in the maintainers. It's - yes, people want to concentrate on license, what foundation are you part of, all of that other stuff. And I think they do play a role. But at the end of the day, it's the maintainers. And if all of a sudden the maintainers withdraw their support from a project, I think a lot of projects that have a tremendous amount of investment and are not just a simple weekend toy, if that makes sense... You know, saying "Oh, we'll just get some maintainers and they'll spend a weekend and then they'll know how to program, let's say, a NATS server." Right? I think that's maybe a little short-sighted.
+
+We have someone who's awesome internally that just got promoted, and is now running the server team... They graciously took that from me so that I can do other things. That person came in and was under our tutelage, and specifically under my tutelage, and it still took them about eight months or so. So we were paying them, and we were tutoring them... And most of the server's infrastructure was built by myself and someone named Ivan, and we were tutoring this person. And it still took them about eight months, almost a year to become like "Oh, I can go anywhere in the server code and be effective, be productive."
+
+And so when you get statements that "Oh, fork it. You guys go do your own thing and then we'll archive it, or try to find new maintainers" - again, I want folks to kind of really think through that and say "Is that the best outcome for the ecosystem?" I don't think it necessarily is, but someone from the CNCF might say "Yes, that's absolutely the way it's supposed to work." But what people are looking at to the CNCF around stability for projects, and longevity... You know, if the maintainers leave, there's going to be some level of disruption, for sure. It depends on the project, but it's not going to be zero. And in some cases, in our point of view - which people might disagree - NATS thrives when Synadia thrives. Because we're driving all of that stuff. At least that's my opinion. And again, not all projects look like that, and that's totally fine as well. But that's another one where I was like "That doesn't feel like it should be the only answer." That there could be a dialogue of "How do we better support these types of projects? What if they're not necessarily a good fit as a foundation evolves? What options are there for doing that?"
+
+And again, I think prior to the dust-up, but also going forward - and again, we'll release a statement next week, probably about whenever the podcast goes out, around our commitment... You know, most people might not understand, of our total R&D budget, 70% is going into open source. And it's funded by the VCs who fund Synadia, and it's funded by our customers. And the ones that can pay, that aren't paying, that are like "Give me, give me, give me. I'm entitled to this. Hey, get off your ass and change this thing for me", which we see every once in a while in GitHub issues... That's kind of when I'm like "Maybe we need to figure out a way to have them change their perspectives a little bit." I'm not saying I don't want hobbyists to be able to use our stuff for free, or embed it all over the place. My original tagline for NATS was "Connect everything", meaning it had to run everywhere, and was openly accessible. But again, it's that production problem, which is non-zero. It's like "How do we solve that problem for different levels of projects and open source?"
+
+**Adam Stacoviak:** For sure.
+
+**Derek Collison:** And I don't have all the answers, obviously, as you can tell, but I do believe --
+
+**Adam Stacoviak:** I mean, I think the important thing is you have a lot of thoughts, you have a lot of care. And I think that's the thing you have to really come back to, is you've got a lot of care for what you're building, what you've built, the community, the spirit of open source, your customers, your employees, your investors... I mean, there's so much things I'm sure that you're holding that we're not privy to that we can definitely see that you care. But it's cool having you on this show, talking about this stuff, and even more cool that you're a listener and you know --
+
+**Derek Collison:** Oh, yeah. I love you guys.
+
+**Adam Stacoviak:** ...rug pulls are not cool.
+
+**Derek Collison:** I know. And no pay, not okay. \[laughs\]
+
+**Adam Stacoviak:** That's right. No pay, not okay.
+
+**Jerod Santo:** We're gonna have a T-shirt with one on the front, and then that on the back. That's right. A little balance.
+
+**Adam Stacoviak:** That's right. Represent both sides. Get it together.
+
+**Jerod Santo:** That's right.
+
+**Adam Stacoviak:** Derek, thank you so much.
+
+**Jerod Santo:** Thanks, Derek.
+
+**Adam Stacoviak:** Nice to meet you.
+
+**Derek Collison:** I appreciate it. Thank you guys, and congrats on all the success. Really, really cool to see.
+
+**Adam Stacoviak:** Oh, man... Thanks, Derek.

Pivoting to Retool (Interview)_transcript.txt

@@ -0,0 +1,275 @@
+**Adam Stacoviak:** We're here with David Hsu, founder of Retool. Big fan... David, as you know -- I know a lot of our audience is probably users of Retool. They are aware of Retool because you've been a sponsor for many years... But we haven't had you on the podcast officially yet, so... Come on, let's get you on here, right? Let's talk about your past, your present, and your future. What do you think?
+
+**David Hsu:** Yeah, excited to be here. We've been a big fan of the Changelog. So...
+
+**Adam Stacoviak:** Big fan of y'all as well, obviously. I think we may have been like the earliest of days with Retool. I think you're around seven years old, roughly... I was in preparation for this conversation, I kind of wanted to go back in time a little bit; not so much to bring up like the old ways of things, but I kind of want to revisit that initial video that I found on your website. I think it's on your About page, if I recall correctly. It's somewhere. I don't know. It's somewhere on the Retool website. But I found the initial -- no, actually it's on your Y Combinator profile page. That's where I found it. And I'll link it up in the show notes, but it's the initial pitch that you gave... I think it's the initial pitch. So take us back there. We know what Retool is, for the most part today, but tell us about those humble beginnings in that video.
+
+**David Hsu:** Yeah. Wow, that was a while ago. I think that was around probably 2017 or so when we first got started. So when we first started Retool, my co-founder and I, we were both pretty lazy engineers, you could say. And we were tired of rebuilding internal tools all the time. Both of us had started a previous company before this... That didn't really go very far. We spent so much time building internal tools at that company, and we thought there's gotta be a better way. Why are we sort of reinventing the wheel every single time? Why are we populating tables, getting data from our database, figuring out how to build CRUD forms on top of that table...? It was just so much work to do something really pretty simple, just CRUD on top of our database. And so that's where the initial idea for Retool came from, is could we, as lazy engineers, build a better way, a faster way for engineers to build internal tools? So that's where the idea first came from.
+
+And so we went through YC in around 2017, and initially it was pretty slow. When we tried to explain Retool to people, people wouldn't understand what Retool was. "Why would I use something like this? I'm pretty happy writing code from scratch. Why would I use a product like Retool?" So it was pretty uphill for a long time. And I remember pitching VCs was very difficult, because we'd tell them we're really focused on engineers, and they would say "Why are you focused on engineers? Why don't you focus on democratizing programming? That feels way sexier." So it was really a -- happy to go more into detail than any one of those, but it was uphill for a long time.
+
+**Adam Stacoviak:** I think what's wild about hearing that is that -- the narrative roughly for this podcast, hopefully, is where you began, and that story arc of what got you here. And today, the here is a lot of agentic stuff... We'll obviously talk about the new AI platforming that you're doing nowadays, but... It's kind of interesting hearing that initial sentiment to build for engineers, and democratizing programming, basically, given where we're at today. Just want to throw that in there, but let's go back there...
+
+So it did seem like, even initially, if I'm being honest, it did seem like even when we were first part of that sponsor pool - I don't even know when it was. It certainly wasn't 2017. I want to say it was like 2018, maybe 2019, if I can recall correctly, that we began to work together... And I thought, "Okay, internal tools make sense." It seemed like it. But I recall also that you mentioned back at the YC day that you had another batch... They call them batches, for everybody listening. So when you go through YC, they're called batches. You've got, I think, summer and winter batches, and you go through this Y Combinator incubation process as a two-person, three-person, whatever number person team, and out the other end you hopefully come out with funding, or opportunity, and things like that. So they call those batches.
+
+And so when you were in that first batch, I recall you saying that Brex was also a part of that same batch, which is kind of funny, because you guys are both in this moment unicorns, right? But in those moments, you were just simply a hair on the potential of a unicorn, you know what I mean? You were not the futuristic thing that you are today.
+
+Go back to that moment where you were bringing on larger talent like that, even; larger opportunities like Brex. How did you go from building and solving these problems like that, to solving their problems? And why was it uphill?
+
+**David Hsu:** It was really cool, actually. So we were part of the winter 2017 batch, and I think it's been one of the best batches for, I think, YC over the past few years. So actually, Brex is a part of that, Rippling was also a part of that batch, we were part of that batch, and a few others, too. So it was really awesome seeing a lot of these companies that today are so successful, so early.
+
+\[08:20\] I remember when we first talked to Henrique and Pedro from Brex, they were just two guys starting a company... I think, actually, at the moment in the batch, I think they were actually building, I think, VR software, if I recall... And at the end of the batch, they're like "Oh, actually, we pivoted into working on corporate cards for startups." And I remember thinking, "Well, good luck with that. It doesn't seem like a market to me." But they've obviously done incredibly well.
+
+**Adam Stacoviak:** \[laughs\] That is hilarious.
+
+**David Hsu:** So it's really cool to see \[unintelligible 00:08:50.20\] succeed in that way. They're so talented as founders.
+
+**Adam Stacoviak:** Yeah. Was it part of that batch process they began -- you'd mentioned they pivoted during the batch process, which is really... Maybe you could speak to the details here, but it seems like some folks may go into Y Combinator with a pretty clear idea. And then that incubation process simply hyper-focuses them on that product development, the user research, the networking... Whatever it might be happening in there. And then some folks come in there with an idea that morphs dramatically. Can you speak to that? Was that at all Retool's scenario? Or were you pretty laser focused on your first one, and you just iterated, and that incubation process was really a baking process of the real product?
+
+**David Hsu:** Yeah, so my co-founder and I actually were working on a different idea at the very start of YC, actually. It was a P2P payments company based in the UK, and it turned out that a P2P payments company is not a particularly good business, especially when you are losing money every time someone sends the money. And so it was a negative margin business, and we were losing money pretty fast. But the lucky thing was, as part of building that product, we built a lot of internal tools. And you can imagine the kind of internal tools a P2P payment company would build. It's tools to go manage fraud, tools to go approve signups, tools to manage underwriting... And so we built a lot of internal tools. And so we were actually quite lucky that through that process, we learned that as lazy engineers, building internal tools was really pretty challenging. And it's an incredible amount of work required.
+
+And so that's actually where the idea for Retool came from, was entirely based on the fact that we had worked at a different startup prior to that, actually. And so throughout YC, we really refined that idea. And we were lucky that as part of YC there were many other companies part of the batch, such as Brex, such as Rippling, such as many other companies, too. And I think Brex was probably our first, maybe second, something like that; maybe our first five customer for us. And I remember talking to Pedro and Henrique about it, the founders of Brex, and as they were working on their credit card company, they had a very similar challenge to us with our P2P payments company, which is building a credit card company is very operationally complex. And so we told them about our story, and they were very interested in a product like Retool, because it could save them a lot of time, but more importantly, allow them to scale a lot faster.
+
+Today, it's hard to say what percent of time Brex spends on building internal software, but it's probably half of the surface area of Brex is internal facing. And so having a product like Retool that allows them to move a lot faster, adapt quickly to changes, launch new product lines a lot faster... I'd like to think that we played a small role in Brex'es success, which is really exciting to all of us here.
+
+**Adam Stacoviak:** Yeah. So that gives folks a context... I think most folks get this, but just to be super-clear - what would you consider an internal tool? What exactly is internal tooling?
+
+**David Hsu:** \[12:07\] Yeah, so a really surprising fact about internal tools is that more than half of the software in the world is internal facing. When people first hear that stat, they're generally pretty surprised, because --
+
+**Adam Stacoviak:** Yeah, I'm surprised...
+
+**David Hsu:** ...a lot of software engineers, we work in Silicon Valley. And Silicon Valley's job is to go build software for other people. So if you look at traditional Silicon Valley companies - Airbnb, Google, Netflix etc. these are all Silicon Valley software companies. Google is a software company that builds a software platform that connects advertisers and consumers. Same with Meta, same with Facebook, same with Airbnb, same with Uber. And so pretty much most software engineers at these Silicon Valley companies are building external facing software, the core product for Airbnb, Netflix.com, Google, the search product etc. But if you think about most companies in the world, most companies in the world actually aren't software companies. If you go look at the Fortune 500, for example, probably 10, 20 of them are software companies. 480 of them are not. 96% of the Fortune 500 are not software companies. But they actually hire a lot of software engineers. And pretty much all these software engineers do day in and day out is build internal tools.
+
+And so we think of internal tools as basically business applications that help you run your business, that's not being used by your end consumer, if you will. And it's actually the majority of the software in the world, actually, which is pretty surprising.
+
+**Adam Stacoviak:** So you have a lot of companies that are not traditionally software companies, that are writing more and more software internally. They're not buying their software from a vendor, they're not buying it from a box, they're not buying it from maybe even a SaaS, where that's kind of internal, but it's more of a service at that point, so it's not like software they're maintaining; it's just something they buy, for the most part, of some sort of subscription. It's a month, or a quarter, or a year. How did you go from this place where you're building this P2P payment company, learning that internal tools kind of suck to rebuild over and over and over, and learning that this whole entire world existed? Did you discover it during the process of Y Combinator? Because you mentioned the pivot - was that when you sort of like "Okay, this clearly has got negative margin. We've done this thing several times; let's not build these internal tools anymore. What if we build a platform that does it for other people?" Is that kind of how it worked out?
+
+**David Hsu:** Yeah, to be honest, it was mostly internal conviction driven. My co-founder and I were engineers, and to us, it was obvious a product like this should exist. And so to be honest, even before doing any market validation, we built it for ourselves. So as part of our previous company, actually, we had built an internal version of Retool actually ourselves, because we were just engineers that like building things. And when we were thinking about what to work on next, this became one of the major ideas. "What if we productize this?" And it was in talking to Brex, it was talking to some other customers that we realized there was a really big opportunity here.
+
+So another one of our -- maybe customer number three or number four is this company called Tyco Electronics. And Tyco gives you a sense of how large the market is... Which is when we first launched Retool, Tyco Electronics, a customer of ours now, a person named Alessio came inbound and said "Hey, I'm really interested in using Retool." And we were trying to understand what kind of internal software do you want to go and build, how much internal software do you actually build every day, every year... And we were shocked that a company like Tyco Electronics - I think it's fortune number 250, or something like that. So it's solidly in the middle of the Fortune 500. So an average Fortune 500 company. They every year spend around half a billion dollars building internal software, which to us was shocking. We thought, you know, we were working on Retool, and it was a cool thing for us engineers to use, a cool thing for us to use, to go build a few internal apps every now and then... But when we heard that Tyco spent half a billion dollars every year building internal software, we realized the opportunity here was ginormous. And specifically, we didn't believe the half billion dollar number at first, but if you do the math, it actually makes sense.
+
+\[16:28\] So Tyco as a company has, I think, around 120k, 130k employees. And of those 130k employees, around about 2% to 3% of them are software engineers. So they have around about 2,000, 3,000 software engineers. And if you do the math based on the average software engineering salary of let's say 200k per year, 200k multiplied by 2,000 software engineers actually comes out to $400 million every year that they're spending on software engineers. And Tyco has no external-facing software. Pretty much all they're building is internal tools, day in and day out. And so that gives you a sense of how large the market is just for one single company.
+
+So when we heard that, that was really pretty shocking to us, and we realized that "Oh, wow", when we were interviewing more and more customers, we realized that actually the majority of software in the world is internal facing. And the majority of internal software kind of looks the same, actually. It's pretty much just buttons, forms, tables, drop-downs on top of Postgres databases, on top of Salesforce, stuff like that. And so that's when we realized the opportunity for Retool really is rather large. And we were really excited about the mission of potentially changing how developers built this large class of software.
+
+**Adam Stacoviak:** When did you personally see a new piece of software be built on top of Retool, and you personally had this "Oh my gosh, what do we build?" moment? I mean, because if you were tired of the old way, which was from scratch essentially, and you could build on top of Retool with maybe an API key, sprinkle some code here and there for Postgres or some database calls, things like that... When did you personally feel like "Oh my gosh, this is the next big thing"?
+
+**David Hsu:** It was probably a few separate moments. One moment was actually when Alessio, that customer from TE, described Retool as similar to BI, but a much larger opportunity. And what he meant by that was if you think about BI, which is actually a parallel space to Retool, to internal software, today if you ask any engineer, you ask anyone really to go build you a BI dashboard - BI is business intelligence. So maybe you say, "Hey, give me a chart of signups over time", for example. If anyone asks me that, my first reaction is "Oh, I'll go use a BI tool." I'll go use Tableau, I'll use Metabase, I'll use Power BI; whatever it is, I'll go use that. Because when I use a product like that to build that chart, all I have to do is to go write a few lines of SQL, drag a chart on, and I'm done. It's so simple.
+
+I wouldn't even consider building that from scratch. To build a BI dashboard from scratch, I would probably have to go import some database drivers, I'd have to go build an authentication/authorization framework, I'd have to go find a bunch of frontend charting libraries, and try to debug them and see which one is best... Then I'll find out "Oh, it's a little bit slow, so I'm going to go build a caching layer..." It's so much work to go build BI dashboards from scratch, and no one does it today.
+
+And so when Alessio said "Hey, actually, this is just like BI, except there's 10x, honestly, more use cases than BI", that's when the idea really popped up in our heads, "Wow, this is a product that is like Tableau, but 5, 10 times larger, actually, in terms of opportunity." And the fact that half of all software is internal means that one day half of all software could be built in this way, which is really exciting. So that was one moment.
+
+\[20:01\] Another moment was when another customer actually made an analogy to AWS, which - she said that they had gone all in, this company had gone all in on AWS. This was maybe in 2016, or something like that, so a little bit earlier. And the way that she viewed it was AWS saved her a lot of time, and her company a lot of time... Which is instead of building your own data center, trying to figure out how to go manage racks of servers, when a power supply blows out you have to go figure out how to go replace the power supply, when a hard drive spindle fails, you have to go replace the hard drive... Now all they do - now we all do today - is use AWS. We press a few buttons, and servers magically appear. And if we want to go launch in a new region, we'll press a few more buttons, and we'll launch it in another region.
+
+And all this commoditized, or this undifferentiated heavy lifting I think is what AWS calls it, has been commoditized. And we as Retool and that company don't have to worry about having to go worry about all this heavylifting anymore. We just go move a lot faster actually, because AWS is handling all the stuff for us.
+
+And the analogy she gave is Retool is kind of like AWS, but the AWS of the application layer, which is there's so much stuff in the application layer that is so bespoke right now that you have to build from scratch. So when you're building an internal tool today, for example, you have to worry about "What table library should I go and use? Okay, I'm going to go use this table library. Oh, that's not quite right. Okay, let me go find this other table library. Oh, this table library comes with virtualization, so when I load big data into it, it's actually quite slow. Okay, now that I have that built... Okay, now I have to go worry about authentication, authorization. Who can actually go use my application? How do I go hook it with single sign on, for example? Oh, now I need audit logs, because I need to go verify -- when something bad happens, I want to know who did what, for example." There's so much crap that you have to deal with just to get a simple CRUD dashboard going. And it's so much work. And so the idea that Retool can sort of commoditize all that away, such that the engineer can focus truly on what is unique or differentiated about what they're building I think is so powerful. So that was maybe a second moment, too.
+
+So I think it's a few moments like this when you see customer use cases, the customers go build things that you could not have imagined... And they actually go pitch a product back to you, in ways that you couldn't have imagined either... That's, I think, what was really validating, for us to realize "Wow, this is a really large opportunity for us to change how developers really go build software."
+
+**Adam Stacoviak:** One thing you did say, though, was that it was a... Remind me if I'm -- I'm going to paraphrase... I think you said that it was uphill, was the way you described it. Like, to describe how it was to raise awareness for Retool, to preach the mantra of "Don't build your own internal tools, use a platform like Retool to make it 10x faster", et cetera. How long did you spend, or are you still in that phase where it's still kind of hard to -- it's still sort of uphill in a way, or hard to sell Retool and to help people understand that this is a real problem, or this is a real solution you all solve very easily.
+
+**David Hsu:** I would say it's gotten a lot easier recently, especially with AI, which is helpful. But in the early days, I think many people just weren't sure. "Should I go trust a new company like Retool?" Today, Retool has tens of thousands of customers at this point, and so...
+
+**Adam Stacoviak:** Trust is too easy now, yeah.
+
+**David Hsu:** Yeah. Companies ranging from NBC, all the way to the US Army, all the way to Amazon... Companies like that all trust Retool now, so that helps quite a bit. So that's great. But the AI has been a real tailwind for us too, because now a lot of people are really looking at "Can I go adopt new ways of building software?" The ways that we used to build software are no longer cutting it. Are there better ways? So that's been quite helpful, too. So it's been less uphill recently, but the first few years really was a lot of educating the market, and a lot of explaining to people why this is better than building software from scratch.
+
+**Adam Stacoviak:** \[24:17\] How has Retool changed over the years? Has it always just been like internal tools, and just incrementally getting better? Have you added more user interface styles? Where has been the advancement of the platform over the years?
+
+**David Hsu:** Yeah, the advancements we're most proud of really are around AI over the last three to six months, honestly. So just a few months ago, a month and a half ago, we launched this product called Retool Agents, which really builds on top of a lot of the platform that we've built. It has enabled customers to go do some pretty incredible things.
+
+So in particular, to use Retool to go build a frontend application, basically you can build the frontend, but then you write a bunch of queries to connect it to your databases, to your APIs, stuff like that. And one idea we had around six months ago - and this was just really an idea at that point in time - was "What if we could expose all the tools that you have built in your company in Retool, and expose them to an LLM?" That could be really powerful, because then it could allow the LLM to go do some pretty incredible things in your company.
+
+The thesis came from -- the main idea here basically is we think that if you look at the power of LLMs today, LLMs are really quite smart. I think they've conclusively passed the Turing test. So I was a philosophy major in college, and back in the day we would debate "How do you know if an AI is smart? What does it mean for an AI to be smart?" And I would say a good amount of philosophers thought, and to this point still believe, that if an AI could pass the Turing test, that we would consider that smart or intelligent. And the Turing test is basically, if you're talking to an AI, can you distinguish the AI from a human? And today with LLMs, you really can't. If you talk to any state-of-the-art model, whether from OpenAI, from Arthropic, DeepSeek, Ollama, honestly, they sound more human than humans actually a lot of the time, actually. And so they've conclusively passed -- at this point, I think we can all say that AIs sound just as human, if not more human actually than even humans do. So that's great.
+
+But when we look at how much value are people getting out of these LLMs today, it's really pretty limited, and it's really just one use case. It's really all just around chat. And it's kind of surprising that if the LLM is able to do everything, or can basically pass as human, why is it that we've invested so much energy and money as an industry into AI, and yet there's not that much ROI today.
+
+So to give you an example, I think over the past few years the industry collectively has invested somewhere between a trillion and $2 trillion into AI, which is a lot of money. A lot of that has gone to GPUs, and to training models etc. But if you sum up all AI revenue in the market today, I think it's maybe around 20, $25 billion. So that's like a 2.5% ROI, based on how much money was invested, which is pretty surprising.
+
+**Adam Stacoviak:** That's interesting to think about, yeah.
+
+**David Hsu:** Yeah, it's kind of shocking, actually. And our thesis was, it's not actually about how smart the models are. The models are plenty smart. It's actually about what the models can actually go and do, which is when you use chat, the model can't really do anything.
+
+\[27:56\] I use chat frequently, for example, to go help me write things in Google Docs, for example... What I'm doing is I'm just conversing within my ChatGPT app, and I have to copy and paste it into the Google Docs... Or if I'm just building a slide deck, for example, I have to go copy-paste the data into a slide deck... And then, "Oh, actually, it's not quite right, so let me copy-paste it back again..." It's so much work going back and forth. And so the idea is LLMs today are plenty smart. Really, we just have to give it the arms and legs for it to actually go do things. And this is, I think, what's so powerful about something like Cursor, about the idea like Cursor, which is theoretically, if Cursor didn't exist, you could theoretically do - or Copilot didn't exist - you could do what Copilot does with just ChatGPT. You could say "Hey, I'm going to go copy-paste this code from my IDE, paste it into ChatGPT, and have it write it, and then copy-paste it back. Oh, but it doesn't work. Okay, let me copy-paste the error message. Okay, let me copy back again."
+
+And theoretically, that would work, but you can see how Cursor or Copilot is so much better. The fact that you could expose the IDE capabilities to an LLM and actually allow the LLM to go do everything makes your life 10 or 100 times better, actually.
+
+And yet, if you look at how non developers, i.e. most of the world is using ChatGPT today, are using LLMs today, that's what they're doing. They're kind of pasting back and forth, actually. And so our idea was, can we go expose an LLM, and allow it to actually do things in your business systems? So it could actually go run queries in your Postgres database, it can actually go refund customers in Stripe, it can actually go modify Salesforce, it can actually go send emails on your behalf, for example. If we could do that, that would almost be like a Cursor for everything, basically.
+
+And so we shipped that a month and a half ago, and the reception has been really positive. So we can talk more about that, but that's probably what we're most proud of shipping, I think, over the last few years, is this product, Retool Agents, which - yeah, it's really quite powerful.
+
+**Adam Stacoviak:** It is powerful. And it kind of goes back to the initial framing, which was the investor had said "Why would you focus on developers", or engineers I think he may have said; assuming gender, but... Yeah, I think you get in this place now where AI is kind of everywhere, it's table stakes. Like, can you really be a modern day software platform like Retool and not have AI-focused features? I think the answer is hard no, right? But you didn't just simply add AI, like you had said, with chat... I'm really curious how you do it in a way -- because it's non-predictive, right? Like, when we do generative AI, when we use LLMs, we can predict to some degree what it's going to do, but it's non-deterministic, right? Like, what comes out is non-deterministic. How do you unleash that on what I would potentially consider the crown jewels of all crown jewels, which is the internal software of Fortune 500 companies? And if you've got even a small majority of Fortune 500 companies customers, then when you unleash that kind of thing to that kind of audience, there's got to be some concern. How do you do -- how do you roll something like that out that's non-deterministic in a way that covers your butt? ...is the easiest way to ask that question.
+
+**David Hsu:** Yeah, so this is the power behind this form of software called agentic software. This is basically what Cursor is, and many other products are, which is the core idea of an agent... Today, the most popular architecture for agents is called a React style agent. And so a React style agent is basically a reasoning and an acting agent. And so what it does is that the LLM reasons, and then it acts. And the way it acts is you give it tools. So you say "Hey, you can go search the web", for example. Or you can -- in Cursor's case, "You can go modify some code" or "You can go run some code", for example.
+
+\[32:10\] There's this Open AI product called Operator, which is a good example of this, which is - Operator is basically an agent with one tool, which is it can go browse the web. And it's pretty cool, actually. You can tell it to go do things. So you can say "Hey, go buy me a pair of shoes, in this particular size, this particular color", and you can watch it go do things. It says, "Okay, well, let me think about what to do there. Okay, well, I'm going to go visit this website. I'm going to go search up shoes, this color, this size. Click this link. Okay, great. Oh, that's the wrong size. Okay, let me go back, actually. Let me go click this other link." And so you can sort of watch it reason and act, basically.
+
+And what we've discovered is, to what you're saying, something like that, Open AI Operator, is pretty effective in a consumer context. So if you're trying to buy a pair of shoes, buy a T-shirt, for example, it could be good at that. But in an enterprise context, it's actually really, really bad at it. And the reason is, the one tool that it has, web browsing, is such a non-specific tool. For example, if you tell Operator - you can try this yourself, actually - to go onboard an employee that just joined yesterday, for example. What it's going to do - it's gonna go and search "How do I go onboard an employee that joined yesterday?" And it has to go find like a wikiHow article... And it's like "Okay, great. Well, a wikiHow article says to go onboard an employee that joined yesterday is I go do these things." And that's not what you want to do at all. If you work in a company like Amazon, for example, there is one right way of onboarding an employee, which is you have to go add them to Workday, you have to go send them a laptop via Amazon Logistics, you have to go do this, you have to go do this, you have to go do this, you have to call these very particular API endpoints. It's not "Let's go browse the web, and go google "How do I onboard an employee?"
+
+And so that, I think, gets at the power of an agent, which is with this kind of architecture of reasoning and acting, what you can do is you can actually outsource the reasoning to the LLM, but the acting needs to be deterministic. And that, I think, is the power of something like Retool Agents, which is you can go build these deterministic tools inside of Retool.
+
+In fact, our customers have built millions, if not tens of millions of tools for agents to go and use. And so for example, for a company like Amazon, they've built tools to say "Hey, here's how you go add someone to Workday... Is you call the Workday API endpoint with the first name, the last name, the SSN, and the address", for example. Great. Okay. That's the first thing you've got to go and do. "Here's how you send a laptop. You call this custom internal API that we've built", maybe a node or wherever else, "and it expects these parameters. It expects the first name, last name, address, and computer model, for example, as well as the start date, so we know how fast to ship that laptop, for example."
+
+And so once you've formalized those tools, then what you tell the LLM to do is you say "Hey, LLM, don't just go browse the web and do whatever you want. Instead, you only get these five tools to go onboard this employee." And then the AI is so much more accurate. And so in our testing, our AI is so much better when you give it these very specific tools.
+
+And that, I think, is the power of a platform that allows you to outsource the non-determinism to the AI... But really, you want to minimize that. Really, you want to maximize determinism. And to do that, you need a sort of tool building platform, if you will.
+
+**Break**: \[35:37\]
+
+**Adam Stacoviak:** So you have the idea of an agent, which is some sort of demarcated name and role. So Add New Employee, let's just say. That's the agent's name, and it has a task of X, but it's only got limited tooling... So it's kind of like saying "Get to Rome." Everybody says this old adage, "How do you get to Rome?" "Well, all roads lead to Rome", right? But what if only five roads led to Rome, and those are the only five roads you had access to. You can only go to Rome. You cannot go to -- I don't know, where else was popular back in Roman days? Some other city, right? How do you get there? Well, you can't, because you don't have access to those roads. You can't go there. You can only go to Rome.
+
+**David Hsu:** Yeah, totally. And so we've discovered it's just so much more effective that way. And if you look at actually, for example -- this is why Cursor is so effective too, actually... Is cursor, when you go use agent mode, it only has a few things it can do. And so it's not just endlessly googling, and "Well, let's google for this, let's google for that, let's see what happens." It says, "Well, I've only got a few things. I've got to modify the code, and I've got to run it. So let's see -- I've gotta go modify the code and run it. Oh, I got this error. Okay. Well, let me go fix that error." So by constraining the AI, it becomes substantially more effective. But also by giving it tools to actually go do things, instead of telling you to do things...
+
+**Adam Stacoviak:** Nice...
+
+**David Hsu:** Which is actually, if you look at what people are using ChatGPT for, it's actually kind of crazy how inefficient it is... Which - today, the way I use ChatGPT is I copy and paste so much stuff into and out of ChatGPT. And to some extent, it's almost like I've become the API for ChatGPT, right? I've sort of become this manual worker is doing what ChatGPT tells me to do. Whereas if you could glue ChatGPT to actually doing the work and give it the tools, the arms and legs to actually go do the work, you could just sit back and drink a beer, or something. It'd be so much more effective.
+
+**Adam Stacoviak:** For sure. What you're reminding me of - and maybe this is something you're already onto - is as a business owner, I've been more exposed to the idea of systems, right? I will tire out, I will procrastinate as an individual, just me as Adam. Unless I can build some systems, I cannot scale my enterprise, or lack thereof, let's just say. We're not a big outfit. But in order to scale beyond the one, which is me, to do the same job I do to begin to hand off parts of it to somebody, I can't do that unless I build some version of systems. And we use a lot of Notion... We don't really use a lot of Retool, honestly. We don't have a lot of internal software, but like a lot of stuff we're doing in Notion is in those kinds of ways. But what you're making me think of is really how systems is what's crucial, and they've already built -- so if you're a deep or even shallow-ish customer of Retool, you've got a couple internal tools, right? And if you can systematize those things, the agent... I imagine agents are similar to systems, or systematizing these things... It's a form of automation. Now you have the same software you've built out for yourself in your enterprise doing certain things, whether it's adding an employee, or updating a BI like dashboard that's built in Retool, et cetera. You need to -- you essentially need to trust what's happening there. How do you then expose not just this agent world, but this ability to create systems for these enterprises? Because I've got to imagine if you've got one or multiple internal tools, the next thing you want to think about is "How can I systematize it around these tools?" Is that the way they've also been thinking about this? Has this been sort of like "Build an agent, automate some things"? How do you sell this idea to these folks?
+
+**David Hsu:** \[42:21\] Yeah, so the great thing is that for a lot of existing Retool customers, they've built millions of tools already. Every API you have, every database table you have, to some extent, every single query you've written is a tool, that you can now expose to an LLM in this controlled way, to actually go and do things. But I think this is where the industry really is going, is how can we compose or build the tools such that we can expose them to LLMs, and have LLMs decide which ones to use, to actually go automate more and more and more work.
+
+I think tool building is going to be a critical skill, if you will. And thinking about the level of abstraction at which you build tools - is it very high level, is it very low level - is going to be really interesting. We see our customers today, for example, those that have built thousands of Retool apps, for example, and actually composing agents, in fact. And so an agent scales up to probably a dozen tools or so... But what you can do is you can actually build agents that have other agents as tools. And so maybe an agent has 12 other agents as tools that it could call, and it says "Hey, if it's this kind of problem, I'll call this agent. And if this agent says, "Okay, well, if it's this particular subtype of problem, I'll call this agent", and that agent has itself 12 tools that it actually uses. So I think this kind of Mandelbrot-like way of scaling complexity is really going to be pretty -- this is, I think, going to be how AI automates more and more labor over the next two, three years.
+
+**Adam Stacoviak:** Mm-hm. How long have you been working on this agentic platform? What do you call it - do you just call it Retool Agents? What's the proper nomenclature I should use to reference it?
+
+**David Hsu:** Yeah, Retool Agents. And the way we pitch it is we basically say "Hey, have you heard of Cursor? Great, you've heard of Cursor. Well, the reason why Cursor is so effective is because it's an LLM with tools that can actually go and do things." So it can sort of reason, and actually act. And that is such a powerful loop that you can give an LLM.
+
+**Adam Stacoviak:** The reasoning and the acting, yeah.
+
+**David Hsu:** And if you could do that for all other aspects of your business, such that it's not only developers getting more effective, it's your HR team getting more effective. It's your finance team getting more effective. That is so powerful. But actually, the people building all these agents are still developers, actually, which is really cool. Because in order for an agent like this to work, you have to build the tools, and the tools have to be built by developers. So I think the importance of developers, actually, if anything, has gotten more important over the last six, nine months for us.
+
+**Adam Stacoviak:** Can you give me a glimpse into the kinds of agents being built? Can you expose any one customer, or any particular specific scenario that gives us some clarity?
+
+**David Hsu:** Yeah, sure. So one pretty cool use case is we have a customer who has actually built an agent to go replace management, actually, inside their company, which is really cool. So they have a sales team, actually, that is doing a lot of cold calling, and for them, it's important for the salespeople to be getting feedback in a timely manner on "Hey, you could have done better here. Here's where you did really well, actually; keep on doing that." And you can think of it -- it's pretty interesting, because as a sales manager, it's actually hard to go and listen to all the calls your salespeople are doing. Because if your salespeople are working, let's say 40 hours a week, it's actually hard for you as a sales manager to be supervising and to be listening to 8 times 40, so 320 hours of calls happening every week.
+
+\[46:00\] And so what this customer has actually done is they've actually built an agent, actually, to go manage this team of salespeople, where the agent actually, after every call finishes, goes and analyzes that call, and then thinks about it and gives feedback, actually, on "Hey, this could have gone better, but you did a really good job here. Keep on doing this", etc. And then actually, in future calls, it can actually provide, in context, while the salesperson has the call, it can actually give them live feedback, tell them what to say, actually.
+
+So it's really pretty cool seeing the power and what you can do, actually. And I think a lot of jobs, actually, if you think about sort of an agent doing this, the agent's actually probably doing a better job than the manager was before, actually... Because the LLM is able to take in a lot more information, actually, than an ordinary human is, that I am actually able to. And so I think really leaning into what LLMs are good at, which is consuming information, summarizing information... But that's actually a large percentage of knowledge work, actually; a large percentage of my job, actually, is doing those kinds of tasks.
+
+**Adam Stacoviak:** Yeah.
+
+**David Hsu:** And so I think in the future we're going to see a lot of -- I think the labor market is really going to change quite a bit, when agents become more and more powerful, and have more tools, and actually go do things in your business, like giving feedback, writing performance reviews, stuff like that.
+
+**Adam Stacoviak:** I think some of these things -- I'm thinking this case in particular... There's some things that I'm personally okay with and ready to hand off to something that isn't... Let's just say isn't me. "Delegate" is the easiest word, that everybody uses. It's the actual proper word. And I really don't care necessarily how that task gets done; whether it's procrastination and it just solves itself, or somebody else does it, or literally now some version of a robot or machine or AI can do it for me. I don't really have that problem. And one case in particular, I think, which you just mentioned, which is a great example, is where a salesperson is attempting to do their job, which is to -- you know, sometimes it's selling, but sometimes it's just simply warming the waters of an opportunity. It's just answering questions, providing feedback, providing knowledge... And they need somebody else's support, whether it's a manager, a technical salesperson, somebody more knowledgeable, with a bit more experience or depth... And if I -- not so much if I can remove them, but if I can speed up my personal feedback loop if I'm in that role, attempting to warm the waters of an opportunity, not just simply sell, but... I would want those answers to be fast, one. And two, I think my manager over time might get over my third or fourth request of "How do we do this?" Or "What do we do there?" That person will just grow tired, of me, potentially even like individually me... And then my question, or just generally answering questions like that of people like me. Even if it's their job, they may just be like "This is not -- it's not a good use of my time the company's paying me for", for example. And so in those cases, I'm kind of happy to let those kinds of things go to the agents.
+
+You alluded to this -- and I think we all sort of see it coming, in a way, is this major shift in the labor markets... Which we're just not able to truly predict yet. And here I am, sitting with someone who may be, in the levers of time, David Hsu, co-founder/CEO, Retool, all the things you do, could be responsible for the labor market change. How do you feel about that? Has it always been just whatever it takes to innovate? And how do you sit at that wheel, so to speak?
+
+**David Hsu:** \[49:53\] You know, I think there's going to be an enormous change in the labor market, especially for knowledge workers, myself included, over the next few years, because like we're just talking about, honestly, LLMs are really quite smart at this point, and in many cases can do a better job than I can, actually, at consuming information, summarizing information, and even making decisions at times. Maybe they're less biased than I am, for example... So I think there's so much potential there.
+
+And you add that to what you said, which is - I think ultimately we're outcome-oriented. I think businesses, business owners, we're outcome-oriented. We don't actually care; even all employees, I think we're all... As long as the work gets done, whether it's done by a human or a robot actually doesn't really matter that much, if it's being done to a high quality. And oftentimes, with robots it can be done to a higher quality, faster than humans. So that I think is, from a philosophical perspective, a little bit unsettling, but the way I think about it - and perhaps this gets to your original question, is I think it really is a... I think in the short term it's going to be painful, in the sense that we're going to have to adapt. I think the labor market will have to adapt, a lot of humans will have to go and adapt... But those that really embrace AI -- I don't think it's something you can really hide from at this point. Those that embrace AI, the builders that are able to go build the building blocks such that their companies can really leverage AI, I think those are the ones that are going to do really well over the next few years.
+
+And we see so many examples of this in our customer base today, where someone takes action, and they are able to go write some SQL queries, they're able to go build some building blocks, build some tools for LLMs, and they go automate a whole process, and they get promoted three times in a few years... I think those are the people, the sort of action-oriented people that are hands-on keyboard-able to go do things. I think there's going to be a premium to that in the market over the next few years. But I also think in the long term, it will lead to a lot more abundance. I think Democrats have been talking a lot about this in California, around an abundance agenda... But that's ultimately what I think AI will do for us, is there will be enormous deflation, where the cost of providing a lot of services will go down. And that will allow humans to go afford a lot more of those services, which would be really cool.
+
+So I think overall, it's going to shake out really well overall for humanity in the 5, 10, 20-year timeframe, but there's going to be a lot of, I think, pain and gain for those that can leverage AI. There's going to be a lot of gain for those that can really leverage AI in the next few years, and a lot of pain for those that can't, and end up having to retool themselves, or to move to a different job family. So yeah, it's going to be a pretty somber few years.
+
+**Adam Stacoviak:** Interesting. Yeah, certainly a possibility. I can see a lot of things shifting, but even -- I kind of come back... I don't think this example multiplies well, potentially. And I could be wrong... In the case back to the manager giving feedback directly to people - over time, I'm going to get slower at it, potentially; I'm not going to be the best at it. I may give some attitude... I may not give the best response... And that really is -- the context and the intelligence of the response can still come from someone like me, feeding, essentially, the LLM context, so that a salesperson, or an AE, or however you want to categorize that person or title them, they're able to leverage my knowledge, but without me having to be present to literally handspoon, handfeed it to them, on the case by case basis. To me, that seems -- I'm cool with that. I'm cool with that.
+
+\[53:54\] Let's dig into the specifics of like -- I'm thinking there's probably a certain situation or certain framework, or parts to internal tools that every company needs of some shape or form. Is there a stampable blueprint, so to speak? "Your company, you likely have these kinds of needs, for these kinds of internal tools. You should come to Retool and build them. And then on top of all that, we're going to give you smart AI to react and take action and reason on all this stuff on top of this platform." How close is that to being reality, where you just have like "Here, take this, and it's done", where you don't have to individually build each one? You can sort of like templatize what it is to have an internal tool set for your business.
+
+**David Hsu:** Yeah, a lot of that is happening today, where especially for smaller, medium-sized companies, a lot of our templates work off the shelf... Which is "How do you do customer support", for example? How do you automate that? Well, you have to plug in your intercom database, plus your Salesforce, then you give that to an LLM... And it's actually pretty capable of answering. You plug in your docs page maybe, maybe you vectorize your previous support tickets... And it's very good at answering support tickets, actually, already. That's a template that we have on our website, that you can check out for free, yourself. And so that's going pretty well, I would say, especially for smaller or medium-sized companies. For the largest companies, like the Amazons or the US armies of the world, there it gets tougher, because their templates don't necessarily work. Because Amazon is not going to just adopt to how someone else does something.
+
+**Adam Stacoviak:** For sure.
+
+**David Hsu:** They have a very specific way, to go back to employee onboarding, for example. They have a very specific way of "Here's how our workday is set up." It has these columns. And these are the required fields \[unintelligible 00:55:51.25\] workday, for example. And there's no way that anyone else can really tell them how to do that. It's sort of proprietary to Amazon, if you will. So I think that the -- this is why I think the developer skillset is so important... But I think that the developer plus the line of business or understanding the requirement from the line of business, the context is going to be ever more important, if you will... Which is, I think, the sort of skill of translating a requirement into code. That, I think, is \[unintelligible 00:56:23.21\] commoditized. And we're starting to see that already, with app gen products like ours, or AI IDEs like Cursor or Copilot, for example. That literal act of coding is less interesting nowadays. But the ability to go interview stakeholders, to go understand "Okay, well, this is the API in particular I should use", that skill, I think, is going to be really, really important, at least for the next few years, I think. So yeah, that's the most important skillset, we believe.
+
+**Adam Stacoviak:** I think so, too. It's interesting... I've seen the templates, I just didn't think they were that expressive -- I didn't think they would be expressive to an enterprise, that's for sure... But when you look at your audience, you have the idea of I think a lot of startups or a lot of upstarts that start their scale-ups now, so to speak. Like, Brex is no longer a startup, rightt? They're as old as you are, so they're knee-deep and they're good to go... But they're still using Retool. There's no template to say "Out the other end, Brex." But if I'm a particular kind of company, for example a grocery store, maybe there's particular things that a grocery store might need. How frequently does a grocery store open up? Probably not too frequently. So that's probably a bad example. But let's say a barber, or a nail salon, or a massage therapy, or let's say a golf studio. Or a range where you can go hit golf balls. There's bespoke, different, unique businesses, but they all have employees, they all have probably payroll, they all have unique things that each and every business has. Has it ever been wise? Or does your template sort of cover that chasm, so to speak? So it says "You're starting a new business, start on Retool. Don't build these tools yourself. You're probably going to use these four or five web services... Let us complement that with Retool." Is that something you do?
+
+**David Hsu:** \[58:20\] Yeah, ish. But oftentimes no, actually. So the way we think about it is, there are times when internal tools are really quite valuable. And I think for large companies, for example Amazon, internal tools are -- this is why Amazon is such a successful company, is due to the internal tools they've built. And so it's so important for you to -- same with Brex, for example. Same with Ramp, another one of our customers. Same with Stripe, for example. Internal tools are the lifeblood of their business. To be honest, if you're a barbershop, you probably don't need the world's best internal tools. You can probably go use some barbershop management software, a CRM built specifically for barbershops, and you're probably fine, honestly.
+
+If you're a cafe, for example, probably just go buy Square. And with Square Terminal, you can manage your POS, you can go manage your inventory, and that works great for you. So we don't think Retool is the right fit for everyone. We think Retool is the right fit if you truly need custom internal software. The good news is there's a lot of custom internal software being built across the world every day - it's half of all software - but it's also not 100%. And in many cases, you would actually be better served just by buying something off the shelf.
+
+So that's something we're pretty honest with ourselves about, and pretty honest with our customers about too, which is if you're a cafe, probably just go buy Square instead of trying to build your own POS system. It's going to be a lot of work, and honestly, not going to get you a much better result than just buying Square. So...
+
+**Adam Stacoviak:** For sure. Yeah, I think what I mean is -- obviously, a barber is probably going to use Square. My barber uses Square, for example. I love Square, by the way. Shout-out to Square. I always enjoy using Square, whether I'm at the barber or a coffee shop. I always love it. What I was thinking of was more like you know I'm going to use something like a Square,
+
+I'm going to use something like a Shopify to do some version of commerce. And so because of that, you can assume certain things about the kind of business I'm running, and so you can complement that tooling, because there's integrations, and because there's some tooling that might be bespoke to me, that makes sense to build inside of Retool. I was thinking that you may have made some inroads there, where you can sort of predict a set of internal tools for a business. I just was thinking aloud on that. That'd be kind of cool. That'd be kind of cool. I think so. Because I think there's some opacity... Well, there's some opaqueness to, and some opacity level to Retool, where it's kind of hard... There's a lot of abstract. You can go and look at some of your templates, you can go and sort of see some things, but I've never seen a fleet of internal tools beyond what I've built for myself, which is like maybe one or two. It's nothing that's beautiful or amazing. And so I was thinking, is there a way to give me a glimpse behind the scenes to like a fleet of tooling, and how much of that fleet of tooling do 80% of the companies that use Retool always build? Is there a Retool for Retool, essentially?
+
+**David Hsu:** Yeah, totally. We see our customers building a lot of pretty similar internal apps. But to give you a sense, a company like Brex, I think they have maybe around a few thousand tools that run inside of Retool. Coinbase is another example.
+
+I think a few thousand tools that run inside of Retool every day, actually. So yes, there is a good amount of overlap. I think we could be doing a better job showing off some of these use cases. But to be honest, with AI I think a lot of this becomes maybe a little bit less important, too... Because witth AI, nowadays in Retool you can go describe the problem that you have, and we'll go build you an app for that, actually, pretty quickly. And we pull on the knowledge that we have, all the expertise that we have.
+
+\[01:02:08.11\] And so that I think is pretty powerful too, is with AI you almost have to think less about, or you don't need to see examples anymore. For example with ChatGPT, you can go say "Hey, I'll upload an image. Style it this way." And it's not like "Oh, I need to see examples of this.' It's just "Oh, well, here's my image", if you will, because it happens so quickly now. So that I think is quite cool, too. And I think we should do a better job showcasing that on the website.
+
+**Adam Stacoviak:** Yeah. Interesting. I guess while we're on that subject, what about your internal tooling? I mean, you obviously probably use Retool to build Retool, so you've got to have a ton of your own tooling built on top of Retool. Can you give me a glimpse into some things that you may have built, or things that make sense for you to build over the years?
+
+**David Hsu:** Yeah, we have so many cool tools built in Retool. So we have one called Retool Home, which is inspired by Stripe Home. It's basically this internal -- it's for employees of Retool to use, and you can log on and you can actually go see other people... It's like an internal company directory. You can see fun facts about them, you can see -- people generally record a video when they first join Retool, so you can go see that... There are fun games you can play in there, like how many Retools can you recognize in a row, for example... We have a decision log in there, so you can see what are decisions that have recently been made, for example... So it's really cool. So that's one example of an internal tool that we've built inside of Retool.
+
+A lot of our customer tooling is built on top of Retool. So if we want to go, for example, reset someone's password, add credits to someone's account, pretty much all of that is built in Retool. A lot of our finance tooling is now built in Retool, which the finance team loves, because it is much more robust than using Excel spreadsheets for doing \[unintelligible 01:03:52.01\]
+
+**Adam Stacoviak:** For sure. Gosh.
+
+**David Hsu:** So it's pretty cool, yeah. A large portion of Retool, probably 70% of Retool runs on top of Retool. So...
+
+**Adam Stacoviak:** I love a spreadsheet, but... Not always. You know? Not always. I mean, there's times that I'm like "Can I just get back to the simple spreadsheet?" and I can organize things a bit differently. But then there's times I'm like "You know what? A spreadsheet's not that kind of cool. I'd rather have something better."
+
+**David Hsu:** Especially when it's really hairy. When you have multiple tabs, thousands of rows, hundreds of formulas, it's really hard to parse what is going on. And I think a lot of those spreadsheets, honestly, over the next few years, are going to go away, because AI is going to make -- with Retool and our Copilot product, for example, building apps is going to be so easy... And a lot of those spreadsheets that should have been apps, in the future are going to be apps.
+
+**Adam Stacoviak:** Yeah, I wonder... And then I wonder about how do we -- how does that world connect? Because it's nice to have a random bespoke app, but at least I know where my spreadsheets live, and I can query them, and I can search them, I can history them... And maybe you'll eventually have that kind of stuff too, but then I think "Gosh, now I've got applications that could have been spreadsheets simplified. Now they're full-fledged applications." Maybe I've got them in a special URL, and I've got to manage the domain, and DNS, and other stuff like that. Or they're just a simple URL schema inside of Retool that makes it easy to go to customer.retool.com/whatever. Maybe it's like that. I don't know. How do you manage a world once you've got a few of those bespoke apps versus spreadsheets?
+
+**David Hsu:** I think bespoke apps are much easier to manage than spreadsheets, because with spreadsheets you basically have no guardrails. There's files, maybe on the cloud, maybe on your computer... It's hard to do version control on them, it's hard for multiple people to collaborate on them even...
+
+**Adam Stacoviak:** Sure. Google Sheets is pretty easy, but I think a file-based sheet is less easy.
+
+**David Hsu:** Yeah. But even Google Sheets is hard to tell cell history... For example, if someone accidentally fat-fingers a cell, that becomes quite com--
+
+**Adam Stacoviak:** That's true...
+
+**David Hsu:** So I think apps are honestly way better than spreadsheets. Well, in some cases, simple spreadsheets make more sense than apps, but I think as app building becomes way easier, there's going to be a lot more apps, I think.
+
+**Adam Stacoviak:** \[01:06:09.28\] Okay. So today - am I able to go into Retool today? This shows just how naive I am about how Retool works, so bear with me. Despite y'all sponsoring us, despite me should know a lot more, this is maybe a TMI in a way. Can I go into Retool now and tell it what kind of application or what kind of tool I want to build, or what I need? I have a Postgres database, or this or that. Can I describe the problem today and it builds me the application today in Retool?
+
+**David Hsu:** Virtually. We are launching that later in August. Today you can build large parts of your app; you can go write queries, you can go build some UIs, but it's not a full end-to-end experience yet. This is our next big launch coming in August. We're actually launching a full end-to-end way. It's like app gen, basically. So it's something similar to Bolt or Lovable, where you can just say what you want and it builds the whole app for you.
+
+But what's going to be really incredible is the fact -- it's two things. One is that we can actually build on top of your existing data sources. What we've found is with many of our customers, like Amazon and others -- Amazon has a lot of databases. They use Snowflake internally, they use RDS... They have so many places where data lives. And for them, they're not really interested in using something like Bolt, because Bolt is really great for prototyping. It's great for generating a new application from scratch... But how often are you truly generating a new app on top of new data, where you have no data really? Not very often, actually. In most businesses, the data already exists, but you want to build an app on top of that. And that is where I think Retool is really going to shine. We're going to be the first product in the market that allows you to go generate applications on top of pre-existing data. Whether it's in Postgres, Snowflake, Databricks, Salesforce, or Stripe, you can actually go generate an app entirely based on that, which is going to be really cool.
+
+The second thing is that because we're so focused on internal software, we'll give you a lot of things right out of the box, like authentication, authorization, audit logs, stuff like that. And so you can then say "Hey, actually, these apps could only be usable by these people, or be usable only by these \[unintelligible 01:08:29.26\] groups, actually", which is going to be super-powerful. So we imagine catalyzing a wave of application building with this product, this Copilot product. It's coming in August, so we're pretty excited about that.
+
+**Adam Stacoviak:** Interesting. Yeah, I'm not familiar with all these details of what's coming... I just assumed it was already there, but apparently it's coming. So you'll be able to generate a full-on Retool application from scratch, which is essentially what a lot of other platforms are already doing, but you were saying on top of a database. When you said that, I was thinking, "Okay, so what will you be able to do, I suppose, by just looking at the database?" Will you take a lot of the prompt and context of what to do, but can you also just say "Hey, look at this database and generate the screens that should manage this kind of database"? Is it going to be that easy to generate, or what are some of the processes and guardrails going to be when you do that? Do you know?
+
+**David Hsu:** Yeah, totally. So what's going to happen is once you've connected a database or an API even, the LLM will realize, "Oh, I have all these API inputs I can hit, or all these tables that I can pull data from." And so then, it's kind of what we were talking about before - the LLM is so much more powerful and so much more correct when it actually knows the shape of your data, or is actually able to interact with your database. So the apps it generates look much better, but also actually work immediately. And it can go straight to prod, because of the security guardrails and stuff like that. There's no product on the market that allows you to do that right now, so we think that's really quite powerful.
+
+**Adam Stacoviak:** \[01:10:11.00\] Interesting. So is this a leak you're doing here on the podcast right now? What are you doing \[unintelligible 01:10:14.27\]
+
+**David Hsu:** I actually don't know if I'm allowed to talk about it, but I'm pretty excited.
+
+**Adam Stacoviak:** You mean you're going to be in trouble right now, or what?
+
+**David Hsu:** It's coming just next month, so I hope I'm not in trouble, but we're really, really pumped about it.
+
+**Adam Stacoviak:** Okay. To me, this is the cool part, honestly. I mean, I like the idea of what you've already released, I like the idea of reactive, where you have reasoning and you have action in terms of agents... But to me, this is the coolest, because it's -- for what you just said there, which was the... When you said the agent can only have certain tooling. So the accuracy of the predictive of what it will do is so much more guardrailed \[unintelligible 01:10:52.28\] to be true versus false, or a hallucination, because you've given it parameters. And so I think about the same things... If I'm building on my real data, or my real API endpoints, then the LLM inside of Retool before it even builds out the thing I'm building out - I can have a conversation with it about my API endpoints. I can discover it as a maybe junior-ish or up-and-coming developer for the company, where I'm learning more and more about the different systems; I can plug into Retool, given I have auth, and given I have whatever, to build out whatever tooling I need to, and begin to query the database and the API to learn even... And maybe to ask it, "Hey, we have this problem. How can this internal tool that we need to build solve it? What API endpoints will we use?" That to me is -- that's where you want to be at. And as you mentioned before, the accuracy level is high, because you've literally got the API and the endpoints and the details there. The context is so clear, versus abstract.
+
+**David Hsu:** Yeah, exactly. So we think that a lot of the app gen products in the market today are quite good at UI generation or prototyping, which is awesome. That's very valuable. But it's hard to -- as far as we can tell, no one is really using these app gen products that go straight to production, because in the end, they're not very robust, especially if you're an enterprise. It's very hard to get the stuff -0 it's hard to deploy it in your cloud, for example, on top of your data, there's no security... So that's all really quite challenging.
+
+I think for us, we think the opportunity is if we focus specifically on internal software on top of your data, can we go build an AppGen product that not just generates UIs, but is immediately actually usable, and you can actually deploy it to your customers? And I think with all the context we have for your data from other apps you've built, we can do that. And so we're really excited about that opportunity for us.
+
+**Adam Stacoviak:** Does that change Retool at all? I mean, I know since the beginning it's been internal tools, and you can sort of expand the umbrella of what internal tools means to the point where it just means software... But does this change the trajectory at all of Retool? Does it take you from an internal tools company to something that's -- or does that even bottleneck you from the true potential you can achieve?
+
+**David Hsu:** Yeah, so we've thought a lot about how this changes for our customers. And our hypothesis right now is today, most developers are the ones that are "hands-on keyboard" coding. And so the developer is the one building the application. \[unintelligible 01:13:37.28\] AI, but they're still building it. It's still you that is building it. What we see in the future is we think that today's developers will move to more of a software architect kind of role, where they'll be building a lot of the building blocks for AIs to use. So for example, maybe you work on the API endpoints, maybe you work on the data schema, because that stuff is really important. But for the actual application building, you don't need to be doing that anymore. It's going to be either the AI or even lines of businesses actually building and generating applications based on your building blocks.
+
+\[01:14:17.23\] And so we have this concept that we're internally calling tomorrow's developers, and the idea there is today's developers are the ones that have both the skill and the knowledge to go build applications. But tomorrow's developers may just have the skill and not the knowledge, actually. What I mean by that is -- my wife is actually a good example of someone like this, which is... She was an econ major in college, and she wrote some MATLAB in college, so she's able to code. She's able to think in a coding-like way, if you will. But if you go ask her to go build a web application, she would be pretty overwhelmed, because there's just so much going on. It's so complicated to go build a web app today. You have to worry about what frontend framework do I go and choose? What state management library do I use? What are the pros and cons of using something like Heroku, or Vercel for example today, versus AWS? Should I use a VM, or should I go use an all-in-one platform? Should I use Supabase or should I go use Postgres? What's the difference? It's so complicated. There's so much knowledge that is required to go build an application. And so we think that these kinds of people with AI and with software architects, with software engineers becoming software architects building the building blocks, if you give them the building blocks and you give them AI, they will be able to build really powerful applications they can actually directly use.
+
+And so that's kind of our vision, is we think that software engineers, their role will change a little bit, and a lot of these line of business people that can learn SQL, can pick up SQL, will now be able to ship secure applications, and actually build them themselves and actually use them in production. So yeah, it's going to be a pretty exciting next few years, I think.
+
+**Adam Stacoviak:** It would be interesting to think about - to use your idea of tomorrow's developer - somebody who has the skills, meaning they understand how to leverage the power of an LLM, or however that evolves to be talked about, whatever lexicon makes sense to say LLM. They have the skills to do so, but they don't have the knowledge of how to build software necessarily. They may have some knowledge on how to build software, but not how to build software or internal software for the company that they work for, right? They don't have a full understanding, but they can use the platform to query and learn, and then actually build. And if you give it enough of the security measures - like you mentioned, authentication, authorization, if that's already baked into the way things work, then it's pretty easy to give people more and more trust when it could be either just taking offline or regenerated differently in such a rapid manner. We used to have a fear, I suppose, of shipping software, because the cost and time to change or to blame or to fix or to redeploy was more time consuming, to put it bluntly. And now it can be a lot less time involved there.
+
+\[01:17:24.11\] So maybe that speed to deployment is what's changing things to make it more trusted. It'd be kind of cool to be this future tomorrow developer, not really have the skills to do software development or the knowledge, but have the skills to conduct and architect based on feedback and learnings to deploy something to prod, and not really be a software developer. That's kind of what I think about, honestly.
+
+**David Hsu:** Yeah. We see a lot of data teams doing that now with Retool, and I think that we're pretty far from total non-technical people being able to do this, because I think if you don't know how to write SQL, if you don't know how to debug it, it's going to be a little bit challenging for you to go build full stack applications. But if you are able to write SQL and you have software architects or today's software developers becoming software architects building you the building blocks, then absolutely. I think it's going to be really quite powerful.
+
+**Adam Stacoviak:** Interesting. Alright, David, tomorrow's developers can't wait to ship some stuff to them and get some cool stuff. Anything left? Anything left in closing? I know we're -- I think we have like maybe one-ish minute left. I just realized we're right at time here... What's left that I haven't asked you, that you want to leave on the table here?
+
+**David Hsu:** I think we covered a lot. We covered agents, we covered Copilot, we covered tomorrow's developers... Yeah, I feel pretty good. Yeah.
+
+**Adam Stacoviak:** Yeah. Right on. Glad to finally have you here. Big fan of Retool, as you know... I'm just such a fan of enablement, like you've done. I don't personally in my small business have a vast need for internal software, but I know a lot of teams who do. And I've used -- what's kind of cool actually, to circle it back, is I used a vendor onboarding form recently. And I can tell it was a Retool application I was using... Like, "Okay, cool..."
+
+**David Hsu:** That's awesome.
+
+**Adam Stacoviak:** Because we're sponsored, so we deal with a lot of different brands who work with us and whatnot, and so we have to do vendor onboarding, and who are we, where do we live... Not so much live, but like where we work, kind of thing, and I have to share some data with them. And I filled out a form and it was a Retool application as an intake form for them. I was like "That's pretty cool. That's pretty cool."
+
+**David Hsu:** That is really cool. Yeah. Awesome.
+
+**Adam Stacoviak:** Yeah. After all these years of working together, finally circling back personally to use some internal tools from Retool. Alright, David, thank you so much. It was good talking to you.
+
+**David Hsu:** Thanks for having me on. Yeah, it was a lot of fun.

Programming with LLMs (Interview)_transcript.txt

@@ -0,0 +1,667 @@
+**Jerod Santo:** We are here with David Crawshaw, CTO and co founder of Tailscale. David, welcome to the Changelog.
+
+**David Crawshaw:** Yeah, I'm not actually the CTO anymore.
+
+**Jerod Santo:** Oh, no. Your LinkedIn is outdated.
+
+**David Crawshaw:** Oh, does it still say that? I thought I had updated it.
+
+**Adam Stacoviak:** Are you masquerading, David?
+
+**Jerod Santo:** That's real time LinkedIn updates here... \[unintelligible 00:04:14.23\]
+
+**Adam Stacoviak:** It's good. JIT LinkedIn updates.
+
+**Jerod Santo:** I read it somewhere, but usually \[unintelligible 00:04:19.08\] before the show starts.
+
+**Adam Stacoviak:** Me too.
+
+**David Crawshaw:** I think my LinkedIn -- it might be confusing, because it still lists that I was the CTO. I stepped back from the CTO last year.
+
+**Jerod Santo:** Okay. So what are you doing now?
+
+**David Crawshaw:** I am spending my time exploring sort of new product spaces, things that can be done, both inside and outside of Tailscale.
+
+**Jerod Santo:** Very cool.
+
+**David Crawshaw:** Most of my work inside Tailscale is around helping on the sort of customer side, talking to users or potential users about how it can be useful. And then, because I have such an interest in sort of the world of large language models, I've been exploring that. But that is not a particularly good fit for the Tailscale product. I spent quite a long time looking for ways to use this technology inside Tailscale, and it doesn't really fit. And I actually think that's a good thing. It's really nice to find clear lines like that, when you find something where it's not particularly useful... And I wouldn't want to try and -- a lot of companies are attempting to make things work, even if they don't quite make sense... And I think it's very sensible of Tailscale to not go in that direction.
+
+**Jerod Santo:** Do you mean like deploying LLMs inside of the Tailscale product? Or how do you mean it wouldn't fit?
+
+**David Crawshaw:** Well, yeah... So what would Tailscale do with LLMs is the question I was asking from a Tailscale perspective. I think Tailscale is extremely useful for running LLMs yourself, for a network backplane... In particular because of the sort of surprising nature of the network traffic associated with LLMs.
+
+On the inference side -- so you can kind of think about working with models from both a training and an inference... These are sort of two sides of the coin here. And training is very, very data-heavy, and is usually done on extremely, extremely high bandwidth, low latency networks, InfiniBand style setups, on clusters of machines in a single room. Or if they spread beyond the room, the next room is literally in the building next door. The inference side looks very different. There's very little network traffic involved in doing inference on models, and in terms of bandwidth. And the layout of the network is surprisingly messy. This is because sort of the nature of finding GPUs is tricky, even still today, despite the fact that this has been a thing for years now.
+
+**Adam Stacoviak:** Very tricky.
+
+**David Crawshaw:** Yeah. I feel I should try and explain it, just because it's always worth trying to explain things... But I'm sure you all know this, which is that if you're running a service on a cloud provider that you chose years ago, for very good reasons, all the cloud providers are very good at fundamental services, but they all have some subset of GPUs, and they have them available in some places, and not others, and it's never quite what you're looking for. And if you are deciding to run your own model and do inference on it, you might find your GPU is in a region across the country, or it's on a cloud provider that's different than the one you're using... Or your cloud provider can do it, but it's twice the price of another one you can get... And this leads to people ending up far more in sort of multi-cloud environments than they do in sort of traditional software. And so Tailscale actually is very useful there. So for users, I think it's a great fit. But what does the product actually need as like new features to support that? And the answer is it actually is really great as it is today for that.
+
+**Jerod Santo:** \[08:11\] There's no specific AI angle that you can add to the product and immediately make it more useful.
+
+**David Crawshaw:** Yeah, I think that's right. I mean, we came up with some proposals, but they're not exciting. We'd be doing it because corporate at headquarters told us to find an angle for AI, or something like that. We as a startup have the option of just not doing that, and so we didn't. So...
+
+**Jerod Santo:** Well, we should probably disclaim that Tailscale is a past, and of course, hopefully a future sponsor of Changelog, and that Adam's a huge --
+
+**Adam Stacoviak:** We're working on it.
+
+**Jerod Santo:** ...Tailscale user, and brings it up often... But this is not a sponsored episode. In fact -- well, first of all, we don't do sponsored guest appearances. But also, I had no idea that you were a co-founder of Tailscale when I read your blog post. That made me reach out to you.
+
+**Adam Stacoviak:** Me neither.
+
+**Jerod Santo:** I found it out afterwards. I was like "Oh, cool."
+
+**David Crawshaw:** That's great. I didn't know we'd be talking about Tailscale at all when I came here today, so we're both basically on the same page.
+
+**Jerod Santo:** Yeah, there we go.
+
+**Adam Stacoviak:** "I still work there..." I'm just kidding. \[laughter\]
+
+**Jerod Santo:** Real time update, I did double-check LinkedIn... You are correct, it says 2019 to 2024 was CTO, but you just see co-founder Tailscale and then CTO next to it, and you move on. And that's probably what Adam and I both did.
+
+**Adam Stacoviak:** Same. Exactly the same.
+
+**Jerod Santo:** We didn't realize there was an end date on that particular role.
+
+**Adam Stacoviak:** Yeah. The nomenclature or usage on -- the metadata usage on LinkedIn is the... Their UI is like which date, which month did you begin, and it says Present. So the assumption was there. I didn't read your byline on the role.
+
+**David Crawshaw:** Maybe what I'll do is I'll put something new above it, and that'll make it clearer.
+
+**Jerod Santo:** There you go.
+
+**David Crawshaw:** But I don't want to mislead anyone.
+
+**Jerod Santo:** Yeah. Fair enough. Fair enough.
+
+**David Crawshaw:** I also honestly don't check LinkedIn very often. It's not a big part of my life.
+
+**Jerod Santo:** We usually check it right before a show to make sure we get someone's title right, which is why we're both eating crow right now for getting it wrong. But...
+
+**Adam Stacoviak:** Well, to get back into the mood or the groove, whichever you want to call it...
+
+**Jerod Santo:** Let's get into both.
+
+**Adam Stacoviak:** Well, I'm a Tailscale user, as you know. I just trimmed some machines, because I've been doing some more home labbing... So I use Tailscale really only in my home lab. And thank you so much for this free tier, because I don't want to give you any money, honestly... I'm just kidding with you. I think you're amazing, but I've gotta put a hundred machines on my tailnet before I have to pay you any money. I've got 18. There's no way I'm ever going to pay you money... Based on your tiers, not mine.
+
+**David Crawshaw:** That's totally great. I'm very happy about that.
+
+**Adam Stacoviak:** Which is -- yeah, that's by design. That's by design. And I think, one, thank you, because it's let me be more network-curious, and more home lab-curious. So you as a corporation, Tailscale, have allowed me and enabled me and so many others to do just that, and that's so cool. And I applaud y'all for that choice by design.
+
+**David Crawshaw:** Well, thank you. That's excellent.
+
+**Adam Stacoviak:** That being said, at the same time, I've gotta dig. And it's not really a dig, it's just more like Tailscale is kind of boring in the fact that I don't have to do much to make it work. I put it in, I Tailscale up, and I'm done. You just work. I never have to worry about you working, unless you're not up --
+
+**Jerod Santo:** Do you want it to crash some more, or what are you looking for?
+
+**Adam Stacoviak:** I'm just saying, it's pretty boring. Unless I'm doing like serves, or I'm sharing a disk across the network... I'm not doing that kind of stuff, but... You know ,this whole multi-cloud shared GPU thing is super-cool, because you can have a tailnet on top of a different network, and share that GPU access, which I'm assuming is what you meant by that.
+
+**David Crawshaw:** That is what I meant, yeah.
+
+**Adam Stacoviak:** And that's just so cool. Honestly, it is.
+
+**David Crawshaw:** I mean, I love boring software.
+
+**Adam Stacoviak:** I do, too.
+
+**David Crawshaw:** And so to me, the fact that you're having a boring experience is perfect.
+
+**Adam Stacoviak:** Yeah, man. No surprises. No surprises.
+
+**David Crawshaw:** Yeah. It's a product that's designed to enable you to do more things, not for you to spend your days having to configure it.
+
+**Adam Stacoviak:** \[12:00\] It is so smooth. The DevEx experience on this thing is bar none. I know my machines, I know where they're at, I know if they're out of date... It's pretty easy to do that kind of stuff. And as an avid user and a lover of TailScale - again, not sponsored, just super-passionate - I can't see how an LLM would fit in it either. I just can't see how you would work in AI to make the platform better. I mean, I haven't thought about it deeply besides this 20-ish minutes so far in the conversation, but -- I mean, give me some time and I might.
+
+**David Crawshaw:** Yeah, if you come up with anything, let me know. I'm very excited about the idea of it. But software has to be, in some sense, true to itself. You have to think about its purpose when you're working on it, and not step too far outside that. I similarly wouldn't build a computer game into TailScale. I don't think that would be a particularly good fit for a product.
+
+**Jerod Santo:** Unless it's like an Easter egg.
+
+**David Crawshaw:** As an Easter egg it would be great, actually. A little quiz game or something built into the terminal...
+
+**Adam Stacoviak:** Once you have 100 machines in your tailnet, you get access to -- you unlock a secret machine name that's on your tailnet by default.
+
+**Jerod Santo:** There you go. It's for the Pro plan.
+
+**Adam Stacoviak:** There you go.
+
+**David Crawshaw:** Right. It can ask questions like "What is the oldest machine on your tailnet?" or something like that. That would be that would be a lot of fun, actually.
+
+**Adam Stacoviak:** There are some questions I would probably ask the tailnet... There are actually some things I don't know about my tailnet, that I could discover via a chat LLM interface. I mean, there are some things I can see some value in. But -- I mean, does everybody want that or need that? Maybe. I don't know.
+
+**David Crawshaw:** Yeah, I don't know either. I very much went looking for something I would use features like that for, and I didn't come up with anything. If you do come up with anything, I'd be very happy to hear about it.
+
+**Adam Stacoviak:** Honestly, now that I'm thinking about it in real time... You know, you have a column whenever you're on your admin and you're on your machine's dashboard, essentially... You can see last seen, or ones that are out of date. And unless you're savvy, you probably haven't enabled Tailscale's ability to auto update. Maybe you have, maybe you haven't. I forget which machines I've done it on. Like, every one I install again, once I knew that update was there, I do enable that. But sometimes I forget. So I might be like, "Okay, are there any of my machines that haven't been seen in a while? Or are there any versions that are -- give me a list of the ones that are out of date, that I should probably concern with around security" Because you're probably not emailing me about my security concerns, but my tailnet knows which ones are too far out of date if I haven't auto updated.
+
+**David Crawshaw:** That's true. I think we did actually email customers once about an out of date version where we were concerned about the security. I think that has only come up once. Mostly, keeping Tailscale up to date is sort of proactive, good security practice. It has fortunately not been a significant source of issues, in part due to careful design. A lot of engineers worked very hard to make it that way.
+
+**Adam Stacoviak:** For sure. And you've got a lot of amazing engineers there.
+
+**David Crawshaw:** Yeah, it's a great team.
+
+**Adam Stacoviak:** I guess now that I'm thinking about it, I do have some ideas...
+
+**David Crawshaw:** Nice.
+
+**Adam Stacoviak:** I mean, I think -- this is the idea I have for Notion as well. I use Notion a lot more... Anything where you have a platform where you can create your own things on top of it. A tailnet. One tailnet is not the same as the next. Even though they operate the same, the way I use mine may not be the way you use yours. It would be nice to have an interface where I can just ask Tailscale how to tailscale, basically. Like, I have an idea, I want to create a new group, or a whatever, or I can be introduced to new features... It's discovery. And you're essentially, by not having your own, you force people to go into the public LLMs, essentially; into the ChatGPTs, into the Claude, into the Ollamas, into the DeepSeeks, or whatever you might have out there. And if you can corner the market on your own, I think you'd be, one, better served, because you know your documentation better, two, you know the deterministic nature of it is maybe non-deterministic, but you can probably fine-tune it a bit more to be more focused on your customer base... I'd probably ask my Tailscale LLM more questions if I could.
+
+**David Crawshaw:** \[16:12\] Yeah. So that's like an interesting sort of meta question there about LLMs, around how many models there should be in the world, from sort of a consumer perspective, in a sense... Because that's almost like you're just consuming it \[unintelligible 00:16:28.06\] And this sounds very similar to the question of how does search work on websites, which you could have asked 10 years ago or 20 years ago. You know, do I use the Wikipedia search, or do I go to Google and type in my search and maybe put the word wiki at the end to bring the Wikipedia links to the top? Both of these are valid strategies for searching Wikipedia. And I honestly don't use the Wikipedia search and I haven't in a while, so it may be amazing... But I have as a consumer a general concept that the search systems on individual websites are not terrific. And Google is baseline decent... And so as long as I'm searching public data, I would generally prefer the Google search.
+
+I guess, in a sense, that's a less and less true statement every year, because the large chunks of websites are just not public data anymore. Like, you can't search Facebook with Google, you can't search Instagram with it, you can't find a TikTok with it, or anything like that. And so the existence of those - I think they sometimes get called walled gardens - says that we should have more fine-tuned tools like that. And there's just a lot of similarities there. So should a startup the size of Tailscale build customized models for that, for its users? I think it's sort of a big, open-ended question around how the model space will evolve. And I think my last year of working with models, fine-tuning them, training them, carefully prompting them... You can do more and more just with carefully structured prompts and long contexts, that you used to have to use fine-tuning to achieve.
+
+But all of this, my sort of big takeaway is that they are actually extremely hard to turn into products, and to get those details right, in a general sense, for shipping to users. They're actually quite easy to get going for yourself, and I think if anything, more people should explore running models locally, and playing with them, because they're a ton of fun and they can be very productive very quickly. But in much the same way that it's really easy to write a Python script that you run yourself on your desktop, versus a Python script you ship in production for users, LLMs have this huge sort of complexity gap when it comes to trying to build products for others. And so I agree that that sort of tooling would be fun, and should exist. I also think where we are today, it's quite hard for a team the size of a startup to ship that as not part of their core product experience.
+
+**Adam Stacoviak:** What if it enabled so much deeper and greater usage? Because the one thing you want to do as a startup or a brand like you are, I would imagine at least from the outside, is a deeper customer is better than a shallow customer, right? If I've only got a few machines - well, one, my affinity and my usage is lower, so maybe my value is lower... But if I'm deeply entrenched, and it's as a result of great documentation, which you have... But docs - they are good when you have a particular precision thing and you want to read and understand and discover how a feature works, and they only go so far. And sometimes they are even out of date. Just hypothesizing that, whether or not this -- what would be required, one, in terms of a lift, one in engineering power and two potentially financial power... And then two, what is that costing you by lack of more deep users and shallow users in comparison?
+
+**David Crawshaw:** \[20:18\] I think that is exactly the right way to frame the question for a business, and I don't know the answer to a lot of those questions. I can talk to some of the more technical costs involved. What the benefits would be to the company is extremely open-ended to me. I don't actually -- I can't imagine a way to measure that. Based on talking to customers of Tailscale who deploy it, thinking about the \[unintelligible 00:20:43.07\] And so to go back to something you said earlier, about how you use it and you don't pay for it... I think that's great, because Tailscale has no intention of making money off individual users. That's not a major source of revenue for the company. The company's major source of revenue is corporate deployments. And there's a blog post by my co-founder, Avery, about how the free plan stays free on our website, which sort of explains this... That individual users help bring Tailscale to companies who use it for business purposes, and they fund the company's existence.
+
+So looking at those business deployments, you do see Tailscale gets rolled out initially at companies for some tiny subset of the things that it could be used for. And it often takes quite a while to roll out for more. And even if the company has a really good roadmap and a really good understanding of all the ways they could use it, it can take a very long time to solve all of their problems with it. And that's assuming they have a really good understanding of all the things it can do. And the point you're making, Adam, that people often don't even realize all the great things you can do with it is true. And I'm sure a tool that helped people explore what they could do would have some effect on revenue.
+
+In terms of the technical side of it and the challenges - there are several challenges. In the very broad sense, the biggest challenge with LLMs is just the enormous amount of what you might call traditional, non-model engineering has to happen at the front of them to make them work well. It's surprisingly involved. I can talk to some things I've been working on over the last year to give you a sense of that... Beyond that, the second sort of big technical challenge is one of Tailscale's core design principles is all of the networking is end-to-end encrypted. And the main thing an LLM needs to give you insight is a source of data. And the major source of data would be what is happening on your network, what talks to what, how does it all work? And that means that any model telling you how you could change your networking layout, or give you insight into what you could do would need access to data that we as a company don't have, and don't want. And so we're back to - it would have to be a product you run locally, and have complete control over... Which is absolutely -- my favorite sorts of products are that. I like open source software that I can see the source code for, compile myself, run locally. That's how I like all things to be. But trying to get there with LLMs in the state they are today is actually, I think, pretty tricky. I don't think I've seen an actually shipped product that does that really well for people. There's one, there's a developer tool that I hear a lot of good talk about, that I don't -- I'm just trying to live-search for it for you. Nope, that's the wrong one. That's Magic Shell History, which also sounds really cool. I should use that one.
+
+**Jerod Santo:** Is that Atuin?
+
+**David Crawshaw:** Atuin, yeah.
+
+**Jerod Santo:** That one's awesome.
+
+**David Crawshaw:** Oh, you've used it? Oh, great.
+
+**Jerod Santo:** I'm a daily user. Yeah. No LLMs involved on that one.
+
+**David Crawshaw:** \[23:59\] Yeah, I thought that was one... There's another one that is in the sort of agent space for developers as they're writing programs, and it helps you -- it's like a local Claude, effectively. And it's primarily built around helping you construct prompts really carefully for existing open models. And it's come up several times, and I'm sorry, it's falling out of my head. I will look it up later.
+
+**Jerod Santo:** That's alright.
+
+**David Crawshaw:** But I hear very positive things about it. That's the closest I've seen to sort of a shipped completely local product that does that sort of thing. On which models to use, I think given the state of models that exist today, open models, the major shipped open models are so amazing that it always makes sense to start with one of those sort of models as a -- if nothing else, as a pre-trained base for anything that's happening. Building a model from scratch is a very significant undertaking, and I don't think is necessary for most tasks. The available open models are extremely general-purpose. And so at worst, you would be fine-tuning from one of those to build a product. If you take one of the Llamas, or... I mean, there's a lot of talk about DeepSeek, which produces terrific results... It's a very large model. It'd be very hard to start with it, though I understand there's some very good distilled work coming from it using other models.
+
+**Break**: \[25:26\]
+
+**Jerod Santo:** So you've been using those in your day-to-day programming work for the last year. And back in early January you wrote this post "How I program with LLMs", which I've found to be refreshingly practical and straightforward, your findings. You said you've been actively trying these things. I feel like I've been passively trying them; not really trying to optimize my setup, but just like a Neanderthal kind of poking at a computer box... Like, "Oh, does this work? No? Okay..." For the last couple of years. So I do use these things, but I don't think as effectively as most, or at least some.
+
+And I loved your finding that, of course, you're building something as a result of it. But can you take us on that journey over the last year or so, where you started with LLMs, and what you've found in your day-to-day programming?
+
+**David Crawshaw:** Yeah. I don't think your experience is unusual, actually. I think almost everyone has your experience. And for most software, I am in the same category. I try things at the very surface level when they're new-ish, and see if there's any really obvious way they help me. And if they don't, I put them aside and come back later. A great example of that is the Git version control system. It was 10 years before I really sat down and used it. I was using other version control systems. After 10 years, I was like "Okay, this thing's probably going to stick around. I guess I'll get over its user interface and figure it out."
+
+**Adam Stacoviak:** "Fine... Git it is."
+
+**David Crawshaw:** I was reluctant, but I got there in the end. LLMs really struck me as fascinating. I made this active decision to not do that with them, and set out on a process of trying to actively use them... Which has involved learning just a really appalling amount, honestly. It's very reasonable that most engineers haven't done really significant things with LLMs yet, because it's too much cognitive load. If you're writing computer programs, you're trying to solve a problem.
+
+**Jerod Santo:** Right.
+
+**David Crawshaw:** You only have so much of your brain available for the tools you use for solving problems, because you have to fit the problem in there as well, and the solution you're building. And that should be most of what you're thinking about. The tools should take up as little space as possible. And right now, to use LLMs effectively, you need to know too much about them. And that was my big takeaway 11 months ago or so, which is why I started working on tools with some friends to try and figure this out. Because there has to be a way to make this easier. And my main conclusion from all of that is there's an enormous amount of traditional engineering to do in front of LLMs to get there.
+
+So the first really effective thing I saw from LLMs is the same thing I think most engineers saw, which was GitHub Copilot, which is a code completion -- actually, GitHub Copilot has taken on new meanings.
+
+**Jerod Santo:** It's more than that now, right?
+
+**David Crawshaw:** Yeah, it's an umbrella brand that means all sorts of products... And I honestly haven't even used most of those products at this point. The original product is a code completion system built into Visual Studio Code, where as you type, it suggests a completion for the line, or a few lines beyond that, of where you are... Which is building on a very well-established paradigm for programming editors.
+
+\[32:20\] Visual Studio 6.0 did this 25 years ago with IntelliSense, for completing methods in C++. This is not a new idea. Around the same time we had etags for Emacs, or ctags, I should say, which gave us similar things in the Unix world... And so this is sort of extending that idea by bringing out some of the knowledge of a large language model in the process of completing. And I'm really enamored with the entire model. Copilot's original experience when it came out was magical. There was nothing like this before.
+
+**Adam Stacoviak:** It was.
+
+**David Crawshaw:** And it really, I think, jump-started a lot of interest in the space from people who hadn't been working on it, which was almost all of us. And from my perspective, the thing that really struck me was "Wow, this works really well. And wow, it makes really obvious, silly mistakes." It was both sides of this. It would suggest things that just don't compile in ways that are really obvious to anyone who takes a moment to read it. And it would also make really impressive cognitive leaps where it would suggest things that "Yes, that is the direction I was heading, and it would have taken me several minutes to explain it to someone", and it got there immediately.
+
+And so I spent quite a lot of time working on code completion systems, with the goal of improving them, by focusing on a particular programming language. And we've made some good progress there. We actually hope to demonstrate some of that publicly soon, in the next few weeks, probably in this sketch.dev thing that we've been building. We'll integrate it so that people can see it and give it a try. But so those models are interesting, because they're not the LLM experience that most users have. Like, when everyone talks about AI today, they talk about ChatGPT, or Claude, or these chat-based systems. And the thing I really, really like about the original Copilot code completion model is it's not a chat system. It's a different user interface experience for the knowledge in an LLM. And that's really a lot of fun. In fact, the technology is a little bit different, too. There's a concept in the model space called fill-in-the-middle, where a model is taught a few extra tokens that don't exist in the standard chat model. With fill-in-the-middle, which is a lot of fun, a model is taught a few extra tokens. And it's taught a prefix token, a suffix token, and a middle token. And what you do is you feed in as a prompt to the model the file you're in. And everything, all the characters before where the cursor is get fed right after a prefix token. So you feed in a prefix, all the characters of the file, then you feed in a suffix, and you feed in all the tokens after the cursor. And then you feed in the middle token, and then you feed in whatever goes into the middle to complete it. And that's the prompt structure for one of these models.
+
+And then the model keeps completing the thing that you fed in. It writes the next characters. And you train the model by taking existing code -- there's a few papers on how these models are trained, because Meta published one of these models, and Google published one of these models under the Gemma brand. There's a few others out there. There's one from Qwen, and some other derived ones.
+
+And you take existing code files, you choose some section of code, you mark everything before it as the prefix, everything after it as the suffix, and you fill in everything after it as the middle. And that's your training data. You generate a lot of that by taking existing code and breaking it up into these files randomly, by randomly inserting a cursor. Then you've taught a model how to use these extra characters, and how to complete them. And so it's not a chat model at all, it's a sort of a sequence to sequence model. It's a ton of fun.
+
+\[36:22\] And the advantage of these systems is they're very fast compared to chat models. And that's the key to the whole code completion product experience, is you want your code completion within a couple hundred milliseconds of you typing a character. Whereas if you actually time Claude, or you time one of the OpenAI models, they're very slow. Like, they take a good minute to give you a result. And there's a lot of UI tricks in hiding that minute. They move the text around on the screen, then they stream it in...
+
+**Jerod Santo:** Right.
+
+**David Crawshaw:** Yeah, it's very clever.
+
+**Jerod Santo:** Because you're reading it word by word as it comes out, but it's basically stalling. You're like "Come on, just give me the answer already."
+
+**Adam Stacoviak:** That's right.
+
+**David Crawshaw:** Yeah, exactly. You can really feel it with the new reasoning models, 01, these things, because there's this pause at the beginning.
+
+**Jerod Santo:** The thinking phase. I'm like "Come on, I don't --" And it tells you what it's thinking, which is cool, but...
+
+**Adam Stacoviak:** "Think faster...!"
+
+**Jerod Santo:** "I don't care what you're thinking. Give me the answer." \[laughs\]
+
+**Adam Stacoviak:** I think it's actually kind of cool, when you see that reasoning come up.
+
+**Jerod Santo:** It is cool.
+
+**Adam Stacoviak:** I mean, you get to see like -- I feel like this is the closest we've glimpsed into the future than we've ever been able to, by watching the reasoning in real time. You see the active reasoning that it's happening, it explains the reasoning... Users ask me this, so I'm going to think about that. Okay, I've thought about that, which causes this..." And it's like this step process, and it reminds me of how I think, too. So I'm like "That's pretty dang cool..." But it's also a great trick. I agree.
+
+**David Crawshaw:** Yeah, it is a ton of fun to watch. I agree. And it is a lot of insight into how the models work, too... Because the insides of the models are a large number of floating point numbers holding intermediate state, and it's very hard to get inside of those. But the words, you can read them. You can make some sense of them.
+
+**Jerod Santo:** Right.
+
+**David Crawshaw:** So code completion is, I think, extremely useful to programmers. It varies a lot depending on what you're writing, and how experienced models are with it, and just how sort of out on the edge of programming you are. If you're really out in the weeds, the models can get less useful. I used a model for writing a big chunk of AVX assembly a few months ago, and the model was both very good at it and very bad at it simultaneously, and it was very different from the typical asking a model to help with programming experience. It would constantly get the order of operations wrong, or overcomplicate things, or misunderstand... It was a very different experience than typical programming.
+
+**Adam Stacoviak:** What model was this? How did you find it?
+
+**David Crawshaw:** I used all of them for that.
+
+**Adam Stacoviak:** Okay...
+
+**David Crawshaw:** And this is what I meant by --
+
+**Jerod Santo:** Just try 'em all.
+
+**David Crawshaw:** ...I'm spending a lot of time actively exploring the space. I'm putting far too much work into exploring the model space as I do work.
+
+**Jerod Santo:** It makes sense that there are specific models that are good for autocomplete, versus search, versus chat... But have you found the correct one for each particular subtask, or what's your advice there? Is it like "Use them all" or "Just stick with this. You'll be good"?
+
+**David Crawshaw:** I can't advise people to use them all.
+
+**Jerod Santo:** To use a bunch of them?
+
+**David Crawshaw:** Yeah. And this, I think, is the big problem. You mentioned that most programmers are probably using this... As far as we can tell, not one-fifth of programmers are using these tools today.
+
+**Jerod Santo:** How can you tell that?
+
+**David Crawshaw:** Through surveys. A couple of people have done surveys of programmers, and it seems to come back that most people are not using these tools yet.
+
+**Adam Stacoviak:** Wow.
+
+**David Crawshaw:** \[39:58\] Which is both shocking to me, because they're so useful, and it also makes a lot of sense, because it's a lot of work figuring out how to use them.
+
+**Adam Stacoviak:** I have an analogy that I'd like to share. If you're a runner, you probably wear running shoes, right? You're probably not going to run barefoot. I think it's like admitting to running barefoot. Like, you wouldn't do that. You would run a marathon with rocks on the road, and debris, and things like that barefoot? ...versus running shoes designed to aid you in the process of running to make it more speedy, comfortable, agile etc. I feel like that's where we're at. I've changed my tune, let's just say, because I feel like it's not going to go away. And to hear that one-fifth -- I haven't dug into these surveys, but that's surprising. One-fifth is using it... And it seems like, I guess, then the four of the five are saying no, or for the time, for the moment, denying it. I mean, do either of you disagree with that analogy? Is it way off, or is it -- Jerod, I saw you kind of like shake your head a little bit. What's your thoughts on that? No?
+
+**Jerod Santo:** Yeah... I mean, I don't think that these tools have come to the place that the running shoe has. I also think there's probably plenty of world-class runners who run shoeless, and would never run with a shoe on, because that's for fools. But I'm not going to go there.
+
+**Adam Stacoviak:** \[laughs\] Well, would you run a New York City marathon with no shoes?
+
+**Jerod Santo:** I wouldn't be so foolish as to try a New York City marathon.
+
+**Adam Stacoviak:** Okay. \[laughter\] You have to admit though that having the shoes on is probably better for you than worse for you.
+
+**Jerod Santo:** Well, at what point has the shoe proven itself to be useless? Because these tools routinely prove themselves to not just be wrong, but dramatically wrong, in ways that if you follow them, you will be like Michael Scott, who drives directly into the pond.
+
+\[41:49\]
+
+*"No, no, no, no, no. Look, it means go up, to the right, bear right over the bridge, and hook up with 307." \[Make a right turn\]*
+
+*"Maybe it's a shortcut, Dwight. It said go to the right."*
+
+*"It can't mean that. There's a lake there."*
+
+*"It knows where it is going. The machine knows...!"*
+
+*"This is the lane. THIS IS THE LANE!"*
+
+*"Stop yelling at me! Stop yelling!"*
+
+*"No, there's no road here!!"*
+
+**Adam Stacoviak:** Oh, yes...
+
+**Jerod Santo:** Because his voice assistant or his GPS tells him to keep going straight, and he just keeps going straight.
+
+**Adam Stacoviak:** I concur on that one, too. I concur.
+
+**Jerod Santo:** So I can see why you could get frustrated and throw up your hands and say "I want to come back to this in a year, or two years. But I'm going to let all the frontiers people, like David, figure all this stuff out. In the meantime, I've got code to write." I can see a lot of people saying that. I could see myself saying that. I haven't, because I am curious, and I don't want to fall behind... But I still don't feel like this is a must have for everybody today. But there are moments where I'm like "That was amazing."
+
+**Adam Stacoviak:** For sure. I'm actively trying to fit it in into everything I do, is I guess my perspective. If it's home lab, it's that; if it's contracts, agreements, proposals, if it's thinking, if it's exploration, if it's coding, if it's... You pick it's, your if it's, kind of thing. I'm trying to fit it in, and I'm just -- so I'm sitting down on my bench, I've got my socks on and I'm trying to put the shoe on, let's just say...
+
+**Jerod Santo:** \[laughs\]
+
+**Adam Stacoviak:** ...to kind of extend my analogy.
+
+**Jerod Santo:** If the shoe fits, you're going to wear it.
+
+**Adam Stacoviak:** Yeah, I believe that I'm going to put this shoe on and I'm going to wear it for every scenario that makes sense, because I can tell you I move faster, I think differently when I'm in those modes. Are they wrong? Do I always check it? Of course. But I know that it's coming for almost everything we do. Every task we do that's productive - coding, thinking, writing, whatever - it's coming for it, in a positive way.
+
+**David Crawshaw:** I mean, I totally agree that it is coming for it. I also think it's very early days, and a great reason to not learn this technology today is that it's changing so fast.
+
+**Adam Stacoviak:** Yeah.
+
+**David Crawshaw:** And that you can spend a very long time figuring out how to make it work, and then that can all -- all of that accumulated skill can be sort of made useless tomorrow. By some new product.
+
+**Jerod Santo:** Right. Well, if you remember, Stable Diffusion first dropped probably two years ago now.
+
+**Adam Stacoviak:** Yeah.
+
+**Jerod Santo:** \[44:05\] And we were enamored with prompt engineering. And what was that artist's name that you'd always -- if you added it to your Stable Diffusion prompt, it would automatically get awesome? And then he got mad, because everyone's using him to like make better pictures... Like, that whole technology, that magical incantation is just completely moot at this point. Like, it's probably easier now to get better pictures without being such a wizard... And whatever name you're invoking in the past is just -- that name doesn't do what it did on the last version of Stable Diffusion. Just as one instance. Prompt engineering has changed dramatically. And anybody who's ignoring it all, and just listening to us talk about it on the Changelog and just like staying with their regular life, they've saved themselves a lot of time, than those of us who dove in and decided they were going to memorize all the magic words.
+
+**David Crawshaw:** Yeah, absolutely. A year ago, a common technique with open models that existed was to offer them money to solve problems. You start every prompt by saying "I'll give you $200 if you do this." And it greatly improved outcomes.
+
+**Jerod Santo:** Yeah. Or "Let's take this step by step." That phrase was one of those magical things that made it better.
+
+**David Crawshaw:** Yeah. All of those techniques are gone now. If you try bribing a model, it doesn't help. There was a great example I saw of that where someone kept saying "I'll give you $200 if you do this", and they did it in a single prompt several times, and they got to the nth case, and it said "But you haven't paid me for the previous ones yet." \[laughter\]
+
+**Jerod Santo:** "Listen, fella... We had a deal."
+
+**Adam Stacoviak:** "No means no."
+
+**Jerod Santo:** That's great.
+
+**Adam Stacoviak:** No money means no. Yeah, there you go. Alright.
+
+**David Crawshaw:** Yes, they're very funny models. So I spent a long time believing, and I actually still believe this in the long term, that chat is currently our primary user interface on the models, and it's not the best interface for most things. The way to get the most value out of models today when you program is to have conversations with models about what you're writing... And that's -- I think it's quite the mode shift to do that. It's quite taxing to do that, and it feels like a user interface problem that hasn't been solved yet. And so I've been working a lot with Josh Bleecher Snyder on these things, and we have been sort of -- we spent a long time looking for how can we avoid the chat paradigm and make use of models. That's why code completion initially was so interesting, because it's an example of using models without chat. And it's very effective. So we spent a long time exploring this.
+
+To give you another example of something we built in this space - because we've just been trying to build things to see what's actually useful. We built something called Merd, merd.ai, which I think we've put up a few weeks ago, and it does merge commits for you. So if you try and push a git commit or do a rebase and you get a merge conflict, you can actually use LLMs to generate sophisticated merge commits for you.
+
+It turns out that's a much harder problem than it looks. Like, you would think you'd just paste in all of the files to the prompt, and you ask it to generate the correct files for you... Even the frontier models are all really bad at this. You almost never get a good merge commit out of them. But with a whole stack of really mundane engineering at the front - mundane is not the right word, because a lot of it's actually really very sophisticated. But it doesn't involve the LLM itself. It's about carefully constructing the prompt.
+
+**Jerod Santo:** Traditional.
+
+**David Crawshaw:** Traditional is a much better word. Yeah. You can actually get very good merge commits out of it. And that user experience seems much better for programmers to me; you could imagine that being integrated into your workflows to the point where you send a PR, there's a merge conflict, it proposes a fix right on the PR for you. And in fact, we attempted a version of that where there's a little GitBot that you can @ mention on a PR, and it sort of generates another PR based on it, that fixes the merge conflict for you.
+
+\[48:07\] And that sort of experience doesn't require the chat interface to be exposed to the programmer to make use of the intelligence in the model. And that is where I dream of developer tools getting, so that everyone can use them without having to learn a lot about them. Like, you shouldn't have to learn all the tricks for convincing a model to write a merge commit for you. There should be a button. Or not even a button, it should just do it when GitHub says there's a merge conflict. So it's actually -- it works pretty well. We've seen it generate some very sophisticated merge commits for us. I'd love to see more people give it a try and let us know what the state of that is.
+
+But so... Just because that is such a hard state to get to, we built Sketch, which exposes the traditional chat interface in the process of writing code. Because we're just not -- we don't think the models are at a point yet where we can completely get away from chat being part of the developer's workflow.
+
+**Jerod Santo:** So at what level of granularity is Sketch working at? And do you imagine it moving up eventually, wherever it is? Because the panacea, the silver bullet is what some folks are trying to do with Devin, for instance, where it's like you describe at a very high level a system, and it goes and builds that system v0 - from Vercel is another one - that's doing these things. And they're very much at, in my opinion, the prototype/demo level of quality, not the production level of quality in their output... And it seems like they're very difficult -- in my limited experience with these things, they're very difficult to actually mold, or - what do you do? I'm losing a word here. Sculpt. I don't know, like a sculpture. To actually sculpt what they come out with, and change it into something that you actually would write, or like.
+
+But those are like the very high level of "Well, it should have a contact form that submits to this thing." But maybe you're looking down more where I use them currently, which is "Yo, write me a function that does this particular thing." And at that level, it seems a lot easier to even chat to it, if I have to. I would rather not chat to it... But spit out code that I could copy-paste and modify, versus being like "I'm going to throw this away and rewrite it."
+
+**David Crawshaw:** Right. Yes, I think that lines up really well with the way Josh and I think about these things, where today if you open up a model, a cloud provider's frontier model, or a local DeepSeek, or even a Llama 70b, you can ask it to write a Python script that does something. It could be a Python script to go to the GitHub API and grab some data and present it neatly for you. And it will do a great job. These great models can basically do this in a single shot, where you write a sentence and out comes a Python script that solves a problem. And that's an astonishing technical achievement. It's amazing how quickly I've got used to that as a thing, but...
+
+**Jerod Santo:** \[laughs\] Yeah, you're not even impressing me right now, you know? Yes, it can do that. We all know.
+
+**David Crawshaw:** Exactly.
+
+**Jerod Santo:** But it is amazing.
+
+**David Crawshaw:** Exactly. Me five years ago, if you told me that, I would struggle to believe it. And yet now I just take it for granted.
+
+**Adam Stacoviak:** Yes.
+
+**David Crawshaw:** \[51:38\] And so that works. We've got that. We've got a thing that can write really basic Python scripts for us. Similarly, these systems, at least the frontier models, are good at writing a small React component for you. You can give almost any of them -- you need more than a single sentence. You need a few sentences to structure the React component, but out comes some HTML and some JavaScript in the React syntax, the JSX syntax, or the TSX syntax... And it's pretty close.
+
+It might need some tweaking, you might have some back and forth to get there, but you can get about that out of it. And clearly, models are going to improve. There's no evidence to suggest we're at the limit here as the models keep improving, every month at this rate... And part of what we're interested in Sketch is getting beyond helping you write a function - which I also use today; I get frontier models to write functions for me - to sort of "How can we sort of climb the complexity ladder there?" And so the point we chose is a point that is comfortable for us, and is helpful for us, is the Go package. How can we get a model to help us build a Go package to solve a problem?
+
+And there's an implicit assumption here in that the shape of Go packages looks slightly different at the end of this... Packages are a little bit smaller, and you have a few more of them than you would in a sort of traditional Go program you wrote by hand. But I don't think that is necessarily a bad thing. Honestly, in my own programming, as a Go program, I tend to write larger packages, because there's a lot of extra work involved in me breaking it into smaller packages. And there's often this thought process going on in my mind of like "Oh, in the future, this will be more maintainable with more packages." But it's more work for me to get there today, so I'll just, I'll combine it all now and maybe refactor it another day.
+
+And switching to trying to have LLMs write significant chunks of packages for you makes you do that work upfront. That's not necessarily a bad thing. It's perhaps more the way we'd like our code to end up. And so Sketch is about taking an LLM and plugging a lot of the tooling for Go into the process of using the LLM to help it. An example is I asked it the other day to write some middleware to Brotli-compress HTTP responses under certain circumstances, because Chrome can handle Brotli encoding, and it's very efficient. It's not in the standard library, or at least it wasn't the last time I looked... And the first thing it did was it included a third party package that Andy had written, that has a Brotli encoder in it. And so Sketch go gets that in the background in a little container as you're working, and has a little go.mod there that modifies... So that as you're editing the code, you get all the code completions from that module, just like you would in a programming environment. And more importantly, we can take that information and feed it into the model, as it's working.
+
+If we run the go build system as part of it, and if a build error appears, we can take the build error, feed it into the model. It's like "Here's the error", and we can let it ask questions about the third party package it included... Which helps with sort of some of the classic problems you see when you ask Claude to write you some Go code, where it includes a package and then it makes up a method in there that doesn't exist, that you really wish existed, because it would solve your problem...
+
+And so this sort of automated tool feedback is doing a lot of the work I have to do manually when I use a frontier model. And so I'm trying to cut out some of those intermediate steps where I said "That doesn't exist. Could you do it this way?" Anything like that you can automate saves me time. It means I have to chat less. So that's the goal, is to slightly climb the complexity ladder in the piece of software we get out of a frontier model, and to chat less in the process.
+
+**Jerod Santo:** Are you achieving that by having a system prompt, or are you actually fine-tuning? How are you as the Sketch.dev creators taking a foundation model and doing something to get here?
+
+**David Crawshaw:** Today it is almost entirely prompt-driven. There's actually more than one model in use under the hood, as we try different things... For example, we use a different model for solving the problem of - if we want to go get a package, what module do we get to do that? Which sounds like a mechanical process, but it actually isn't. There's a couple of steps there. So a model helps out with that.
+
+\[56:16\] There's very different sorts of prompts you use for trying to come up with the name of a sketch than there are for answering questions... But at the moment, it's entirely prompt-driven, in the sense that a large context window and a lot of careful context construction can handle this, can improve things. And that can include a lot of tool use. Tool use is a very fun feature of models where you can instruct -- so to back up and give you a sense of how the models work, an LLM generates the next token based on all the tokens that come before it. When you're in chat mode and you're chatting with a model, you can at any point stop and have the model generate the next token. It could be part of the thing you're asking it, or its response. That meta information about who is talking is sort of built into just a stream of tokens.
+
+So similarly, you can define a tool that a model can call. You can say, "Here's a function that you can call", and it will have a result. And the model can output the specialized token that says "Call this function, give it a name, write some parameters", and then instead of the model generating the next token, you pause the stream. You, the caller, go and run some code. You go and run that function call that it defined. You paste the result of that function call in as the next set of tokens, and then ask the model to generate the token after it. So that technique is a great way to have automated feedback into the model.
+
+A classic example is a weather function. So you define a function which says "Current weather." Then you can ask the model, "Hey, what's the weather?" and the model can say "Call function current weather." Your software that's printing out the tokens pauses, calls current weather, says "Sunny", you paste "Sunny" in there, and then the model generates the next set of tokens, which is the chat response saying "Oh, it's currently sunny." And that's the sort of easy way to plug external systems into a model.
+
+This is going on under the hood of the user interfaces you use onto frontier models. So this is happening in ChatGPT, in Claude, in all these systems. Sometimes they show it to you happening, which is how you know. You see it less now, but about six months ago you could see - in the GPT-4 model you would ask it questions and it would generate Python programs and run them, and then use the output of the Python program in its answer. I had a really fun one where I asked it "How many transistors fit on the head of a pin?" and it started producing an answer and it said "Well, transistors are about this big. Pins are about this big. And so I guess...", the magic little emoji appeared, "that this means this many transistors fit on the head of a pin." Some very large number. And if you click on the emoji, it shows you the Python program it generated to do the arithmetic. It executed that as a function call, it came back with the result, and that saved it the trouble of trying to do the arithmetic itself, which LLMs are notoriously -- they notoriously struggle with doing arithmetic. So this is a great thing to outsource to a program.
+
+**Jerod Santo:** That's a funny workaround, because if you're a calculator for words, you're not necessarily a calculator for numbers...
+
+And when you can't do those reliably, then you could just write a program that does it and returns the same thing every time.
+
+**David Crawshaw:** \[59:51\] Yes. They're very good at writing programs to do the arithmetic, very bad at doing the arithmetic. So it's a great compromise. The thing we do with Sketch is try to give the underlying model access to information about the environment it's writing code in using function calls. So a lot of our work is not fine-tuning the model, it's about letting it ask questions about not just the standard library, but the other libraries it's trying to use, so that it can get better answers. It can look up the Godoc for a method if it thinks it wants to call it, and use that as part of its decision-making process about the code it generates.
+
+**Adam Stacoviak:** Can you describe "let it ask"? I mean, you've said a couple of times and I've been curious about this. When you say "let it ask", what does that mean? Like, decompress that compressed definition.
+
+**David Crawshaw:** So at the beginning in your system prompt or something like your system prompt - it depends on the API, on exactly how the model works - there is a function call which is get method docs, and it has a parameter which is name of method. And then you can construct a question to an LLM that says "Generate a program that does this", with the system prompt, which explains that there's a tool call there. And so as your LLM is generating that program, it can pause, make a tool call that says "Get me the docs for this." And so the LLM decides that it wants to know something about that method call. And then you go and run a program which gets the documentation for that method from the actual source of truth, you paste it into the prompt, and then the LLM continues writing the program, using that documentation as now part of its prompt. So this is the model driving the questions about what it wants to know about.
+
+**Jerod Santo:** And it just blocks and waits for that to come back, effectively.
+
+**David Crawshaw:** Yes.
+
+**Jerod Santo:** Yeah, so it's like an embed.
+
+**David Crawshaw:** If you step back to like running llama.cpp yourself or something like this, you can sort of oversimplify one of these models as every time you want to generate a token, you hand the entire history of the conversation you've had, or whatever the text is before it, to the GPU, to build the state of the model. And then it generates the next token. It actually generates a probability value for every token in its token set.
+
+And then the CPU picks the next token, attaches it to the full set of tokens, and then does that whole process again of sending over the entire conversation, and then generating the next token. So if you think about that very long, big, giant for loop around the outside of every time there's a new token, the token is chosen from the set of probabilities that comes back, it's added to the set, and then a new set of probabilities is generated for the next token.
+
+You can imagine in the middle of that for loop having some very traditional code in there that inserts a stack of tokens that wasn't actually decided by the LLM, but then become part of the history that the LLM is generating the next token from. And so this is -- that's how those embeds work. You can effectively have the LLM communicate with the outside world in the middle there by it driving that. Or you don't even have to have it drive it. You could have software outside the LLM that looks at the token set as it appeared and then insert more tokens for it. So this is all the fun stuff you can do by running these models yourself.
+
+**Jerod Santo:** It's wild.
+
+**David Crawshaw:** Yeah, I know. It's so fun.
+
+**Break:** \[01:03:41.08\]
+
+**Jerod Santo:** Is Go particularly well suited for this kind of tooling, because of the nature of the language, or is it just your favorite, or...? Why Go?
+
+**David Crawshaw:** Yeah, that's a really good question. The best programming language for LLMs today is Python, and I believe that is a historical artifact of the fact that all of the researchers working on generative models work in Python. And so they spend the most time testing it with Python, and judging a model's results by Python output.
+
+There was a great example of this in one of the open benchmarks I looked at - and I believe this has all been corrected since then. This is all about a year old. There was a multi-language benchmark that tested how good a model is across multiple languages... And I opened up the source set for it, and looked at some of the Go code, because I'm a Go programmer... And it had been machine-translated from Python, so that all of the variable names in this Go code used underscores instead of camel case. And the models were getting a certain percentage success rate generating these results.
+
+So Josh went through, actually, and made these more idiomatic in the Go style of using camel case and putting everything in the right place, and the model gave much better results on this benchmark. so that's an example of where languages beyond the basic ones that the developers of the models care about are not being paid as much attention to as what you would like. And things are getting a lot better there; the models are much more sophisticated, the teams building them are much larger, they care about a larger set of languages... And so I don't think it's all as Python-centric as it used to be. But that is still very much the first and most important of the languages.
+
+As for how well Go works, it seems to work pretty well. Models are good at it by our benchmarks. Like we said, if we took the benchmarks and made them more Go-like, the models actually got better results. They have a real tendency to understand the language. We think it's a pretty good fit.
+
+There are definitely times when models struggle, but it's a garbage-collected language, which helps, because in just the same way that garbage collection reduces the cognitive load for programmers as they're writing programs, it reduces the load on the LLM in just the same way. They don't have to track the state of memory, and when to free it... So they have a bit more thinking time to worry about solving your problem. So in that way, it's a good language. It's not too syntax-heavy, but it's also -- it doesn't have ambiguities that humans struggle with. Yeah, it seems to work well --
+
+**Jerod Santo:** Pretty small...
+
+**David Crawshaw:** Yeah. There aren't a lot of -- I haven't seen much research into what is the best language for an LLM. It does seem like an eminently testable thing. Like, there's some interesting -- in fact, it may end up influencing programming language design, in a sense, of imagine you are building a new programming language and you develop a training set that's automatically generated based on translating some existing programs into your language, and you train models for it... You could imagine tweaking the syntax of your new language, regenerating the training set, and then seeing if your benchmarks improve or not. So you can imagine --
+
+**Jerod Santo:** Exactly.
+
+**David Crawshaw:** Yeah. You can imagine driving readability of programming languages based on your ability to train an LLM to write this language. And so yeah, there's lots of really fun things that will happen long term that I don't think anyone started on work like that yet.
+
+**Jerod Santo:** Right. So the level that you all are working at with Sketch, with Go in particular, is the prompting you're doing, and the contexting and everything else that you're building... Is it at a layer of abstraction where you could replace Go relatively easily with \*insert general programming language\*? Or is it like "Well, that would be a new product that we would build"? Like, how hard is that?
+
+**David Crawshaw:** \[01:09:54.18\] Yeah, it's a good question. All of the techniques we're applying at general, but they are all very -- each technique requires a lot of Go-specific implementation. So in much the same way that a lot of the techniques inside of a language server for a programming language - these are the systems inside VS Code for generating information about programming languages... The techniques are general for like what methods are available on this object. They are very similar in Go as they would be in Java, for example. But the specifics of implementing them for both languages are radically different. And I think it's a lot like that for Sketch. The tricks we're using for Sketch are very Go-specific. And if we wanted to build one for Ruby, we would have to build something very, very different.
+
+So yes, I consider it very much a Go product right now... And I really like that focus that that gives us. Because Go is a big enough problem on its own, let alone all of programming.
+
+**Jerod Santo:** Yeah, yeah, yeah. I'm just asking that because I wonder how valuable and important tooling like this would be for each language community to either provide, or fund, or hope that somebody builds... Because if the LLM-related tooling for Go, because of Sketch, just hypothetically, becomes orders of magnitude more useful than just talking to ChatGPT about my Elixir code, for instance... Well, that's a real advantage for Go and the Go community. I mean, it's great for productivity for Gophers.
+
+And going back to maybe the original question about "Should Tailscale have its own little chatbot built into it?", like, does each community need to take up this mantle and say "We need better tooling"? Or is it like VS Code should just do it for everybody?
+
+**David Crawshaw:** I mean, that's a really good question.
+
+**Adam Stacoviak:** Good job, Jerod.
+
+**David Crawshaw:** Yeah. So I very much admire VS Code. I use it... Which, I don't actually have to admire a program to use it.
+
+**Jerod Santo:** That's better than admiring, is using, I think.
+
+**David Crawshaw:** Yeah, that's right.
+
+**Jerod Santo:** \[laughs\]
+
+**David Crawshaw:** But actually, I do both. I both admire it and use it.
+
+**Jerod Santo:** Okay, fair.
+
+**David Crawshaw:** But to look at the inside of VS Code, which I've been doing a bunch of recently... VS Code didn't actually solve language servers for all programming languages. They built JavaScript and TypeScript, JSON, and I think they maintained the C\# plugin... They started the Go plugin, I think, and then it got taken over by the Go team at Google, who now maintain the Go support in VS Code. I don't think the Microsoft team built the Ruby support in VS Code. I don't know who did the Python implementation, but a lot of the machinery in VS Code is actually community-maintained for these various programming languages. So I'm not sure there is another option than imagining a world where each of these communities supports the tooling in some form. I don't know if each programming language needs to go out and build their own Sketch... Maybe there is some generalizable intermediate layer, some equivalent of a language server that can be written to feed underlying models.
+
+Given now -- we're just starting to explore this space. Sketch is very new. We basically started it sometime near the end of November. So there's not much to it yet... Yeah. But so far, what we've found is it's far more than the sort of language server environment that you get with VS Code. More machinery is needed to really give the LLM all the tooling it needs. The language server is very useful; we actually used the Go language server in Sketch. Gopls is a big part of our infrastructure. It's really wonderful software. But there's far more to it than that, to the point where we need to maintain an entire Linux VM to support the tooling behind feeding the model. So what each community needs to provide - I think that's the research in progress, is figuring that out.
+
+**Jerod Santo:** \[01:14:08.23\] Yeah. It's an interesting question, and one I think will be open for a while. I do not want to see a world where Python continues to proliferate merely because of its previous position. I do see with tooling like Devin, and Bolt, and V0... These are very frontend, JavaScripty companies that are producing these things, which is fine. But it's like, if you are just going to go use that, it's going to produce for you a React and Next.js frontend, with a Prisma-based back-end... It's all very much like "These are the tools it has", and that's all well and good. But that's going to proliferate more and more of that one thing, where I'd love to see a diversity where it's like "Yeah, is there a specific thing for Rails people? Is there one for people who like Zig?" Now I'm moving outside of the world of web development, but you know what I'm saying.
+
+And I think your answer might be right, which is like "Well, every community is going to have to provide some sort of greasing of the skids for whatever editor is popular or used", in order to make their tooling work well inside of these LLM-based helpers... Beyond just being like "ChatGPT knows about you", which is kind of like what people are at right now, is like "Does ChatGPT know about me?" It's the new "Am I googleable?" It's the new SEO at this point.
+
+**David Crawshaw:** I've heard people talk about that. A startup founder, who I won't name, mentioned that they were busy retooling their product so that the foundation models under things like V0 and Bolt would be more likely to npm include their package to solve a problem.
+
+**Jerod Santo:** It's super-smart to do that right now.
+
+**David Crawshaw:** I agree.
+
+**Jerod Santo:** Did they divulge any of the how? Like, what are the mechanical steps that you do that?
+
+**David Crawshaw:** I was actually really happy that they said that their plan was to make it really easy to npm \[unintelligible 01:16:08.27\] package and not require a separate signup flow to actually get started.
+
+**Jerod Santo:** Oh, that's nice.
+
+**David Crawshaw:** Yeah, I thought it was wonderful. Their solution to make their product more chatgpt-able, I guess you might say, is just make their product better... Which, if that's --
+
+**Jerod Santo:** \[laughs\] How avant-garde of them. Yeah.
+
+**David Crawshaw:** Yeah. I'm sure one day we'll end up in the search engine optimization world of frontier models. But today --
+
+**Jerod Santo:** There's gonna be some black magic for sale. You know, "Here's how you really do it."
+
+**David Crawshaw:** Yeah... I don't see why a frontier model couldn't run an ad auction for deciding what fine-tuning set to bring in... I had -- again, to talk about experiences, I was using one of the voice models and talking to it as I was walking down the street and I asked it some question about WD-40, because I had a squeaky door... And I think I described in my question WD-40 as a lubricant. And it turns out I just didn't understand that it's not a lubricant, it's a solvent, and the purpose of it is to remove grease.
+
+**Jerod Santo:** It took me years to realize that. I think someone finally told me... Because I've been using it as a lube all these years.
+
+**David Crawshaw:** Yeah...
+
+**Adam Stacoviak:** Oh, my gosh.
+
+**Jerod Santo:** Why do you have to keep reapplying it, you know... \[laughs\] It's not very good lube.
+
+**David Crawshaw:** Well, I just had your experience, but it was an LLM that told me.
+
+**Jerod Santo:** How hilarious.
+
+**David Crawshaw:** And it mentioned in passing, it's like "Yeah, you could also -- you could use WD-40 and then use a lubricant like", and then it listed some brand name. And the moment I heard the brand name, I was like "Oh, I see. A frontier model could run an ad auction on fine-tuning which brand name to inject there." And that would be a really--
+
+**Jerod Santo:** 100%.
+
+**David Crawshaw:** \[01:17:51.29\] Yeah. It wouldn't require doing it into the pre-training months ahead of time. You could do that sort of on an hour by hour basis. So that world is coming... And then once there's a world of ads, there's a world of SEO and all the rest of it.
+
+**Jerod Santo:** Well, the more paramount they become - and Adam, you can probably speak to this, because you're injecting it into every aspect of your life... Like, if the answer includes a product right there, you're just going to be like "Alright, I've got to get that." Sometimes you don't even realize Kleenex is a product. You think that that's a category. But no, that's a product.
+
+**David Crawshaw:** Yeah, absolutely.
+
+**Adam Stacoviak:** Hard to tell, honestly...
+
+**David Crawshaw:** Kleenex is an easy one for me, because we don't have Kleenex in Australia, where I'm from.
+
+**Adam Stacoviak:** Is that right?
+
+**David Crawshaw:** And so I came here and everyone started calling tissues Kleenex, and it was a bit of a surprise to me.
+
+**Jerod Santo:** \[laughs\]
+
+**Adam Stacoviak:** It's like Coke or Coca-Cola, something like that, you know?
+
+**David Crawshaw:** Right. Exactly.
+
+**Adam Stacoviak:** Yeah... I don't know if -- I've gotten some hallucinations, let's just say, on products... And even limited information of what's the true good option when it comes to product search. I haven't done a ton of it, mainly on like the motherboard search... Like, I want to do something that has the option for either an AMD Ryzen, or Threadripper, with more of a workstation, enterprise-class CPU, and I want to like maximize some PCIe lanes... So I'm just trying to like figure out what's out there... I'd prefer the chat interface to find things, versus the Google interface, which is search, by nature, to find things... But thus far, it hasn't been super-fruitful. I think eventually it'd be cool, but it's not there yet.
+
+**David Crawshaw:** I imagine in the YouTube video of this a little Intel Xeon banner will appear, just as you say "other options..."
+
+**Jerod Santo:** \[laughs\]
+
+**Adam Stacoviak:** That's right. Yeah, on the \[unintelligible 01:19:33.11\]
+
+**Jerod Santo:** Yeah, exactly.
+
+**Adam Stacoviak:** I'm a fan of Intel Xeons, too. I've got the Intel Xeon 4210...
+
+**Jerod Santo:** Well, now it's really popping up.
+
+**Adam Stacoviak:** There you go. Bing, bing, bing. It's like in Silicon Valley when Dinesh was talking out loud and they had the -- I think they had AI in the VR, and it was doing some cool stuff. It was pulling up ads real time. It was cool. But yeah, Intel's cool.
+
+AMD is cool. But PCIe lanes are even cooler, you know? Give me the max. Buy 16.
+
+**Jerod Santo:** David, maybe we close with this, for those who aren't Gophers out there... Of course, brand new, hot off the press, still in development, three months old Sketch.dev. Check it out if you're into Go and those kinds of things. But let's imagine you're just a Ruby programmer out there, and you came across your blog post about what you've been doing, these three methods of working with AIs; you have autocomplete, you've got chat, you've got search... Where should folks get started if they haven't yet? First of all, is today the day? Is it worth it now? Or should I wait? And then secondly, if I am going to dive in and I just want to use it in my local environment to just code better today, what would you suggest?
+
+**David Crawshaw:** Yeah, good question, especially for non-Gophers. I would suggest trying out the code completion engines, because they take a little bit of getting used to, but not a lot. And depending -- if you're writing the sorts of programs they're good at, they're extremely helpful. They save a lot of typing, and it turns out -- I was surprised to learn this, but what I learned from code completion engines is a lot of my programming is fundamentally typing-limited. There's only so much my hands can do every day. And they're extremely helpful there.
+
+The state of code completion engines is they're pretty good at all languages, with the caveat that they're probably not very good at COBOL, or Fortran... But all the sort of general languages, especially like Ruby, I'd expect them to be decent at. I suspect the world of code completion engines will get better at specific languages as people go deeper on the technology. It's a thing I continue to work on, and so I feel confident that it can be improved.
+
+The other place that I think most programmers could get value today if they're not a Go programmer is writing small, isolated pieces of code in a chat interface. So you could try out a ChatGPT or a Claude, or if you really want to have some fun, run a local model and ask it to solve problems. Try llama.cpp, try Ollama, try these various local products...
+
+Grab one of the really fun models...
+
+\[01:22:19.18\] It's especially easy to try on a Mac with a unified memory. If you're on a PC, you might have to find a model that fits in your GPU. But it's a ton of fun, and use it to say like "Write me a Ruby function that takes these parameters and produces this result." And I suspect the model will give you a pretty good result. So those are the places I would start, because those require the least amount of learning how to hold the model correctly, and you'll get the most benefit quickly.
+
+**Jerod Santo:** Love it. Good answer. I'm just wondering out loud, and feel free not to know, but when it comes to prompting - I know we're past the age of magical incantations, but as you guys have been building out a product, which is basically sophisticated prompting, are there guides that are useful, or they're like... I remember finding a site - I can't remember right now - there's like people just sharing their system prompts for certain things they do... Maybe there's like a Ruby prompting guide which makes it a little bit easier to get quality results out faster... Does either of you guys know?
+
+**David Crawshaw:** Yeah, I've seen people write guides like that. I would say the guides I've read are now out of date. Like we were saying earlier, guides go out of date. The thing I find most useful is to think of the model I'm talking to as someone who just joined the company. Sometimes I think of them as an intern, though every now and again the models produce much better code than I can... But interns have done that, too. That happens. And then as you're writing the question for it, imagine like - I'm talking to a smart person who knows nothing about what I'm doing, and they need some background. And that gets me really far with the current frontier models. So that would be my general piece of advice that I think applies to any programming language.
+
+**Adam Stacoviak:** I agree with that, too. The random "Just give me X" - you'll get a result, but you'll have to massage it further. It'll ask you more... I will often give it context; like you've said, this intern, this smart person that's new to context, they don't have the background awareness that you want somebody to have... And I'll often give it a lot of that, have a particular request, but then also say "Is there anything else I can give you, or any other information you need to give me a successful answer?" Just some version of like be successful with our goal.
+
+And it's strange even talking like that too, as I even say it out loud, "our goal", as if it's human... And Jerod, you know where we're standing with this. Jerod and I have some history; so I've been very kind, "please", and "thank you."
+
+**Jerod Santo:** He's very nice. He talks to it like we're on the same team, and stuff.
+
+**Adam Stacoviak:** If it gives me a great result, I say "Fantastic."
+
+**Jerod Santo:** High fives.
+
+**Adam Stacoviak:** Yeah, high fives. Why would I be any different?
+
+**Jerod Santo:** Do you offer it money?
+
+**Adam Stacoviak:** No, I haven't done that yet. I've got to try that. I've got to try that, honestly.
+
+**Jerod Santo:** \[laughs\]
+
+**Adam Stacoviak:** But I will ask, "Could you be more successful? Is there anything else I can give you, any more information I can give you to get us to our goal?" And I've found that that's it. Like, it's context. It's a full circumference of the problem set, as much as you can, that makes sense, and then you will have a more fruitful interaction. I will also say that I'm more exclusively using ChatGPT. And so the O1 model, while it's expensive, let's just say - I don't have the expensive plan... I still don't feel like I can be the person that spends 200 bucks a month on this thing. I would much rather buy a GPU than spend 200 bucks a month. Somehow that math makes more sense to me. But then O1's been pretty successful with thinking, and iterating, and more precision... Whereas 4.0 was a bit more brute. But that's just my personal take.
+
+**David Crawshaw:** \[01:26:23.19\] I think you might be onto something with being nice to models. I caught myself being pretty curt with models a few months back... And discussing this a lot with Josh, he mentioned -- the conclusion we came to was one of the challenges of not being nice to models is it sort of trains you to not be nice to people.
+
+**Adam Stacoviak:** Yeah.
+
+**David Crawshaw:** Because you're using all of the same tools... And so it might just be good for you to be nice to models.
+
+**Adam Stacoviak:** I mean, I just don't -- if it's humanistic, even, similar to a human, why not just be kind? Why not?
+
+**Jerod Santo:** I hate to break it to you, Adam, but this is not similar to a human.
+
+**Adam Stacoviak:** The iteration is. It is. It certainly is. If I were collaborating, if that was a human over there giving me the answers back, it would be very human volleyball iterative.
+
+**Jerod Santo:** If.
+
+**Adam Stacoviak:** If it was.
+
+**Jerod Santo:** Right.
+
+**Adam Stacoviak:** I get that it's not, but I'm also like -- like David, "Why not?"
+
+**David Crawshaw:** I think I'm somewhere between your two positions, because I do think it's just a machine, and it's just a tool...
+
+**Adam Stacoviak:** I don't think it's human. Don't let me think that.
+
+**Jerod Santo:** Well, you kind of just said that you think it is... \[laughs\]
+
+**Adam Stacoviak:** Did I? I mean, is that how you interpreted it?
+
+**Jerod Santo:** Yes.
+
+**Adam Stacoviak:** I just meant that if it -- maybe what I mean by that, just to be more clear, is kind of keying off what David said, which is like being kind. Just, why not? I don't know. I'm not overly like "Thank you so much. You're amazing."
+
+**Jerod Santo:** Because it takes time.
+
+**Adam Stacoviak:** I think it's just, I'm a kind person...
+
+**Jerod Santo:** Where you nice to Google when it got you a result in the search engine? Like, you just --
+
+**Adam Stacoviak:** Well, it is not a prompt where there's an ebb and a flow, or a back and a fourth. It's just simply returned an answer.
+
+**Jerod Santo:** Well, "How are we doing?" It asks you at the end, "How are we doing?" Do you tell it?
+
+**Adam Stacoviak:** Well, that being said, I'm not like "Thank you very much." I'm just --
+
+**Jerod Santo:** I'm just antagonizing him at this point...
+
+**Adam Stacoviak:** Yeah, I know you are. You're really digging into me. But I can catch myself saying "That's awesome" or "Great job." Or "I agree with that." Things, isms like that that you would say to another person.
+
+**Jerod Santo:** If that makes you feel more like a nice human, then you should just keep on doing that. But I don't think it's doing anything for the computer.
+
+**Adam Stacoviak:** I don't think it is either.
+
+**Jerod Santo:** It is actually costing resources from the world...
+
+**Adam Stacoviak:** \[laughs\] Listen to you... That's true, too.
+
+**David Crawshaw:** I mean, I think you're right, it doesn't help the computer.
+
+**Adam Stacoviak:** That's true, too.
+
+**David Crawshaw:** I say please and thank you to the models now, so that I remember to say please and thank you to humans. That's it. It's purely -- I don't want to get into the habit of...
+
+**Jerod Santo:** Yeah. You're training yourself.
+
+**David Crawshaw:** Exactly. It's all about training m.
+
+**Jerod Santo:** That's fair.
+
+**Adam Stacoviak:** Self-training. Yeah, I don't feel like I have to. I feel like it's just a natural Adamism. How I do things, how I operate.
+
+**Jerod Santo:** \[unintelligible 01:28:59.16\]
+
+**Adam Stacoviak:** It's who I am in my core. I'm a kind person.
+
+**Jerod Santo:** Well, David, thanks so much for joining us, man...
+
+**David Crawshaw:** Thanks for having me.
+
+**Jerod Santo:** ...and sharing all your knowledge. You've learned a lot, and I've learned a lot from you, o we appreciate your time.
+
+**David Crawshaw:** This was a ton of fun. Thanks for having me.

Reaching industrial economies of scale (Interview)_transcript.txt

@@ -0,0 +1,487 @@
+**Adam Stacoviak:** Can we go back as far as you'd to? Because I kind of want to know how we got here. I go to your homepage today, Beyang, and I see "Industrializing software development AI agents."
+
+**Beyang Liu:** Yeah.
+
+**Adam Stacoviak:** And that was not at all the headline, even a few years ago. And I feel Sourcegraph has taken 15 years to become, not so much successful, but -- it's been 15 years for you. And I would say seven of those years may have been really hard years. And I kind of want to just not let the people coming to Sourcegraph today think it's been the easy road for you. Because I kind of want to know the permutations of code intelligence to AI agents... How did we get here, for you?
+
+**Beyang Liu:** Yeah. Yeah, so first off, I think we're in year 12 right now. So we've definitely been around...
+
+**Adam Stacoviak:** Okay. I gave you three. I gave you three.
+
+**Beyang Liu:** Yeah, yeah. We're working our way up to 15. I think we'll be here for the next 50 years, so it's still early days. Actually, it's funny that you should mention industrialized software with agents and how that's kind of a shift... I don't know if you have show notes that you can link to, but I can link you to a version of our seed deck back in April, May 2013, when we were pitching the company for the first time... And it has this phrase, industrialized software engineering. And so that part of the mission has stayed largely constant.
+
+The high-level goal for us was really about basically making professional software engineering as enjoyable and as efficient as hacking on your side project is... That was really the motivator for us in starting this company. It was the delta between every programmer starts from a point of love or delight, at some point. The reason that you get into programming is there is this joy of creation, this spark of creation that everyone experiences at some point, whether it's at Hello World, or when you first get your first working program to run, and it's cool, and it does something, and you share it with your friends.
+
+I think everyone who's working as a programmer is, to some extent, trying to chase that original high; that's the dopamine rush that makes the job joyful. And it also maps to doing useful stuff. Like, you get joy out of shipping features that impact your users' lives, and actually get used.
+
+But then you contrast that with the day to day of a professional software developer, most of whom are working in a large existing codebase that they didn't write themselves, that is the product of the contributions of hundreds or thousands or tens of thousands of shared hands. And that experience is very different. And what we wanted to do is solve a lot of the problems that create toil for professional software engineers in large production codebases, and make it so that it's possible to focus on this creative part of the job.
+
+\[07:49\] So the initial way that we solved that with the technology landscape at the time was focusing on search, because that to us was the thing that we spent a lot of our time working on. We got our career started out at Palantir, but Palantir by extension meant very large banks and enterprise codebases, because that's the team that we were on. And so the core problem there was just figuring out what all the existing code does, and figure out what the change you need to make, how that fits into the broader picture, and what sort of dependency or constraints that you're working with. And so that was the first thing that we brought to market.
+
+AI was always sort of in the back of our minds. I had done a concentration in machine learning while I was at Stanford. I was actually -- Daphne Kohler was my advisor, and then published some research there in the domain of computer vision... In those days, the state of the art models weren't nearly as good. This was pre-ChatGPT, pre-Transformer, pre even deep neural net revolution. So in those days, the conventional wisdom was neural nets worked in the '80s for limited use cases, but they're mostly toy use cases, and real machine learning engineers use support vector machines, and boosted decision trees, and things that. So it was a very different world.
+
+We're always keeping our eye on the evolution of this technology, and we actually started integrating current large language model embeddings based signals into the search ranking starting in early 2022. And so that was something that we'd kind of been playing around with, and then once ChatGPT landed, we were actually pretty well situated in terms of our understanding of the technology to be able to roll that into our platform.
+
+We launched the first sort of context-aware chat, chat that pulls from the context of your broader codebase, and pages that into the context window and uses that to steer the generated code or the question answering that the LLM provides for you. And that helped launch us into the age of AI, so to speak, because that was -- I think now it's table stakes, right? Like, context-aware code generation. Everyone has that, because everyone realizes that that is absolutely essential for making AI useful in the enterprise. But we were first to market there, and that helped us establish our presence inside folks Palo Alto Networks, and Leidos is a big government contracting agency, and Booking.com, the largest travel site in the world... All these very large enterprises with complex coding needs that have adopted Sourcegraph and our AI coding system, Cody, as their preferred essentially platform for accelerating, automating, industrializing software development.
+
+**Adam Stacoviak:** When you say industrializing, what does that mean to you? I think I understand what that means. What does it mean to industrialize software engineering, practically?
+
+**Beyang Liu:** Yeah, that's a great question. This word industrialize, it's something that we've gone back and forth on, because --
+
+**Adam Stacoviak:** It sounds cool. I dig it.
+
+**Beyang Liu:** I mean, I it, too. I think if you go back to 2013, I think there were a lot of -- a lot of people read sort of negative connotations into it... Like, "What are you trying to do? Turn every development org into a feature factory?" Because that's what I think when I think of the Industrial Revolution.
+
+**Adam Stacoviak:** Sure, yeah.
+
+**Beyang Liu:** These days it's a little bit different, because the whole world is talking about reindustrializing, with all the geopolitical shifts going on in the world, the macroeconomic shifts... I think a lot of people are thinking about "How can we build the gigafactories of the future, and restart that process, that kind of tech tree exploration in the West?" So now I think it carries a very different connotation. But what we meant by it and what we mean by it today is really -- it started with this observation... When you think about software engineering as an industry, software engineering is actually quite unlike every other industry in the world, in one specific dimension, which is every other industry, as you scale production, you get economies of scale. Like, you build a larger factory, you can produce the next incremental thing that you're building at a cheaper cost, or more efficiently.
+
+\[12:01\] So things get better and cheaper as you scale up. In software engineering, it's the exact opposite. In fact, at best case, they asymptotically approach some upper bound. Worst case - and I think the reality in a lot of organizations - productivity goes down after you scale past a certain point. And as a consequence, every piece of software that becomes successful very quickly becomes a victim of its own success. Because if your software gains users and gains revenue, it will also acquire feature requests and bug reports... And as a consequence, you'll hire more developers, you'll increase the number of contributors to the codebase, the codebase will grow over time, you'll lose that cohesiveness of vision... And then it's a frog in boiling water approach. Every incremental day you don't really notice it, but one day you wake up and the codebase is a mess, there's all these constraints, it takes forever to get a simple feature across the finish line... You wake up and now you can't get anything done. And there's kind of two consequences to this that are very salient.
+
+One is that means that the vast majority of software that actually gets used day to day sucks... Because any -- again, any successful piece of software with a lot of users becomes this. So almost definitionally, any piece of software that you're using, you're using it because it was a successful piece of software. But by virtue of its success, it can no longer move quickly. The developers building it can no longer move quickly and build new things and keep that up to date.
+
+The second consequence for this is the disruption cycle actually occurs much more quickly in pure software than in any other industry. Every other industry, we talk about the innovators dilemma. It's like, okay, yeah, I get it, every once in a generation something comes along down the line that shakes the foundation of this very industry, and then you get a whole new generation of companies that emerge. In software, it's literally every 5 to 10 years that happens naturally, because whatever -- you can literally see, in the past five or six years alone we've gone through multiple generations of accounting software for small businesses. Every couple of years there's a new YC company that basically does the same thing, but it's "Oh, we're going to do it better, because the old thing sucks now."
+
+So this disruption cycle -- there's something more than the innovators dilemma in software disruption. It's this kind of phenomenon of as the codebase grows due to its success, it falls victim to the very forces that made it successful, because you're going to grow the number of lines of code, you're going to add more functionality, you're going to have a hard time preserving the cleanliness and sane architecture of the original codebase. And then at some point you can no longer make forward progress. And then some two-person startup comes down the line and does it a lot better, and then eats your lunch.
+
+**Adam Stacoviak:** Well, they eat your lunch for a bit and then you acquire them, or they totally eat your lunch and they destroy you, right?
+
+**Beyang Liu:** Yeah. Yeah, exactly.
+
+**Adam Stacoviak:** Because that's sometimes what happens, too.
+
+**Beyang Liu:** Exactly. So that's what we mean by industrialized software engineering. I think if we can tackle enough of the toil, and also give your senior technical leadership a lever to project their vision for how the architecture should evolve, we can actually get software as an industry to be a real industry. actually have economies of scale, where things can actually get more efficient. Imagine that. It's mind blowing. It's like, oh, things actually get more efficient as the codebase grows? I think that's just very antithetical to the way people think right now, but I think it's certainly within the realm of possibility, given modern AI in conjunction with all the other code-based understanding technologies that we built over the years.
+
+**Adam Stacoviak:** \[15:57\] So given your history with machine learning way before it was artificial intelligence or AI, way back, and you're here now... You've iterated, you've sort of stuck to your gun, so to speak, of industrializing, or solving this complex codebase problem from the beginning, with search now to agents with AI... How did you know? What tea leaves did you read? Or were you just super-passionate because of your time at Palantir? Because they obviously had complex codebases, and you were "Well, this sucks. \[unintelligible 00:16:35.13\] solve this problem. I'll just keep solving it over and over and over until we reach critical mass."
+
+**Beyang Liu:** Yeah, I would say -- I'm not going to say that I have the ability to kind of see the future. I'm definitely not going to claim that I could see how the landscape of AI would unfold from the vantage point of being in 2013. There's certainly a lot here that frankly I just did not anticipate.
+
+**Adam Stacoviak:** That's 12 ago, so...
+
+**Beyang Liu:** 12 years ago is a lifetime, and the technology was very immature and different back then. But what I did know, I think Quinn and I both felt deeply was, number one, we're both really, really passionate about programming and building software... That seemed to us the biggest leverage point. And it was really a combination of "Hey, this is a craft that I care deeply about, and I could see myself spending the rest of my life refining it, and moving it forward", because it does seem -- it almost rises the level of being something spiritually fulfilling. If you want to get philosophical, there's this whole it from bit philosophy of "What does it mean to exist?" and "How does information relate to physical being?" and all that. And you can get into the whole discussion of Turing completeness, what does it mean, how does it relate to our theory of mind, and the soul... Anyways, there's kind of --
+
+**Adam Stacoviak:** Do you have thoughts on that you want to share?
+
+**Beyang Liu:** We can get into it... I don't want to go down that rabbit hole just yet. But the point I want to get across here is that we both feel that this is a really cool thing to be working in. And that's one of the things that we chose this path as the professional path that we want to pursue.
+
+The flip side of that was our experience at Palantir showed us that the pain that we felt in trying to practice this craft in a professional setting is felt across the entire industry, and it gets exponentially more critical the more impactful your software is. Because the more impact your software is driving, the more users and revenue it has, or it's generating, and the messier the codebase is. So it was kind of the marriage of these two things. It was personally fulfilling to be thinking about advancing the state of this craft that you feel a deep personal connection to, and then knowing that it's tied to just this huge amount of potential economic impact.
+
+And so there, our thinking was like with the technology that we have today, if we focus on the fundamentals, building core technical capabilities, and then rolling those into delightful product experiences, there will be a huge company here that we could spend the rest of our lives working on.
+
+**Adam Stacoviak:** Fun. I think it's fun, man. Wow. To have seen - not the future, because you said you're not a clairvoyant, you're not a future teller...
+
+**Beyang Liu:** Yeah.
+
+**Adam Stacoviak:** ...but you can certainly see the direction it's going. And you said, "I'm willing to keep doing this problem, keep solving this problem, for the rest of my career."
+
+**Beyang Liu:** Yeah.
+
+**Adam Stacoviak:** \[19:44\] Wild. Okay. And now we're at the age of literally AI agents... We can essentially deploy these agents similar to the way you would deploy a cron job, but now they're smart. They're so much smarter now, not just a dumb cron job that requires logging, and metrics, and observability, and failures, and fallovers, and system crashes... Now you've got agents that can actually do some cool stuff. You began with search, now you're here... What do you think it is that is attracting folks to the Sourcegraph platform today? Is it the promise of what AI can do deployed at scale?
+
+**Beyang Liu:** I think there's a couple of different dimensions here. I think the most direct thing that attracts people is that we are solving for the challenges that developers face inside production codebases. So contrary to a lot of the other AI coding things that you see, our focus has less been on the indie developer who wants to show off a cool Flappy Bird clone on Twitter... Which, to be frank, it's really, really cool.
+
+**Adam Stacoviak:** So cool.
+
+**Beyang Liu:** It's really cool that you can one-shot that now. But our mission, the thing that gets us out of bed every morning is really to enable the software that drives the world to be built much more quickly and much more robustly. And so that's the thing that we're targeting, and that's the reason why the customers that have chosen Sourcegraph have chosen us... Because they have large, complex codebases and they appreciate the other technologies that we bring to the table, that help us be effective in that setting.
+
+And that brings me to this kind of second dimension, which is - our technical strategy here is transformers, LLMs, AI... That's all really useful, it's game-changing technology, but we think that there is a step function more capability and power you can unlock by combining that technology with some of the other technologies that we've built around information retrieval, code search, mapping out the reference graph and the knowledge graph of the code... We think these technologies are highly complementary. In some sense, they represent the two sides of what traditionally has been called artificial intelligence. Before the current incarnation of what AI means, there was this classic debate between the formal methods, like context-free grammar, Chomsky kind of party of artificial intelligence, where everything's kind of rule-based, and there's some philosophy behind it that these grammars emerged from the mind of maybe some higher order being in our goals to discover the platonic ideal, the hidden structures that connect all these things... You know, very symbolic-based.
+
+And then on the other end is the statistical approach, which is like we're not going to be opinionated about defining down a complete set of axioms or rules for describing what language is; we're just going to try to model it statistically and learn it from the data. I think the age we're living in, everyone sees clearly the limitations of the Chomsky approach, and it's very apparent to everyone the advantages of the data-driven approach now, with modern technology. But I think there's still actually highly complimentary. There's a bunch of productivity-oriented use cases that essentially combine a knowledge graph or some sort of symbolic representation with generative models. So you have the generative model providing some notion of reasoning, but you combine that with a symbolic representation to be very precise about how things are connected. And the composition of those two yields way more useful use cases than if you were to just use one or the other.
+
+**Adam Stacoviak:** Maybe, for context, give me, if you can, give me a picture, if you can paint a picture of a day in the life of an engineer working at one of these large-scale enterprises that has this complex codebase issue, and you're removing their toil. What are the things that they have access to as a developer in this kind of ecosystem? Let's say Palo Alto Networks, for example; you mentioned them... I know they're a logo on your homepage, so maybe you can use them... Pick anybody. I don't care if it's them or not. Somebody. Give me an example of a developer or an engineer on a team like that. What kind of tooling do you give them or do they have access to when they buy Sourcegraph, essentially?
+
+**Beyang Liu:** \[24:24\] Yeah, so the equivalent of a day in the life for an engineer is what does it take to take a feature from idea all the way to getting into production? It's like one whole loop of the software development lifecycle, which we kind of model as this two loops that are connected. There's kind of an inner loop, which is your kind of read/evaluate print loop in your editor, where you're kind of quickly iterating on the idea, and making it come together... And then there's the outer loop, which is sort of like the ops loop. You plan out the feature, you actually write it, you push it up, it gets reviewed, it needs to pass CI/CD, it gets rolled out, and then after it gets deployed, you've got to monitor and observe it to deal with any kind of production bugs or things that emerge.
+
+So a day in the life of an enterprise developer or a developer working in the context of a very large existing codebase, it's kind of like you have an idea for a feature - what's the first thing you have to do? Well, the first thing you have to do is you have to go and acquire a bunch of context about the existing codebase. So in the pre-Sourcegraph world, what does that entail? It's a bunch of grepping for stuff in your existing repository, maybe you've got to clone down a couple of additional repositories if you're very intentional about it... Sometimes people just don't bother, because they're like "Ah, it's too much work. I will just assume that this thing that I'm building doesn't exist yet, so let me just go build it." Sometimes it involves pinging your colleagues for help. So if you're onboarding to a new part of the codebase, or you're a new engineer, you're going to go bug the senior engineer who was around when that code was written, who's very short on time, and kind of crotchety because there's like five other people you who have asked them a very similar question in the past two weeks... So that is a whole kind of tooth-pulling exercise. And that can take you, I don't know, weeks, even months... In some of the codebases that we worked in while we were at Palantir, it'd literally take months just to orient on "Hey, this is the part of the code that actually needs modifying." And there's a bunch of false starts along the way, because you start writing something only to realize "Oh, you should be modifying this other piece over there."
+
+Okay, so that's just getting to the point where you're ready to start writing the code, right? Once you start writing the code, you're in your editor, most of the code you need to write is probably boilerplate, or what we call boilerplate plus plus. It's not exactly something you can copy and paste, but you basically want to find an existing example of using that one particular API that you need to use. And then you'd kind of turn your brain half off and just pattern-match against that example, because that's the reference point. So provided you've found that example, then you can pattern-match against it. If you don't find that example, then that's kind of this arduous "Hey, let me discover how to properly use this API, because there's no docs on how to use it." And someone's probably done this probably like a dozen times before, but I've never done it before, so now I have to essentially go and rediscover how to do that.
+
+So you go through that exercise, there's multiple rounds of kind of iterating on it, and you finally get to the point where "Okay, I'm ready to push it up." You push it up to review... Maybe there's multiple stakeholders that have to sign off on on the change. There's the people you work directly with, maybe that's part of your code review process, maybe you also made some modifications to some other shared libraries, and now other teams that own that code have to also sign off on it...
+
+\[27:57\] A lot of times those other teams don't necessarily care too much about the change that you're trying to land, because they have their own priorities, and they're incentivized to work on those, and not anything else that happens in the company. So who knows how long it'll take to get all the stakeholders to sign off on that change. Or maybe you realize through the review that you get from a senior engineer that this whole approach was wrong, because you didn't do enough search and discovery.
+
+So it's almost like this game of chutes and ladders. You have a potential to shoot all the way back to scratch; you just wasted a week, a month's worth of work because he built the wrong thing. Best case, it takes a lot of time getting people to approve it, you finally get it approved, it rolls out to production, maybe it breaks something, maybe there's a test that breaks that you didn't catch locally, and so on and so forth... Once it gets into production, there's a long tail of issues that could arise too, in terms of it triggered some bug, or maybe there was some security vulnerability that got introduced... Anyways, I don't have to sketch this out, it's already painful just describing this.
+
+**Adam Stacoviak:** I wanna quit right now. I'm like "Gosh... Save me."
+
+**Beyang Liu:** Yeah. And this is just a simple feature. It could be the simplest thing, like center this div, or add support for this authentication protocol, or whatever. And most of your time is spent not on thinking about the end user problem and mapping that to the algorithms and data structures that you can use... Like, that's the fun part of the job. If I could do that all day, every day, I would do that all day, every day. But most of it is spent on box checking, writing boilerplate, finding answers to questions that you know have already been answered many times before...
+
+And so Sourcegraph essentially targets all the kind of critical pain points and slowdowns in that process. So in the planning and search and discovery phase, that's where people use a lot of our code search functionality; or now we kind of have a deep research for your codebase feature that's built into the web application, where you can ask our AI a question about the repository, it will go and do a bunch of searches and file readings on your behalf, and print out what is usually a pretty reasonable summary of the relevant files to read, and how those fit to the tasks that you're trying to accomplish.
+
+So automatically, that shortcuts you from having to have and wait on all these conversations with humans that -- let's be honest, developers as a collective group, we're not the most personable or affable people. I mean, many of us are, but speaking in terms of averages, we tend towards introversion and spending time with our computers. That's why we're doing what we do, right? Plus, even the ones that want to help, if you've answered the same question from three or four different people in the past week, you're at the point where you're like "I just want to do my freaking work. I don't want to have to deal with any of these other things", even if they're reasonable requests. So we can shortcut that process and just help you figure that out yourself through a combination of search and question answering.
+
+Then within the editor there's all this stuff that reduces to boilerplate plus plus, which is like in the sweet spot for what AI can do. So you tag in the relevant file, or you ask it to do a search for the relevant code snippets on your behalf, and then ask it to generate code, it generates the code, you can refine that in the chat window, apply it into the file... And that is just such a -- it adds value in a couple different ways.
+
+The first is the raw time saved, because you don't have to manually type out all the characters for that boilerplate plus plus. The second thing we've noticed is that it kind of changes the psychology a little bit, because it keeps you in flow state more. If you don't have to page out all the working state in your brain, because now you need to go down this rabbit hole and complete this side quest to get your feature built, if Cody can essentially do that for you, or even get you 90% of the way there, you essentially don't have to page out the context for doing the main thing. As a consequence, you can kind of remain in flow state for much longer; you don't have this context switching that is very destructive. So that is where we probably see an insane amount of productivity boost.
+
+\[32:17\] I think the number that Palo Alto quoted at us, using their own kind of internal metric or proxy for developer velocity - depending on which number you use specifically, it's between 40% and 57%, which is a crazy speed-up; that's a game-changing speed up. And it turns out it just makes the job more fun too, because you don't want to spend all your hours writing boilerplate or pattern-matching against stuff that's already written. You want to get to the fun stuff of the job, which is really thinking about the user experience, and then the underlying algorithms and data structures.
+
+Once you push things up to review, what we built in collaboration with a couple of our customers and is now in early access - and by the time this is released, we'll have open access for it - is a review agent. So this is code review that kind of understands how your organization does code review. So the idea here is, one, we want to automate the process of code review, because there's a bunch of comments that humans don't think to leave, or just, they don't get around to leaving. There's the old adage, if you put a push up a 10-line pull request, you get 10 comments. But if you push up a 10,000-line PR, you'll just get a rubber stamp "Looks good to me", because no human...
+
+**Adam Stacoviak:** It's too much.
+
+**Beyang Liu:** Yeah, it's too much. It's like "Well, what else am I supposed to do today, other than review this PR? I've got a life to live, I've got work to do." Well, an AI can just sit there and spend the time to surface a bunch of common things that you as a human, it's probably a waste of your time to check. All this stuff is automatable now with modern LLMs. So we want to do that.
+
+There's also kind of a psychological component here, where we don't see our code review agent as being a complete replacement for human review yet, but it's at the point where if it leaves at least a first pass of comments, number one it allows you to you as the author of the change to kind of iterate quickly against that, and also it incentivizes the reviewer to go in and do a more thorough review, too... Because the comments kind of reflect a summary of the change itself. And just having a comment exist on the PR, it just makes the task seem already halfway done.
+
+**Adam Stacoviak:** Momentum.
+
+**Beyang Liu:** Yeah, momentum. It leads people to quicken the review cycle, too.
+
+**Adam Stacoviak:** Psychology, baby. I love that.
+
+**Beyang Liu:** Yeah. It's a combination of like if you introduce efficiencies, you also get these nice second-order psychological effects that further speed up the development cycle.
+
+**Adam Stacoviak:** Right. Especially after a few cycles of this, you begin to gain some trust for the system, let's just say. I don't want to anthropomorphize this thing, but we tend to. Sometimes we even say "please" or "thank you" to our agents... Which is cool, whatever; you do your thing. You're just wasting tokens, basically...
+
+**Beyang Liu:** Yeah.
+
+**Adam Stacoviak:** But you begin to trust it. So it's like "Well, I at least know I've got to start somewhere." And so this analysis of the code review makes me feel like, even if I only had to do 10%, it's better than 100%, because now I can actually do my job today, I can do the two meetings I have planned, and actually potentially get a review in to potentially get this into production. And that's a great thing.
+
+**Beyang Liu:** Totally. It provides -- it's almost like a smoother on-ramp. There's like an activation threshold you have to get beyond to get started on a particular task, like in this case code review... And without AI, it's like "Oh, I'm dreading this, because I need to get over that threshold." Now we can do the first one or two steps, and then that kind of leads you into -- it's the same thing as how some writers, they say making a mark on the page, writing a...
+
+**Adam Stacoviak:** Anything.
+
+**Beyang Liu:** Yeah. Anything.
+
+**Adam Stacoviak:** A date...
+
+**Beyang Liu:** \[36:10\] A very bad first draft, that's just like god-awful... Or if an artist's just drawing a squiggle on the paper, it kind of unblocks you. It gets you past this blank slate mental block that I think a lot of people suffer from, especially if it's a task that you associate with toil, which I think a lot of people do.
+
+**Adam Stacoviak:** Daunting, yeah. It's like "I can't even do this." I've been there. I've definitely had to do some code reviews and I was like "Yeah, rubber-stamp that thing. It looks good."
+
+**Beyang Liu:** Yeah, yeah.
+
+**Adam Stacoviak:** You know, "Let's just -- let's test in production." Right?
+
+**Beyang Liu:** Yeah. It's like, "You wrote a lot of code. I think you probably understand this much better than I do. Let's just -- we'll say it looks good, and then we can both get on with our lives." I think some version of that conversation happens all the time inside enterprises, and it happens out of necessity, but it's not good that it happens... Because you have reviewing standards for a reason, and that's how bugs and security vulnerabilities and bad architecture and loss of cohesion happen.
+
+**Break**: \[37:17\]
+
+**Adam Stacoviak:** So is that where -- I think we're here because we talked about the day in the life of... Where do we go from after code review? You're painting the landscape of Sourcegraph, what an engineer may have access to to reduce or remove some of the toil in the process.
+
+**Beyang Liu:** Yup. So broadly speaking, I think where the day in the life of a developer moves to is -- yeah, I talked about inner loop and outer loop before. Our vision for these two loops is essentially like accelerate the inner loop. We should provide facilities that automate boilerplate and toil inside the inner loop, such that it feels like we're building leverage around human creativity... Because that is still the thing that is, quote-unquote, out of distribution of the models, and something that is not yet replicable with today's LLMs. And I don't see a clear line of sight to actually replicating that. I think that is truly the essential part of the job. That is the fun part of the job. It's like thinking up a new idea, connecting that with a user need, and then finding just the ideal, best data structures and algorithms that fit within your existing architecture to solve that. That is the essence of the job, and that's the thing that we want to enable people to spend 99% of their time on, as opposed to 1% of their time on.
+
+So we're accelerating in the inner loop... And then the outer loop, I think is -- we want to automate as much of that as possible. Because everything in the auto loop is kind of like a factory. You spend all this time refining the creative aspects in the inner loop, and now the outer loop is just like "How do I get this thing that I have produced in my development environment into production?"
+
+So the whole process of doing the review, and refining it, and making sure there's adequate test coverage, doing an automatic rollback, adding the necessary telemetry... In ideal world, that is all within the realm of automation. It's kind of like we built this Gigafactory, in a way, where if you got it working in your development environment, everything else to get it into production should be pattern-matchable by the LLM.
+
+So in the future -- I think the future that we want to work towards is it doesn't matter what size codebase you're working in; it could be the smallest thing, or it could be the largest, oldest codebase in the world... You express your intent, we accelerate you getting to the experience that you want in your development environment, and then thereafter the machine just takes your change and lands it in production, guaranteeing security, reliability, and robustness.
+
+**Adam Stacoviak:** Love those buzzwords, love those buzzwords. I'm being serious. I'm seeing this, where you're going... I was just half-joking there. That was some buzzword city there, though. Robustness.
+
+**Beyang Liu:** Well, that's what you want, right?
+
+**Adam Stacoviak:** You do want those things.
+
+**Beyang Liu:** Our tagline is "Move fast, but don't break."
+
+**Adam Stacoviak:** Okay, sure. I like that, too.
+
+**Beyang Liu:** And that is now possible. With today's technology, it is possible to enable an extremely fast development cycle, that augments human creativity. But the things that could break your your system, like a security vulnerability, or too much technical debt, or a bug that wasn't caught because you didn't think to add the appropriate test... All these things are in the realm of automation now. They are automatable, we've just got to go and automate them.
+
+**Adam Stacoviak:** \[44:09\] For sure. And you know what? I'm all for that, because there's scenarios where developers want to deploy code, and they'll forget "Oh, I didn't write or consider this one security thing that I should know about. But because I'm a human, I've got limited even emotional will, willpower", which is a finite resource to any human being. We all have brains, and those brains give us a certain amount of power. Regardless of how smart you are, you could have not had your sneakers that day, so maybe you're Joe Pesci for a moment or two, and you're just not thinking straight. Or it's four o'clock and you're trying to get this thing over the line, because - time. And you don't write the test, or you forget about a critical thing.
+
+We should have automations and checks and balances in our flow from the REPL, which you were talking about, this inner loop, which says "Okay, now it's out of my safe world. It's out of my safe idea world. My innovation world. I've tested it, I've done all the things I can." Now it goes out to the outer loop, which - that's where automation should live, because you want to confirm "Is it a threat or is it an ally?" Is this thing that has been done a threat or an ally? And if it's an ally, to prod. If it's a threat, let's check those threats.
+
+**Beyang Liu:** Yup. That's exactly right. And I think this willpower is a finite resource. That is a really good way of phrasing it. I think I'm going to steal that. We've seen it play out not just on the timeline of a day in the life, but we've seen it play out on the span of months and years.
+
+To give an example, we have a customer where we partnered with them to build a sort of code migration agent that was targeting a project to retire all these dead feature flags that they had strewn about the codebase. So feature flags, you add them to do some experiment, and then once the experiment is over, you decide "Okay, we're going to either turn it to a hundred or turn it to zero", but then no one is really incentivized to remove the feature flag from the source code. So over time, these things kind of grow linearly and they add to the complexity of the codebase, because no one is quite sure just looking at the source whether it's fully turned on or not.
+
+**Adam Stacoviak:** Right. Will it break then...?
+
+**Beyang Liu:** Yeah, will it break? Which one of these branches actually execute? ...depending on what feature flag configuration I have currently in production. So they had a migration project... The target date for the project was 10 years down the line.
+
+**Adam Stacoviak:** Gosh...
+
+**Beyang Liu:** And literally, this is the sort of thing where the only people with the context to lead this conversation are the very most senior engineers in your organization, because they are the only ones who know the entire codebase, they know all the nuances of how these feature flags can break if you remove them improperly... And they're the only ones you trust to initiate a migration that literally touches every single repository in your codebase.
+
+So the joke is -- if you ever read Kerouac, "I saw the best and brightest minds of my generation." The developer equivalent is "I saw the best and brightest developer minds of my generation go insane, because they had to spend 10 years of their lives working on a dead feature flag removal." Istead of anything interesting or fun, like building new features for your users or customers.
+
+**Adam Stacoviak:** Stuff that really matters, basically.
+
+**Beyang Liu:** Yeah, exactly. So when we ran that project with them - it's still ongoing, so it's not through to completion yet, but we're basically able to say "80% of these feature flag sites, you can automate within the next month or so." So it's kind of like -- and sitting down and just figuring out how to compose, build an agent to do this migration... We've gotten you within the realm of retiring 80% of these within essentially a month. And then the remaining 20%, they're a little bit trickier, and those will probably involve some amount of human in the loop... But already, we got 80% of the way there within a month. This thing is not going to take 10 years any longer.
+
+\[48:19\] And just think of the -- that's a good chunk of your life, right? For multiple people, for multiple really, really smart people, who can now go and focus on other things that actually move the needle for the business and for the user experience.
+
+**Adam Stacoviak:** Yeah, that really is the best use case really for automation, or intelligent automation... Because one thing I like about the idea of, I would say using today's newest frontier models, or generation, or reasoning in problems like that is because it can run scenarios. If it's sophisticated enough, it can say, "Run the scenarios if these five flags were removed. What would happen?" Because it has awareness of the database, it has awareness of its scalability of it, what the database actually is, or what the language might be, or how it's sharded or not sharded... Or whatever the scenario is for the application, it knows that and it can, to some degree, quantify, or try to quantify some of that reasoning... So that when you get there, the PR, you're mostly there; instead of having to go 100% of the way, maybe you only go 50%, or in this case 80%. You've gotten it 80% done, and you only have to deal with 20%. So hey, smart fellas, smart folks, go and tackle this problem. Now you only have to do 20% of the work, versus 100%. I would be so much happier if I were that intelligent person, having to deal with that issue. Because it's important, but not so important that you should spend 10 years of your life doing it.
+
+**Beyang Liu:** Yeah. And the remaining 20% is the fun 20%. That's the interesting problems, the ones where you can actually wear your computer science hat, and God forbid, actually make use of some of the things that you learned in college.
+
+**Adam Stacoviak:** Yeah. Okay, so that's one more layer of the day in the life. This is a long day in the life. Wow. Code migration, code review... What's left? What's left in this day in the life, this outer loop that you're going to automate? Is there more?
+
+**Beyang Liu:** You know, I think we've covered most of it. There's specific pieces that we can get into, but broadly speaking, it's accelerate the inner loop through an amazing in editor AI experience - help automate the boilerplate, help automate the grunt work of doing simple things - and then in the outer loop fully automate as much as we can. So I think the last remaining piece here would probably be some sort of automated remediation of production incidents that happen. And I think there --
+
+**Adam Stacoviak:** So you want to automate incidents now, too.
+
+**Beyang Liu:** Yeah. I think the entire software development lifecycle -- we essentially want to be the platform that performs all of that. We're essentially building an agent construction mechanism into Sourcegraph, where the code review agent is just kind of patient zero for that. There'll be like a first-party agent that is customizable and composable, so that people can tune it to the specific needs of their enterprise... But that's just the beginning of the long and fat tail of toil that exists inside enterprises. So the common points here will probably be things like code review, test generation, issue and incident remediation... Those probably are the most salient things for us. But then inside every org, there's kind of a long tail of very specific things that you need to do for that organization. Like, if you're a big bank, you might have compliance requirements, or checks for proper handling of PII. If you're in healthcare, there's a lot you have to do around data privacy, and making sure that doesn't get leaked to a part of the system.
+
+\[52:17\] Every industry has its own set of things that they needed to enforce in this life cycle, and so the way we've approached this automation problem is we're not trying to build a one size fits all thing that just does code review. What we really want to build is an agent authoring platform to enable our customers to go build the automations they need to build to address the toil that exists in their organization.
+
+**Adam Stacoviak:** You're the runtime, they still write the smarts to it. Interesting.
+
+**Beyang Liu:** Yeah, we're the runtime, we provide the common Lego bricks and building blocks, like a common inference point, a common context fetcher, a common way of composing these agentic chains... And then their developers are essentially able to go and with very low friction, assemble these blocks into the things that target the toil that they experience.
+
+**Adam Stacoviak:** I love that. I kind of joked earlier about smart cron jobs. Do you feel that's kind of what this is? I mean, maybe at some level there's some things you do with cron jobs where you script something and you automate it based on certain timing, and it runs, and then you observe it, and "Did it fail, or did it pass?" and there's logs, and there's obviously the effects of the change that you construct... Is there a day where we will no longer write simple, let's just call them dumb cron jobs, and we'll write smart ones that are very agent-like?
+
+**Beyang Liu:** I think there will be. And I think the point that we want to get to -- the way we've constructed our code review agent is to make it so that for the most part you don't have to think about the imperative logic of the review agent itself. What you can do with Sourcegraph is you can define these rules that should hold in a certain part of your codebase.
+
+So a rule is just a declaration of some invariant that should hold across a certain set of files in your codebase. So maybe it's something like -- it could be as simple as "Hey, if you are writing the name of your company, make sure to camel-case appropriately", because a common misspelling is, you know, people forget the camel case, and that's not good. Or it could be much more complex than that, like "Hey, here's a complete style guide for writing code in Go. We've copy and pasted the open source Google style guide, because that's treated as canonical. Make sure all the Go code in the organization follows this style guide to the best of your ability."
+
+Almost every sort of invariant that you want to describe is describable in a statement that, and we want to give people the power to define these rules in one place, and then have them be enforced across their codebase. And the mechanism of enforcement, there will be at least three places where this surfaces. One is inside the editor. So make the rules part of the context that guides the generation of code inside the editor, so that you feel confident that if most of your developers are vibing their code into existence, that vibed code actually follows your organizational standards. So it's not going to be like -- you're not going to have to rework 90% of it because it's using some framework that you don't want to use; you want to use this other framework or whatnot.
+
+\[55:48\] The second layer is code reviews. So at review time, if anything falls through the cracks or if someone wrote that code manually, enforce these rules to make sure that nothing gets committed that causes any of these rules to break. And then the last thing is what you're describing with the cron jobs. You should just have a multitude of background agents in flight at any given point in time, just constantly crawling your codebase to ensure that these rules hold. And if they don't, if there's ever a point where they don't hold, push up a PR and then automate the review of that PR so that a human can come in and be like "Okay, yup, yup, yup. You checked the right things. This looks good, and you've added the appropriate unit test to make sure that nothing breaks. Looks good to me", to ensure that those invariants hold.
+
+I think the combination of these three methods of enforcement will allow your technical leadership, the people who carry the architectural vision of the codebase, to really enforce consistency and coherency across the entire codebase. Not just catching the typo things, but you can actually describe architectural constraints, like "This component, if it talks to this component, should go through this interface", because that's the intention of the architect. It's very important that we maintain this interface boundary, because that allows us to keep the code clean and changeable moving forward.
+
+That's the sort of thing where without this system, without AI, it would have been a losing battle... Because you only have a certain number of senior developer hours in the day, and at some point, some PR is going to sneak past the reviewer, because they've got other stuff to do. And then at that point you've broken the invariant, and then things slowly decay over time. And that to us, to get back to the industrialization point, that's a linchpin of being able to make coding a truly industrial craft.
+
+Another point of reference for us is - if you've ever read The Mythical Man Month, that's the classic...
+
+**Adam Stacoviak:** I have not.
+
+**Beyang Liu:** Oh, you haven't. Oh, you've got to read it.
+
+**Adam Stacoviak:** Oh, man. I'll put it on my list.
+
+**Beyang Liu:** The book was written in the '70s, about challenges that developers faced building mainframe software, but it's 98% applicable to the way development currently works today... Which is - the thing that holds software back and prevents these actual economies of scale from being realized is the loss of architectural vision and coherency that occurs when you scale up the size of the team working on the software and contributing to the codebase. And so this is a perennial problem that has plagued software development since the mainframe software of the '60s and '70s, which this book was written about. But the same general principles apply today. With our system, this system of rules and invariance enforced through AI at different parts of the SDLC, our mission, our vision here is to give the architect of your codebase a lever by which they can enforce the constraints and architectural rules that must hold in the codebase to preserve the overall kind of cleanliness and design of the codebase. Now, for the first time in history - I'm not saying it's a solved problem; we're still in the process of solving it. But now for the first time in history, we have line of sight to solving that. And I think if you're playing around with these technologies and using them day to day, you can see how we get from here to the point where people can define these rules in one place and just constrain and control the evolution of the architecture as code grows. If we can do that, then we basically solve the fundamental problem of the mythical man month, and that ushers us into a new age of software development. We're literally transitioning from cottage industry or artisanal craft into actual industrial processes, where we do have economies of scale and we do have efficiencies where the team can actually, not even not get slower, but actually get faster the larger the codebase becomes.
+
+**Adam Stacoviak:** \[01:00:01.06\] Yeah. I'm seeing this world. I mean, are we there yet? Let's just -- let's stop for one second. Are we there yet? Not quite.
+
+**Beyang Liu:** No, not quite. And we're in the midst of building up --
+
+**Adam Stacoviak:** What's required to get there. Okay.
+
+**Beyang Liu:** For us, it's really about building these factories in collaboration with our partners. So one of the things that differentiates us is, first of all, we're thinking about this problem at all. And second of all, we can sit down with the technical leadership at the booking.coms of the world or the Palo Alto networks of the world and make this system work well for them... Because building the Gigafactory is very different than building something that works inside a small codebase with a couple of contributors. And I think the low-hanging fruit is definitely the small codebases, which is great. I'm glad there's a bunch of people working on that and making that better. But I think where we see a strong alignment with our mission and what we want to do in the world is really enabling software development to go much more smoothly and quickly inside these large, industrial scale codebases.
+
+**Adam Stacoviak:** Well, the picture it seems you've painted - if I can just analyze a little bit - is if I'm an IC in any of these organizations, and like you'd mentioned, there's a certain interface I may have to write code to, it always has to go through that way, so the agent that may confirm that that's a truth or not for my codebase.
+
+**Beyang Liu:** Yeah.
+
+**Adam Stacoviak:** When I deploy code, if I can do my REPL, my inner loop, my creativity, and I have insurances outside of it, like, no matter what I throw out there outside of my circle of trust, let's say within my own brain or my own team's brain of innovation, creativity, moving the ball forward, if I know whatever bounces back holds true against rules that my organization trusts, holds near and dear, is orchestrated and designed by the most thoughtful people... If I can be creative within confines -- it's like, you still have constraints, but now it's not rogue, because you can add more and more smaller teams to be creative, and to throw things outside their circle of trust, and get bounced back against this wall, so to speak.
+
+**Beyang Liu:** Yeah. And the robots should be able to handle somewhere between 90% to 100% of the back and forth. Because before, this would be very painful, because you'd push up your new feature, you're all happy about it, but then some architects would be like "No, it violates this constraint. This part is not clean." You have to go and do a bunch of rework to fix it, and then you push it back up again, there's more comments... And this process just drags on, and both sides get increasingly frustrated, because finite willpower, at the end of the day. But now you can have the robots handle the vast majority of the back and forth, and so... Yeah, now you focus on the fun parts.
+
+**Adam Stacoviak:** Worst case, you're mad at Cody, or whatever you name the next robot, right?
+
+**Beyang Liu:** Yeah. And if you get mad at Cody, the brilliant thing is that - hey, just tag in Cody to go address Cody's comments, right?
+
+**Adam Stacoviak:** Nice. Yeah. You'd mentioned in the pre-call specifically some things you do for enterprises and government agencies... What is it that those -- I mean, we probably talked about some of it, but what is it that those organizations need that has only been uniquely designed and engineered by Sourcegraph to help them do better?
+
+**Beyang Liu:** Yeah, so there's two things here. One is the context awareness, building that in as a first-class citizen. I mentioned that we were first to market for that, and we actually have a way of delivering that really effectively into the enterprise. Sourcegraph has been self-hostable for many years now, because a lot of our early customers were very large enterprises. And so we've built all this DNA around delivering the best context for human developers in the form of code search and code navigation into all these different deployment environments... And we can essentially offer those same context fetchers as a kind of game-changing tool or capability for LLMs, as well; any sort of AI code generation or question asking.
+
+\[01:04:01.08\] The other facet of this is just the security and compliance requirements that a lot of these larger enterprises have to comply with, or are constrained by. So in terms of being able to deliver something into a purely self-hostable environment, I think we're basically the leader in being able to do this. Even Microsoft, or Copilot, with all their enterprise branding, there's still components of their infrastructure that have to be tied to their Claude... And that essentially rules them out for a lot of, frankly, the largest software engineering organizations in the world. They're ones that you haven't heard of, but they're literally powering half the economy, and they have very stringent security and data privacy constraints.
+
+**Adam Stacoviak:** So you can self-host everything we talked about today.
+
+**Beyang Liu:** Yes, yes.
+
+**Adam Stacoviak:** And it's by design; that's not obviously by accident, because you wouldn't do it by accident, because that's very hard. It's very hard to self-host.
+
+**Beyang Liu:** Yeah, yeah.
+
+**Adam Stacoviak:** So does that make it challenging to build Sourcegraph, to sort of have two deployment targets? Or I guess infinite, really, technically... Because you've got your own Claude, I'm sure, which - I can go there and be a Claude customer, you're hosting it, versus my own hosting.
+
+**Beyang Liu:** It's absolutely a challenge and it's something that has slowed us down a bit. The in-editor experience, I think, we've had a bit more challenge in pushing that forward than we would have... Like, we ideally would have liked to move faster on the in-editor UX, but because we're targeting a variety of different editors and deployment constraints, that has constrained what we've been able to do there, to a certain extent. But I think now we're at the point where we've kind of figured out all the most important parts for how to deliver this into large-scale organizations... And so now revisiting the core UX, I think we feel pretty good about the degree to which we can implement all the things that people want in terms of AI code generation and agentic automation within the editor. But it does remain an ongoing challenge. It's not a trivial problem by any stretch to solve, and it's something that a lot of the upstart people trying to do things with like code plus AI they have not solved yet, and I think will take some time for them to solve.
+
+**Adam Stacoviak:** Mm-hm. Is your most popular editor VS Code, or your target, where a lot of the generation is happening?
+
+**Beyang Liu:** I think probably by usage numbers, VS Code is the most popular, but we have a ton of JetBrains users as well, and we also support Visual Studio and Eclipse, which - it might be surprising to some folks in your audience, but there are very large government agencies that are still on Eclipse, and they are constrained by factors beyond their or our control, and being able to deliver this into that editor is kind of game-changing. We're literally bringing the latest technology into a very old-school legacy editor, and it completely upends the speed at which they're able to move inside those codebases.
+
+**Adam Stacoviak:** I know that one of your teammates, I think, went to Zed... I'm thinking, "Why wouldn't you just build your own editor or acquire one? Why not just acquire Zed, and make it easier?" And then deploy the best editor ever out there.
+
+**Beyang Liu:** Yeah. So I think the teammate you're talking about is --
+
+**Adam Stacoviak:** You can mention his name if you want to.
+
+**Beyang Liu:** ...is Thorsten Ball. He's awesome. He actually just bounced back, actually.
+
+**Adam Stacoviak:** I saw that.
+
+**Beyang Liu:** \[01:07:54.06\] He rejoined us.
+
+**Adam Stacoviak:** That's why I felt comfortable mentioning his name in this scenario, because he's back again.
+
+**Beyang Liu:** Yeah. And look, he's awesome. Look, I love the Zed folks, too. We've had one of them on our podcast back in the day; I really love what they're doing. I think it's a fantastic editor, and I sincerely hope that it becomes really big. I think they have a unique approach here. They're not a VS Code fork. They're really kind of rebuilding an editor up from first principles, and focusing very much on performance, which I absolutely love to see as a developer. And I hope they succeed. I would love that to be an editor that I could use someday.
+
+As far as our business is concerned, the way we view the in-editor experience is -- maybe this is a difference in our vision for the world compared to some of the other players in the space, but we don't think that there's going to be a single code editor that is universally adopted for every language and every platform in the future, much less a closed source one. I think open source in the editor remains important, because it's so near and dear to the inner loop and the individual developer workflow.
+
+So when we think about the type of in-editor coding experience that we want to enable, we start from a point of wanting to deliver that experience into the best editor for your particular ecosystem or your platform. I guess that's somewhat of a contrarian view these days, because there's all these VS Code forks floating about, that frankly speaking they've done a really good job on some of the core UX, and kind of pushed the frontier in terms of showing people what's possible with AI. But the long story -- I guess the short of it is for us, we are thinking about building for the large codebases that make the software that power the world. Those are multi-editor environments. They're also environments where it's not just the context inside your editor that matters, it's actually connecting it back to the overall context of your organization and the entire codebase. So the way we're approaching about building for the editor is to bring the best AI UX in the field in a way that feels native to each of the editors that we support. We eventually hope to support all the editors out there.
+
+**Adam Stacoviak:** That makes sense. I mean, in essence, what you said was "Go where the developers are", versus force them into your stadium. You're going to all stadiums. A world tour.
+
+**Beyang Liu:** Yeah, basically. And part of this too is like - when we look at the UX of how people employ AI inside their editor today, we don't see anything that fundamentally requires a different editor... Especially looking forward with model capabilities evolving as quickly as they are, I would say our point of view is that every six to nine months there is a step function increase in LLM capabilities where the assumptions that you had to build around in terms of what sort of user experience you enable, or what are the facilities that you have to build around the model, those entirely change every six to nine months... Because the gaps that you had to fill because the previous generation of models couldn't just one-shot the thing that you're trying to do, or couldn't just find the right tool - those constraints no longer exist.
+
+\[01:11:53.17\] And the constraints that you have today, in another six to nine months won't exist either. And that has been the pattern since the release of ChatGPT. So we've seen a progression of these step function models. There was GPT-3.5, GPT-4, then Claude, Claude 3.5 Sonnet, and now 3.7 Sonnet. I think each of these has moved the needle in a way where a lot of the prior assumptions and constraints no longer hold.
+
+So when we look at the ideal editor UX today, looking forward, we actually think there's a very clean way to integrate AI-generated code into every editor, that doesn't involve a lot of extra UX Chrome for things. It's mostly just relying on the model reasoning capabilities, providing it the right set of tools, in the form of both local tools and tools that call out to codebase-wide context fetching... And then having a constrained set of hooks into the editor; like, how do you render a proposed diff into the editor buffer? How do you render an edit suggestion into the editor? These sorts of things, which... Look, you have to build those integrations so they work well, but it's a very narrow API that we think we can implement really well. I think the thing that we'd like to build is kind of like an AI kernel that you can port to every editor... And maybe even to the command line, because it's a lot of the same usage patterns that you see there, where it's like, you want to describe the change that you want, you want to see what edits are proposed, you want to accept those, and then you want it to continue doing what it's doing, but keep you in the loop as to what it's doing, and how that affects the rest of the codebase.
+
+**Adam Stacoviak:** Yeah, there's a certain framework that happens no matter what. There's certain interfaces you expect, no matter if it's your Zed editor - because I'm telling you to acquire it, or suggesting that you should - or every editor out there that developers use, no matter where it's at. Even legacy ones, or I suppose lesser used - I should not say legacy - like Eclipse. There's still lots of folks using Eclipse, for good reasons, I'm sure. For good reasons.
+
+**Beyang Liu:** Yeah, if you think about -- there's another analog here, which is like, a couple years back a big innovation in the editor space was the introduction of the language server protocol, LSP. And that showed that, hey, you didn't have to build bespoke code navigation for every single editor. Before that, it's like "Oh, I want to use this editor because some hacker out there spent night and day over the course of weeks or months hacking into this compiler to rip at the internals, and now that editor supports really good go-to-definition for my language of choice." LSP came along and basically said, "Look - fundamentally, what do you need to do? You need a go-to-definition, you need to find references, and then there's a short list of other functionality; that's really useful to have. We can abstract that away and make that feel native in every editor, so that you don't have to reinvent the wheel." And I think something that is possible with LLM-driven code generation.
+
+**Break**: \[01:15:02.21\]
+
+**Adam Stacoviak:** Let's dig into, I think, something that you have a unique perspective and vantage point on, which is the frontier models, what is being used, what's being deployed. You obviously give access to all the frontier models, from Claude, to Sonnet... All the different ones. What is most useful to developers? Do you offer them all because they all have unique perspectives on code generation? Is there one you particularly like best? What are you hoping for in terms of frontier models, like open source, non-open source? Are you a DeepSeek user? I don't see it in your list... That's a swath of things. Let's go from there.
+
+**Beyang Liu:** Yeah, so the landscape today, March 5th, 2025, is --
+
+**Adam Stacoviak:** It could change tomorrow.
+
+**Beyang Liu:** It could change tomorrow. By the time this comes out maybe it will be different. That's why I say the date, because I want to make sure that I'm pegging the statement to a specific point in time. The preferred model for our user base is the Claude Sonnet family of models. We have a lot of users still using 3.5, and a lot of users using 3.7... I think there's been some interesting discussion around 3.5 versus 3.7. For instance, I saw a bunch of Cursor users tweeting that 3.7 didn't feel like a big improvement; in some cases it's worse than 3.5. For the Cody users, it's actually been the opposite. And I think it's actually because Cursor has done more legwork around optimizing their prompts around 3.5. We've done less of that, because we've been more model-agnostic in our approach... And what we've found is 3.7 just has certain capabilities, especially around tool use, that have gotten substantially better. So that's why we're building -- the future of Cody is really going to be built around the set of capabilities that are now unlocked with 3.7 Sonnet.
+
+We also see a lot of usage of reasoning-oriented models for trickier problems, or problems more around codebase understanding, where a non-reasoning model might make a hallucination or an incorrect interpretation, a shallow interpretation of some set of context... So we do see some usage of the reasoning-based models like O3 Mini, and things like that. DeepSeek is really exciting. I actually think it is available now in our model dropdown.
+
+**Adam Stacoviak:** Oh, is it? Okay.
+
+**Beyang Liu:** Yeah.
+
+**Adam Stacoviak:** Marketing says no, but I'm not inside the application. I'm on /cody of Sourcegraph.com and it's not there, so... That's fine.
+
+**Beyang Liu:** \[01:20:05.14\] Oh, yeah, yeah. You probably need to be inside the editor. DeepSeek is also really exciting, because it's, I think, at the frontier of the open source models. And the open source models are also very interesting to us, because, one, we always like supporting open source, especially at the model layer; that makes it attractive from a variety of different points of view. I think LLaMA is also an awesome family of models that is available through Cody, and that we also see a lot of future potential around.
+
+The second reason why DeepSeek and LLaMA are very exciting is their finetuneability. There's a bunch of use cases inside the editor, like suggesting the next edit, or applying a proposed change into the editor, where these open source models provide a very good foundation for tuning a very fast model, that's specifically targeted at those use cases, where the use case is very latency sensitive. So we want to preserve the context awareness and preserve the intelligence as much as we can, but since it's constrained to the specific use case, it doesn't have to have as much of the general intelligence that a frontier model has. But because we have relaxed that constraint, we can push the latency and cost down to the point where it feels more instant.
+
+**Adam Stacoviak:** Can you explain to me - or is this proprietary - how you interact with the various models? You mentioned you're agnostic, so I imagine just like you want to go where developers are, which means whatever editor they're using... So that means that you've got an agnostic approach to how you gain your context via Sourcegraph, and apply that to the age or the model that's going to give back the generation, or give back the reasoning. Can you explain to me how that works, from prompt, to Sourcegraph context, to the model and back again?
+
+**Beyang Liu:** Yeah, so model choice was something that we prioritized as a key original design constraint around Cody. We basically made the bet that the model landscape would shift...
+
+**Adam Stacoviak:** Constantly.
+
+**Beyang Liu:** ...and evolve. And I think that prediction has panned out. And so we never wanted to be overly constrained to one specific model, or the current generation of models... And so we architected the code in such a way that we could easily introduce new models, and also customize the prompts that we use for each model, so that we can use the best prompt structure given each new model that comes along. And so because of that - one, it's advantageous if you're a user, if you want to play around with the latest model on the day it drops, most of these models we make available within 24 to 48 hours for users of Cody Pro. And then if you're an enterprise user, this is also helpful, because sometimes you have constraints around which model you can use. There's a bunch of enterprises, for instance, that don't want to use DeepSeek, because the progeny of that model just rules that out.
+
+**Adam Stacoviak:** For whatever reason...
+
+**Beyang Liu:** Yeah, for whatever reason. \[laughs\] And so we don't have to use DeepSeek in that case. Some people cannot use Anthropic, or OpenAI. Maybe they're tied to Azure OpenAI, or AWS Bedrock, so they're constrained to the models that are available on that platform, and we can easily make the functionality available given their model constraints.
+
+**Adam Stacoviak:** Can the person or persons over at the org, Sourcegraph org, I suppose - can they delimit or remove the opportunity to even select a model they can't use?
+
+**Beyang Liu:** Yes.
+
+**Adam Stacoviak:** Oh, that's cool.
+
+**Beyang Liu:** You can do that at the admin level.
+
+**Adam Stacoviak:** That makes sense.
+
+**Beyang Liu:** Yeah.
+
+**Adam Stacoviak:** That's the way to do it, man.
+
+**Beyang Liu:** Yeah. Yeah.
+
+**Adam Stacoviak:** Any plans to -- is there any more on that front? Did I cut you off?
+
+**Beyang Liu:** I mean, so we partner very closely with all the various frontier model development shops. I obviously can't share anything that's not public, but I think we're seeing a lot of exciting things coming down the pipeline, and so we're kind of like building our future in editor experience with an eye towards where we see the puck going in terms of model capabilities... Which is very exciting for the future.
+
+\[01:24:14.16\] I think the thing that I'll say here is that AI has already gone through two to four mini disruption cycles. We typically think of disruption cycles as being on the order of years at least, right? But the shift and the step function improvement in model capabilities that we see every six to nine months or so essentially resets the game for what the ideal UX is, at that cadence; every six to nine months. So rewind to the ancient age of --
+
+**Adam Stacoviak:** One year ago...
+
+**Beyang Liu:** One year ago -- 2023, right? 2023, Copilot was king, they seemed unassailable... Everyone wanted autocomplete. No one even cared about chat, or context awareness in that chat, right? Because autocomplete was where it's at. You want it to be fast, you want it to be snappy... You want it to be a little bit context aware, but people just wanted it to be fast, right? And that was the UX paradigm that was dominant at the time.
+
+And then in 2024 the paradigm completely shifted. I think it first shifted with GPT-4, where - okay, that was a model that had a step function improvement in its ability to one-shot applications. So now it's like, why would I sit there, autocompleting stuff, when I could literally just generate an entire mini app from scratch? It also got a lot better at incorporating context, so now you can do kind of context-constrained code generation, which we deployed, and it had massive success in the enterprise. But then there were still certain things that it couldn't do well. The code was often sometimes broken, or didn't quite compile.
+
+And then when Sonnet rolled out, that was like another step functions. Now all of a sudden these things that were kind of just beyond the capability, are now firmly within -- like, I can trust this thing to reliably emit JSON, for instance. That used to be -- that's a solved problem now. People used to write entire thought pieces around "How do you constrain the model output to produce a valid JSON?" These days you don't even think about that, because the model has been trained to treat that as a solved problem. And I think now we're seeing it in terms of in-editor agents. That is the future, and that's what we're building towards. A year ago, I think if you were pushing agents, there were certain companies that were like "We're going to build the agent that eliminates the developer entirely." I think now people recognize that they were kind of selling beyond their existing capabilities at the time. It made for a great kind of marketing moment, and they got their name in the headlines, but there was a disappointment in the actual product experience.
+
+Now, with the newest set of models, we're seeing this approach, just like -- this is kind of the new paradigm. That's what we're building for. We have kind of like a wait list of people who want to get access to the thing that we're currently building in collaboration with a lot of these frontier shops... And it works very well in large, messy codebases as well, because again, that's the problem that we're solving for.
+
+And so I think the future is very bright, and I don't -- at some point I think we might plateau in terms of model capability, but at least for the past two years, there's just been this regular cadence of step function improvements in model capabilities. It's just been really exciting to see that, and sort of ride the wave of the development and maturation of this technology.
+
+**Adam Stacoviak:** This wait list - is it a secret, these details behind it? Can you enumerate at all? Can you tease at all?
+
+**Beyang Liu:** When is this episode going out?
+
+**Adam Stacoviak:** Next Wednesday.
+
+**Beyang Liu:** \[01:28:00.21\] Oh, next Wednesday. Oh yeah, it's closed right now, but I don't know, DM me on Twitter... We're playing this very ad hoc right now, just because... I don't know, what we've learned over time is that new kind of skunkworks projects like this, it's never good if you make them overly formal too quickly. So we like it to be scrappy, we like it to be just a handful of devs kind of locked in a garage-like environment of some sort, just hacking on stuff... But I think we'll have more to share here soon.
+
+**Adam Stacoviak:** You've already long-betted on the UX inside the editor, so that's not changing. It's an improvement, or maybe something like that with a particular model or model developer.
+
+**Beyang Liu:** Yup, basically. So the way we think about it is basically like this... What we are doing, and what I'd recommend other organizations do is you want to position what you're building inside your organization to ride the wave of model improvements. So what that means is you have to just assume that the reasoning capabilities and the latency will continue to improve at approximately the same clip as they have been. So what do you do? To a certain extent, you have to hack around the existing limitations, but you don't want to invest too much in those, because another six months rolls around and those are kind of obviated.
+
+I think where you do want to invest in right now is very high quality tools and capabilities that can be composed with the rapidly-advancing reasoning capabilities of frontier models. And so it's essentially -- for us it's kind of nice, because we've essentially been building these tools the entire time at Sourcegraph. It's like, we're building tools to help human brains explore the codebase, effectively, and build up a working understanding of it, and write effective code pattern matching against the existing patterns, and validate that that's correct and consistent within the style of your organization and the rules in your organization.
+
+Now increasingly we're seeing the LLM brain being swappable in for the human brain in a growing number of tasks. Every kind of advancement in reasoning capability or context awareness or pattern matching capability advances the frontier of where an LLM serving as the coordinator of these tools, the orchestrator of these tools can now automate something that previously a human brain would have had to attend to.
+
+And so the way we view it is we are really investing in the tooling ecosystem and infrastructure that we build around the model, with the assumption that all the frontier labs are going to be pouring, it seems like all the money in the world - billions, if not trillions of dollars - into pushing the advancement of this reasoning technology as far and fast as they can.
+
+So that's the big wave right now, and the way we're designing our platform is to just provide the best tools for accelerating and automating the software development lifecycle, under the assumption that there's going to be kind of like a -- maybe Moore's Law is not the right analogy here, but I think you can assume some pace of advancement in reasoning capabilities for at least the next couple of years.
+
+**Adam Stacoviak:** Yeah. Well, I'd imagine that you're getting this access and you can call these folks partners and work in this degree because Sourcegraph is a deployable target for all of them, essentially. They can't assume that "I go to the Claude website or to the Open AI website, or to my own ways", that Sourcegraph's Cody is just one more way to deploy and distribute those LLMs and those reasonings, so to speak, to developers where they're at, you know? I'd imagine that's probably true.
+
+\[01:32:14.01\] Can you speak to how saturated your user base is with -- like, of the developers who are developers in quotes to Sourcegraph... I don't know what you call your people, so to speak; your user count? I don't know, I'm trying to get your lexicon in my brain.
+
+**Beyang Liu:** Yeah, yeah.
+
+**Adam Stacoviak:** Do they all use Cody? is everyone using Cody? Is Cody one of your most successful products?
+
+**Beyang Liu:** I would say Cody has been, I think by far and away the most successful product that we've shipped, with code search being kind of a close second, if you track by user count. If you track by revenue, we have other features that drive a lot of revenue as well. Batch changes, which is the large scale refactoring thing - it's very, very useful and very, very valuable for large, messy codebases. But in terms of overall user account, developer count, I think Cody has been extremely successful.
+
+There's not 100% overlap yet. We do still have some people using code search who don't use Cody, or some people using Cody that don't use code search, but we see an increasing number of people kind of flow between generating code with AI, to looking up context myself to make sure that I understand what the code actually does, and what APIs it's using etc.
+
+And yeah, I would say we are seeing an increasing amount of people starting in one point and getting pulled into the other, and for that reason we've also made this AI question answering interface available in a web application, where you can go to any repository in the Sourcegraph UI and ask questions when you're viewing the code in the Sourcegraph web application. So just trying to connect the dots... We kind of developed Cody as a separate, independent thing, but increasingly, our user base is really pulling us to integrate it into kind of the unified Sourcegraph experience.
+
+**Adam Stacoviak:** Yeah. It's hard to see from the outside how Cody is independent, because you can go to /cody, which is what it is, but the headline does not say "Buy Cody." The headline -- I don't even know what it says, but it doesn't say that. I think it's like "The enterprise AI code assistant", you know?
+
+**Beyang Liu:** Yes.
+
+**Adam Stacoviak:** And then somewhere in the line, it's like "With Cody, you can do X." There's not this named thing where it seems like it's a standalone product, that also is deeply integrated.
+
+**Beyang Liu:** Yeah. I'd say candidly our marketing messaging hasn't been the greatest... I think there's some confusion over what the overlap is.
+
+**Adam Stacoviak:** Very much, yeah.
+
+**Beyang Liu:** And that's something that we're trying to address. I think one of the challenges for us is we built so many capabilities around accelerating and automating the SDLC over the years that there isn't a succinct way of describing what the whole platform does, other than it accelerates and automates your SDLC, especially if you're in a large codebase. But that's not as punchy as --
+
+**Adam Stacoviak:** What is SDLC?
+
+**Beyang Liu:** Software Development Lifecycle.
+
+**Adam Stacoviak:** I thought so, but I just wanted to make sure... Unless it was something on the inside. I'm like "What is SDLC?" Okay.
+
+**Beyang Liu:** Yeah.
+
+**Adam Stacoviak:** What's left unsaid? I know that we've talked a lot, you've got a short amount of time, you've got some other places to be... What can we say in closing? What's on the horizon that no one knows? Or knows less of, that you can mention in the closing.
+
+**Beyang Liu:** I think for us it's really mapping back to these two loops - the inner and the outer loop. So expect a huge overhaul and improvement to the in-editor experience that we provide, around what the AI can do when you're coding, when you're vibe-coding, or when you're writing code in your editor... And expect an increasing amount of automation around that long and fat tail of the outer loop of the SDLC.
+
+\[01:36:22.27\] And I think philosophically, the way we're trying to build this is we think the in-editor should be as open and adaptable as possible. So open source, cross-editor... These are the things that we value as developers, and I think pre-AI, every developer would say that they valued that. I think the first generation of AI coding assistants has kind of muddied that a bit, because a lot of folks have kept their solution closed source... But we think the end game here is going to be open, and we aim to build in the open here.
+
+And then in the outer loop, I think the way we approach the automation here is that we're not trying to anticipate all the different ways in which we can compose the different building blocks. We want to build the building blocks and use those to provide a handful of first party automation tools, like the code review agent. But at the end of the day, we want to hand this platform over to our users, who are developers after all, to go and combine these things into the things that tackle the things that eat up your time, and produce toil for you today. So that is our philosophy.
+
+We've been doing this for over a decade now, so we've kind of fully internalized the challenges and what a solution needs to be in order to work in the context of a large, messy codebase, as well as how to deploy that solution into your environment... And like I said at the very beginning, we chose this domain because we felt that we'd easily spend our entire lives working on this. This is where our passion is, and I see a long and bright future ahead for us as a company, essentially bringing all the biggest, most important codebases in the world to the point where they actually can achieve these industrial economies of scale.
+
+**Adam Stacoviak:** Industrial economies of scale. I love that. I mean, I think it sounds cool.
+
+**Beyang Liu:** Yeah.
+
+**Adam Stacoviak:** I'm excited about just the economies of scale really at that point, because it's generally been harder to add more developers to a team, and get better.
+
+**Beyang Liu:** Yeah.
+
+**Adam Stacoviak:** You might go faster, but not necessarily always better. And it's not their fault necessarily, because of willpower, right? Or just like human nature, like "I don't like you" or "You don't like me." It's hard to work together. Or "I had a bad day" or "My weekend sucked." Whatever makes you less effective or productive.
+
+**Beyang Liu:** Absolutely. And you should absolutely go read the Mythical Man-Month, because what you just said there is kind of like the core thesis of the book. It is adding more people to a software development project makes it take longer... Which is -- it is the truth that has held true since the dawn of software development. It is the fundamental challenge that we aim to solve, and it's frankly the thing that's holding back the quality of most software in the world today.
+
+**Adam Stacoviak:** Very cool. Thank you, Beyang, for sharing the depths of Sourcegraph and your journey. It's been fun. Thank you.
+
+**Beyang Liu:** Absolutely. Thanks for having me on again, Adam.
+
+**Outro**: \[01:39:59.10\]
+
+**Adam Stacoviak:** I had forgotten that you open sourced Cody.
+
+**Beyang Liu:** Yup.
+
+**Adam Stacoviak:** And I feel deeply, like, "How do I not know that?" But there's so much that happens in our industry, and I can't remember it all... But I think what made me not realize this is like, it seems like to get into Cody, I have to have a Sourcegraph account.
+
+**Beyang Liu:** Yeah.
+
+**Adam Stacoviak:** And then I don't know how to have a Sourcegraph account, because I try to have a Sourcegraph account and it says "That email doesn't exist." I'm trying to sign up even.
+
+**Beyang Liu:** Yeah.
+
+**Adam Stacoviak:** I was doing it during the show and I was like "What's happening here?" So I don't know how that works or what's going on there, but share with me this wart that I've exposed.
+
+**Beyang Liu:** Yes. We are fixing this. So here's the historical context... A couple of years back we made the really bad decision to eliminate any self-servability in our platform... And at the time, it was just a matter of focus. We were bandwidth-constrained, and a lot of the pull that we saw in the market was for enterprise codebases... Because those are the people for which the pain that we solve becomes existential. It's like, you feel the pain of context gathering in a small codebase, but you're still small, the codebase fits inside your editor... It's not the end of the world. Once you reach a certain scale, then it's like "Holy crap, we can't get anything done. This is existential." So that's where the pull in the market for us --

Refactored in prison (Interview)_transcript.txt

@@ -0,0 +1,301 @@
+**Jerod Santo:** Today we are joined by Preston Thorpe, who's currently incarcerated at Mountain View Correctional Facility in Charleston, Maine. How are you doing, Preston?
+
+**Preston Thorpe:** Doing alright. How are you guys?
+
+**Adam Stacoviak:** Did you just dox him, Jerod?
+
+**Jerod Santo:** I don't think so. This is right there on his website.
+
+**Adam Stacoviak:** \[laughs\] I'm just kidding. You were very specific. I liked it. It was good.
+
+**Jerod Santo:** Well, I didn't want to just say Preston's in the clink, but that's what I was gonna say. Do people still call it that? Is it the clink? I think that sounds cool.
+
+**Adam Stacoviak:** I don't know what they call it these days, man.
+
+**Jerod Santo:** Preston, what do people call where you are?
+
+**Preston Thorpe:** It's certainly not called that here... It's just kind of -- yeah, locked up, or they're put away. I don't know.
+
+**Jerod Santo:** Yeah.
+
+**Preston Thorpe:** I think the clink is a little outdated.
+
+**Jerod Santo:** Yeah, I think it's from the '50s, probably.
+
+**Adam Stacoviak:** It's definitely '80s slang, '90s slang, which is our era, Jerod.
+
+**Jerod Santo:** "If you do that, you'll end up in the clink." Yeah. I think it sounds cool.
+
+**Adam Stacoviak:** I'm cool with it. I mean, as a term...
+
+**Jerod Santo:** Certainly not cool from the inside. Maybe cooler from the outside if you're older... Regardless of what you call it, you're in a unique place. At least first time for us, right, Adam? ...we've interviewed somebody from inside.
+
+**Adam Stacoviak:** Yeah, this is definitely a first.
+
+**Jerod Santo:** And maybe you're surprised that it's possible, you know? Like, I wouldn't even expect -- like, you can podcast from inside a correctional facility? How did this all come together for you? How are you able to do this from where you are?
+
+**Preston Thorpe:** Yeah, so there are, I think, a few podcasts that are put on from correctional facilities, but the content is obviously a little bit more niche. I think California has an ongoing one... But certainly, nothing this, appearing on a normal podcast. At least none that are sanctioned anyway, right? It's like, you might see it where someone's interviewed from a phone that they're definitely not supposed to have, but...
+
+**Jerod Santo:** Right.
+
+**Preston Thorpe:** Yeah, it's certainly a unique situation. There's definitely all kinds of backstory to that. So first, I guess I would want to put a disclaimer there that this is not a common situation, not a common occurrence. There is internet access and laptops at Main facilities, but this is a very special, special occasion/circumstance with the access that I've been given in particular.
+
+**Jerod Santo:** Yeah, we had to ask permission, sign paperwork etc. This has been in the works ever since we had a conversation with Glauber Costa from Turso, and he told us about you and how you have been contributing from inside on open source, and such a cool story. Tell us a little bit of the story, how you ended up where you are. You're serving 10 years, or you're on your 10th year...? Tell us the terms and kind of where you are in your term.
+
+**Preston Thorpe:** Yeah, so I'm currently on my 11th year, and definitely my last. I am approximately a couple months, roughly, away from release...
+
+**Jerod Santo:** Nice.
+
+**Preston Thorpe:** ...to home confinement. So I have to spend some amount of time with a curfew and an ankle monitor and all that. Yeah, so I came about being here in high school. I was just young and made stupid decisions, and chose to get into drugs, just being rebellious... And ended up getting kind of thrown out of my parents' house, but in a manner that they definitely didn't expect. The results that happened, it was a really unfortunate kind of circumstance.
+
+\[08:15\] I was quite a computer geek, you'd say, in high school, and I was really deep into the ware scene, the torrent scene type deal... And at the at the time - it was kind of before what everyone probably knows as the dark web, and the kind of underground, internet black market that exists today. There was a version of that that was around in the pre 2010 era, and I had some proximity to that, because I was involved in the torrent scene and the ware scene; there's some general kind of -- it's adjacent. And that kind of access, unfortunately -- I think what would have occurred in any other circumstance was I would have picked out/ran away. I probably would have learned my lesson fairly quickly, and been super-broke, and it wouldn't have worked out, and then I would have ended up coming home, and apologizing, and turning my life around. That was definitely their intention, you know. But unfortunately, I became very involved in -- you know, I was using drugs and had essentially ended up getting real deep into that particular scene. Large amounts of money and 17-18 year old mentality is just the worst possible combination. So I pretty much thought I was untouchable, and the man... And yeah, I ended up getting in real deep, into drug addiction, and eventually into trouble. And I went to prison for the first time at 20 years old, in 2013.
+
+**Jerod Santo:** So you were using, you said there's lots of money running around, so you're distributing as well... I mean, you had easy access to basically cheap drugs via the dark web, and whatever scene you were a part of there. And were you distributing? I assume you were, because you've got to get the money somehow, right?
+
+**Preston Thorpe:** \[10:48\] Yeah. Unfortunately, that was the case. These things were available right from source countries in ridiculous low wholesale prices. And it felt the movies, if you ever seen Blow... It certainly felt a version of that. It was a very unique time, when certain things hadn't been scheduled yet... Like, the DEA hadn't scheduled certain things, so they were just freely available on the clear web, from distributors. And they weren't illegal. Obviously, some of these things are; I was certainly not not breaking the law, right? But it definitely made the price and availability pretty low. And yeah, it kind of just got -- it took a life of its own, and... Yeah, the ego, the money, the feeling of the fact that I didn't come home and apologize... I kind of felt it was "Hah! I've figured it out." I had some kind of cheat code to life, and the whole time I'm just consuming tons of drugs, and not feeling a drug addict, because I'm obviously not the kind of typical, what you see, broken and struggling for their next hit... It was a never ending supply of everything. So it just all contributed to just making it worse and worse, because I wasn't self-aware of it. It was a bad, bad deal. And surprise, surprise, 20 years old, I ended up getting caught with a pretty substantial package of Ecstasy in the mail from Vancouver, and got a 4 to 10 year prison sentence.
+
+**Jerod Santo:** It would just ship to your house, basically, from a different country? Like, how did they actually -- is this common? It just goes through the postal system that, or what?
+
+**Preston Thorpe:** Yeah, so at the time it was not common, and that's kind of what I said, it was a very unique time in history. Now there's rap songs about it. It's frequently mentioned in hip hop, and it's pretty widely known that this is a thing. Back then it wasn't. It certainly happened, but it was not something that anyone was aware of, or at least not nearly to the degree that it is now. And that just kind of further enhanced the feeling that I was just going to get away with this forever... And it was just -- it was a bad, bad scene.
+
+**Adam Stacoviak:** I think it's wild how much -- I think even today, Jerod, a lot of drugs flow through the U.S. postal system, because... I don't know firsthand, so I can't say that, but I know people who boast about things, even to this day. And it might not be a large quantity of what you described, Preston, but it's something similar, and smaller, or whatever. But I've heard of people talk about moving these things, just to friends. That kind of thing. In passing, let's just say. Do not interrogate my friend group. I'm just kidding... I'm not even around these people anymore. I keep my distance.
+
+It's so wild, because I think that -- and you probably can share some of this, Jerod, maybe not so much in detail, but... I think in adolescence, in that timeframe in our lives, it's already natural to feel some version of superhuman, some version of invincible, immortal... the end and death and suffering and loss will never come, "This is somebody else's life", "I will never get caught, because this is just not what happens. I'm untouchable in some way, shape or form", just because you're in that adolescence spectrum. And you couple that with somehow being able to find access to this stuff... And I imagine -- like, we know that drugs alter your chemical state in your brain. We understand that even before this age group that you're in, your frontal lobe isn't fully formed. They say it's not fully cooked until 25, roughly, so you don't even have a lot of your frontal lobe rationale kicking in it would later in life, to say "Oh hey, Preston, this is a bad choice." Or "Hey, this is a dangerous situation." In those scenarios, it's almost as if your brain can't tell you there's danger, or can't tell you these things.
+
+\[16:04\] Now, that's not the truth for everybody. There's varying degrees of spectrums to that. But man, what a situation to be in, to be in that era and have that kind of access, and then go down the road you've gone. But in the end, though, you're in a place where -- I don't want to glorify your story in any way, and you don't want to glorify your story... So I don't know the full amount of it, but there has been an ability for you to recover in some way, shape or form and to give back. And I think that's the uniquely wild, free thing that is called open source software, and even software development. Free in the fact that if you have a text editor, or a terminal, or access to a computer in some way, shape or form, you could probably find a way to hack around, and poke around, and get curious, and discover things. And the fact that open source is available to you to make that connection - Jerod, you mentioned Glauber Costa from Turso and the things they're doing there... Being able to contribute to, I would say, a high-value open source project. That to me is wild. To go down the road you've gone, but then also to come back in a way to be able to contribute. How does that make you -- knowing your story, and you live this every day, and this is your reality, not ours... How does that make you feel, that fact? How does it make you feel?
+
+**Preston Thorpe:** Yeah, it's honestly amazing, and I feel there's this kind of -- I go back and forth between feeling this isn't real, this day that I live now, and the career that I have, and the things I'm able to do, the access that I have. Half the time I feel "Wow, this is unreal", because compared to two years ago, let alone five years ago, it's just a completely different life. Never in a million years would I ever have guessed that this would be possible from a medium security facility.
+
+And then the other half of the time I look at my past life and I remember things the things that we're talking about, and I think "That isn't real. That happened to someone else." I was kind of recollecting about some stories with it... This kid has been pretty much my best friend for 15 years, and thinking about these actual vivid memories was the craziest. It doesn't feel like it happened to me, because it's so far away from the person I am now, and honestly, the person I was raised. I grew up in a good household, I've got a sister and three brothers, and my parents are together, still together to this day... Great people. And I grew up a good kid, and so it gives me even less excuse. I knew what I was doing was wrong, so... First of all, yeah, I was young and stupid, but I also was super-aware of the consequences of my actions. But the person that I was for those years is so far from who I am now that it doesn't -- there's no sane reality where the person I am today is making any... It's unbelievable that I just didn't care about anything. I don't know, I created this messed up kind of identity and world for myself, and just went with it, accepted it completely. And yeah, it got -- it was out of control.
+
+**Jerod Santo:** \[20:16\] So you said you got 4 to 10 years and you're on year 11. Was there a second thing, or was it extended, or...?
+
+**Preston Thorpe:** Yeah, no, there's certainly quite a bit more there. So I get out after serving three of the four. So when I say four to 10, and the way that \[unintelligible 00:20:32.02\] system works is you have a minimum and a maximum. You serve every day of your minimum, pretty much. So I had a 4 to 10, got a year suspended... There's a clause for doing drug treatment and not having dirty urine, disciplinary tickets. So yeah, so released after age 23, after three years. And my parents had moved back to Michigan, where they're originally from. So this prison sentence I did was in the state of New Hampshire, and the time that I spent in prison - this was my first time in prison, it was my first time ever getting arrested. And honestly, when you're young and you're really impressionable, and the people that I was surrounded with in prison were very different from the people that I was surrounded by on the street... And there's a certain kind of novelty to prison, especially when you're in that lifestyle... Pretty much everyone you know has pretty much been to prison. Prison is like a rite of passion passage type thing... And I just continued making all the wrong decisions, and just seeking out and surrounding myself with just the wrong kind of people, and continued to do the same stuff, and just in a different environment.
+
+**Jerod Santo:** Right. Got busted again and went back in.
+
+**Preston Thorpe:** Yeah. So even still in prison. And when I was released, I had the opportunity to move to Michigan with my family. So I was released with -- I had no identification, I had no social security card, birth certificate... I had nothing. No money. And one of my old friends gave me an old iPhone, and it was just like a rough, rough situation. And I had a two-week stay at a rooming house in the city of Manchester... This particular rooming house is notorious, to this day. There's overdoses every day. There was parole officers and cops walking around this rooming house. And that was my first -- I was released and went there to get my room... You know, you need an address to make parole, right? So just to get out, I had to -- one of my buddies got me that room, and... Yeah, it was just like a major -- it was not a good scene. So I didn't transfer my parole to Michigan. Maybe I told myself it was because it takes a couple of months for the transfer to go through, and I wanted to get out immediately. In reality, I think it was me in the back of my mind just giving myself an out.
+
+\[24:11\] Because going back home would have been a guarantee that I have to abide by the rules, and do the right thing, and try to get a job... And when you're used to the kinds of money that I experienced at a young age, the thought of making minimum wage, or 10 bucks an hour, or even 50 bucks an hour, it was just not... So I was just making excuses for myself, and... Yeah, I very easily justified getting involved in the same stuff.
+
+The time that I spent out in 2016, I was out from -- I was released in March, and by Christmas Eve I was arrested again. That is the case that I just finished. And yeah, and then was on bail, and was arrested again in May of 2017. And I've been in since May 5th of 2017.
+
+**Adam Stacoviak:** As you tell that story, it breaks my heart, because my oldest son was born in March of the year that you were arrested, 2016... And he's nine, and I see his life, and I just think like "Gosh, man... These choices that people make." And I don't want to downplay, because this is your story and I'm not trying to be negative, but... It's just super-sad to hear, because I see the life my son has, and obviously he's born, so he's not your age, but... I see all the new. I see all the new goodness, I see all the things we've done in our lives as a father and as a son for him, and I just think what you've been doing since that same timeframe. I know what we did for Christmas that year. It was his first Christmas, kind of thing. So these are like dates burned in my brain because of my relationship, and it just breaks my heart for you.
+
+**Preston Thorpe:** Yeah, no, it's definitely no joke. I think these things are very commonly portrayed in media, music, television as being like - yeah, very different from the reality that I've seen personally. I've seen quite, quite a bit of it. Yeah, I haven't had a Christmas on the street since 2012. It was the last Christmas that I was free. This life will get you nowhere. Yeah, and I don't think I -- I haven't seen a single person succeed... I haven't met one person that -- I met a lot of very, very successful people in that life, but none of them got to enjoy any... Like, you might get a few years here or there, but it never works out. Never.
+
+But yeah, so... After I came in on, on this sentence, I got -- I was originally sentenced to 15 to 30 years, and receiving a sentence like that was super-devastating. And I was particularly bad into the drugs before I came in on this sentence. I had -- kind of prior to this last 2016, my drug use was not... Like, it was bad, but it wasn't as bad as it was this previous time.
+
+\[28:06\] So my head was just not in the right place. Any hope I did have for a life was now just completely over. I didn't have much hope to begin with, so there wasn't a lot there to lose, but... Yeah, I pretty much gave up and had really accepted the identity of who I was at that time. I really believed that "These are the decisions I made, and I'm paying for them now, and this is the person I am, and I'm going to--" It's just something that -- I accepted that identity, and that's something that really bothers me today, being a completely different person... Going back and thinking "Man, when was it that I decided that I was okay with being someone that I clearly wasn't, and doing things that--" And that's what ended up happening.
+
+So I came in in 2017, was in the \[unintelligible 00:29:22.10\] system for a couple of years, and the case that I came in on was a very high profile case, and I had already been to that prison system before, so I knew everyone... You know, it's a very small system, and my name was a frequently brought up name in the system. And the things that I was involved with -- I was making some poor decisions; I don't want to elaborate too much, but... They ended up getting me on a list for an interstate transfer. They just want to get you out of the system completely. And one of the luckiest things that have ever happened to me - because I had court dates remaining on that sentence, you can only get sent within the New England States. So because they were going to have to transport me back for my sentence review board, and -- so I had some scheduled court dates, and I was scheduled to go to Massachusetts... And I was actually excited about that. I have a lot of friends down in the Massachusetts system. That is what I thought that I wanted. That was the one I would have picked if you would have given me the choice. And I actually went out for a very minor surgery, and when Massachusetts came to pick me up, or they called or whatever, and learned that I was in the infirmary after this surgery, they actually denied my transfer. Like, after previously accepting it. And I had a bed there, and everything. New Hampshire had already accepted the inmate they were going to swap with me. So it was pretty set in stone.
+
+So by just a sheer miracle, they ended up denying me, and I thought I was going to sit -- I had been sitting in solitary confinement this whole time. So it had been over a year at this point that I'd been in solitary confinement. Nothing, just a cell, and books if you're lucky, for 13 or 14 months by this time.
+
+\[31:57\] And then one day they, they came in and told me to pack up. And I was expecting it, but I thought it was going to take a bit longer. We ended up coming up to Maine, and it was certainly not -- I didn't even think it was an option. I thought it was going to be Rhode Island or Connecticut. Those are the three - Massachusetts, Rhode Island and Connecticut are generally the three you would go to. I didn't know Maine was an option. And that was quite literally the best thing that's ever happened to me. It absolutely changed my life.
+
+The Maine system is very different from just about any other system. It's probably tied with Vermont for being the safest prison in the country. And it doesn't have a lot of the prison -- it doesn't have as much of the mentality and the subculture and the gangs. It's not as prevalent; not nearly as prevalent. The New Hampshire system is certainly not San Quentin, but there are units where that's very much how things work, and you're going to be exposed to that, for sure. Maine is very different, and not being around any of those people that I had been -- by this point I'd done six years or more... So not being around the people that I had been influenced by, and the people I felt like I had some kind of a reputation to uphold myself to, and just kind of being in a new place, and with a large sentence, and just... You're kind of focused -- you have no choice but to just focus on yourself, because there's nothing else to do. I didn't owe anyone there anything. I didn't owe anyone back home anything either, but it was just a feeling like I was this person and I was entitled to act this way, and be this person, in the other system.
+
+So Maine has a lot of educational opportunities, and a lot of -- it's just a much nicer system. And after a couple of years I kind of just had an epiphany of like "At what point did I just decide that I was okay with this?" At what point did I decide that I was alright with being the person that I was? When did I decide that I was okay that I just am in prison, and I live in a cell, and this is normal? Why is this normal? Why am I not spending my time trying to better myself and work on being the person that I'm going to be when I get out? Because I can't go back to what I was doing and who I was. And yeah, it was quite literally a single day... I vividly remember the day that I had this epiphany, and from there I just started to make -- that was the beginning.
+
+**Break**: \[35:45\]
+
+**Jerod Santo:** And after this moment, did you just think to yourself "Well, I like computers. I've been a computer guy since the start." Because that is, ironically, how you started your life towards this place, was through the ware scene, and really being adept at using the internet... And you just decided "I'm going to pick up programming", or had you been programming all along? Tell us the technical side of your journey.
+
+**Preston Thorpe:** Yeah, so I had enrolled in college, or had already enrolled in college by this point... And this was pre-laptops. So the college - essentially, you could apply for a Pell Grant, and there was some funding through Doris Buffett, and there was... I think the professors would come in and have in-person classes, or maybe they'd do Zoom calls, but there was no laptops in the cell, like with internet.
+
+So I had applied for college already, and while waiting is kind of when this all kind of occurred... And the first semester that I end up in college just so happened to be the first semester that laptops existed, somewhat as they do today. And I had gone through a couple different avenues in my head about what I'm going to do. And I've always loved learning, I have spent thousands of hours, I read - I mean, hundreds and hundreds of books. Obviously, I've done a lot of time and segregation, and so that's the main kind of thing you do... But I had studied organic chemistry, I had studied - finance was the big one, right before I ended up enrolling in school... I learned everything that I could possibly consume, from getting books sent in, to watching CNBC all day... And I was kind of obsessively learning about the markets, and options trading, because I figured that I could day trade when I got out, and that was a potential kind of career path.
+
+And then when college started, I get the computer, and that was kind of epiphany number two... And I kind of realized, "Wow, I have four years left. Where else does someone get 24 hours a day...?" Well, you've got to sleep, so 14 hours a day, to learn something... I had programmed in high school, kind of before high school. Nothing super-crazy. LAMP, kind of, PHP... I've had an Ubuntu Linux tattoo since 2009. So as I mentioned before, I was definitely a computer nerd, so I kind of had just enough context to understand and remember "If you're really, really good at programming, that's kind of all they care about." You know what I mean? And I definitely did some research and kind of found out that criminal records or college degrees or any of that stuff is negotiable.
+
+\[42:26\] I don't know what year it was, but Google and Microsoft did away with their -- they don't even care if you have a degree, because there's just self-taught people out there that are spectacular. So that was epiphany number two, when I was like "Wow, if I were to spend 12 to 14 hours a day, every single day, for the remainder of my time, I could theoretically be pretty good at this. And hopefully, by the time I get out, I might be looking at the possibility of having a promising career." And that was probably a thousand days ago, and I have spent 12 to 14 hours a day, every single day, since that day. That was pretty much roughly a thousand days ago.
+
+**Adam Stacoviak:** That's amazing. It's amazing to have that realization, and make that choice. Sometimes people make a choice that's sort of dreamlike. "I dream" or "I wish I can do this", and there's some people who make a choice and resolve to do it, and they commit to that resolve, and they show up every day and they do that. I mean, because sometimes boredom, even though you want to make change, can get ahold of you, because we're all human, no matter what our scenarios are. Sometimes we're like "I really want to be this amazing thing, and here's my pathway to do it", but you don't follow the steps to get there, and you've done that. That's wild that you'd be able to do that.
+
+**Preston Thorpe:** Yeah, for sure. I think the environment, definitely -- I think it works well for this environment. The laptops are issued for college, and the network is heavily filtered. As far as what your allowed, the sites you're allowed to visit, it is for school. It's not for your entertainment. They allow Spotify, if you want to stream music or something, but the rules are essentially if you're watching YouTube videos and it's strictly for your entertainment, then that's technically against the rules. So that opportunity that I saw, when I got it, I was always very, very cautious of "There's no scenario where I'm losing this opportunity." So definitely, I would say that the environment of -- there's not a lot of distractions. It's kind of the ideal environment for someone that wants to spend 12 or 14 hours a day learning to do something. Yeah, it certainly helped.
+
+**Jerod Santo:** \[45:49\] Yeah, I saw some Reddit trolls who were, I think, reacting perhaps to the Primeagen talking about you, or something... Or maybe it was Turso, I don't know. And they were jokingly -- of course, it's just a joke; it's not real, guys... But saying that "This is the best life hack ever, ever, if you actually want to learn how to code." It's the best situation you could possibly be in. And it's kind of true. I mean, obviously, nobody would choose that... But once you're there, it has everything that it takes. You have to still put the work in, like Adam talks about, but...
+
+**Adam Stacoviak:** Yeah.
+
+**Jerod Santo:** You are well positioned, with a lot of free time, and very few distractions, especially like you said, because these machines are specifically for education... You know, what else are you gonna do?
+
+**Preston Thorpe:** Yeah, I actually get that pretty frequently...
+
+**Jerod Santo:** Do you?
+
+**Preston Thorpe:** I have a lot of -- so the blog posts that I wrote about my story, and then, again, when the Primeagen reacted to it on stream, the YouTube video, I have a lot of... It's actually been really cool. A lot of people reach out. I get everyone from just people just being supportive, to other people that have done prison time, and that might be in tech... And then a lot, probably more than anything, is college students that are maybe going for their CS degrees, and asking what the secret was; like, what's the hack. And I have been told more times than I can probably count that they are jealous of my situation, because they wish that they could just spend all that time and get really good at programming.
+
+**Adam Stacoviak:** That's a wild thing to be told. I think what's wild, even potentially more wild to that, is that we all literally have a choice. You made a choice that changed your life. And those folks who envy you, or are jealous of your scenario, can also make a choice. They can turn out and tune out different things in their lives. The place they're in is a series of choices that got them to where they are. And we know that because you just shared your story, and it's a series of choices. And so while they have jealousy for you, they can do the same thing; not in the exact same scenario, but they can make the choice to tune out the rest of the world, if they so choose. But wow, it is just so hard.
+
+**Preston Thorpe:** Yeah, definitely. And that's something that I keep in mind, for sure. Like, I've never been much of a social media person, but I understand that most people's -- one of their primary timekillers/distractions is going to be social media. And I'm going to be getting out fairly soon, regardless of what happens; the worst case scenario, it's like a year. But either way, it's soon. And that's kind of something that I've resolved to keep, is if it's not difficult to give up pleasures that you've never really had... So there's a long list of things that I've kind of told myself "I'm not going to start this, because I don't see a reason to change a whole lot." The sites that I visit, what I do on the internet, the people I talk to... There's definitely a feeling of "Oh, when I get out, there's no barriers, and rules..." But at the same time, there's also not a whole lot that I care to do, that I can't do now. I work at my job, for eight hours, starting at 8 or 8:30 in the morning, and I write open source software from 6 to 10:30 at night... And I work out before that. But obviously, I'll be spending time with my family, and all the kind of normal life stuff, but as far as how I spend my time on the internet, I'm probably going to stay pretty close to what I do now.
+
+**Jerod Santo:** \[50:46\] Yeah. Well, you developed good habits, so that will be a real virtue for you, to maintain those habits. I would add options trading to the list of things you should not get into when you get out, because your Robinhood account can also be an entertainment that turns into disaster... And I'm sure you'd like to avoid future disasters. I think your open source work is far more profitable and way less risky than options trading options. But tell us about the open source. How long did it take you to get into? How did you get into it? What do you work on in the open? etc.
+
+**Preston Thorpe:** Yeah, so I guess I could kind of start into just, I guess, programming in general, in here, and how that started...
+
+**Jerod Santo:** Sure.
+
+**Preston Thorpe:** So at the very beginning, there was no -- there was internet access on the laptops, but it was only in the recreation building, where the education department was. That's where the Wi-Fi was. So you could browse the internet in the rec building, in the education hallways, and stuff. But when you go back to the pod, you go back to your cell, there's no internet access. So I would save Python documentation... And especially back then, there was -- it's like Windows Active Directory network, and you don't have any execution policy/privilege. I don't know if you guys are familiar with the old AD Windows networks, but yeah, your execution policy is turned off, so you cannot install anything. You can run bundled, kind of, where all the libraries and stuff are bundled in the one application, the portable ones... But yeah, it was pretty painful. I had to get a particular kind of terminal emulator, that would -- it was all kinds of... I had to use a PIP proxy, that was from a .edu domain, to be able to get Python libraries, and stuff...
+
+And yeah, so after kind of jumping through these hoops, and at the same time kind of constantly bugging staff to install things, and asking for "Hey, I need Git, and I need..." And also volunteering to help out with -- you know, they don't exactly have a bunch of full-time IT staff, and this was a new thing at the time. So I ended up volunteering, and eventually working for the education department at the prison I was at, and ended up helping out quite a bit, just like running the network. And eventually, I think they just got sick of me asking to install things, or for certain privileges... I eventually ended up getting that job, and being able to at least have some level of root access to my laptop. And that certainly made it a lot easier.
+
+\[54:25\] And yeah, I was just like writing Python, and C, and I wrote my own kind of C standard library type deal... I knew enough to know that I wanted to start from the bottom, and not just learn React, or whatever. I think the usual kind of starting point today would be considered that pathway. I knew enough to know that that's not what I wanted to do. That I wanted to start from fundamentals.
+
+After some time, I'm in school, and I am working for the education department, maintaining the network, helping fix things, and get laptops ready, and re-imaging and kind of doing the daily maintenance... They started a -- I can't say they started, but people that were enrolled in college, that had graduated and got their bachelors, those -- there was a couple that were either in the process of getting their masters, or they were going for their PhDs. And they had gotten this special permission to work, usually with the university that they were attending.
+
+One person was a fellow, and he would teach, essentially -- I don't know, substitute teacher. He was teaching, or whatever, over Zoom, for the school. And people were talking about "Oh, they might start allowing remote work." They were playing around with that idea, of "Hey, if people can get jobs remotely, we might just sanction it."
+
+So I actually was the only one that was not a master's student. And I put in a request to seek remote work, after -- I'd probably been learning to program at that point for probably six months or so, seven; maybe seven, eight months. Really kind of completely out of touch with what I actually needed to know, and what was going to be expected of someone, but I just... I just knew I wanted to look for work. Like, I'm spending 12 hours a day, and I thought "Why not?"
+
+And they actually granted my application to -- I put in a special request to seek remote work, and it got granted. And in another brilliant stroke of luck, I ended up kind of cold messaging my CEO (now), on LinkedIn. I had heard about the company from someone else that was on my pod, that does work in prison reform... Like, he's big into that kind of community -- criminal justice reform, prison reform stuff... And yeah, I just kind of got a random tip, and ended up getting on a call, Zoom call with my CEO, co-director now, and they honestly had no -- probably shouldn't have hired me considering the situation, but... Yeah, they gave me a shot. And that was pretty huge. That was a great opportunity.
+
+\[58:45\] And while I was -- I was doing open source stuff at the time, and I was writing a lot of Go, and C, and Python... And I was learning Rust at that point. Right around that time is when I got involved in -- that was right after Exa was deprecated. So that would have been, I would say, my primary introduction to the open source community. I had certainly made contributions to a few projects before that, but that was the first thing I would consider to be -- like, being active in a project. And I don't know if you guys are familiar with the ls replacement command line...
+
+**Jerod Santo:** We are. We are, actually. We talked about it. When was that, Adam? With Nick Janetakis, a couple of years ago...
+
+**Adam Stacoviak:** Probably with Nick, yeah.
+
+**Jerod Santo:** We don't personally use it, but we are aware of it, as a like an ls on -- I was going to say on steroids, but I don't know. On better.
+
+**Adam Stacoviak:** On plus plus.
+
+**Jerod Santo:** Yes, there you go. So you got involved with Exa, and began contributing there?
+
+**Preston Thorpe:** Yup. It was officially -- Exa was deprecated... Because that was in the Debian stable repo; it was --
+
+**Jerod Santo:** \[unintelligible 01:00:19.27\]
+
+**Preston Thorpe:** Yeah. It had been around since 2015... And yeah, the author disappeared. I'm not even sure to this day if anyone's heard from him, but the guy pretty much disappeared. Really unfortunate. Super-talented programmer. And Exa was one of his -- probably his most popular tool; a couple other ones, too. And it being deprecated, it ended up on the frontpage of Hacker News. And I don't know, it was at a stage when if I found some issues that I felt I could fix, it made my day, and I was super-excited to contribute to anything, and have commits in any project. It was just the most exciting thing for me.
+
+Actually, I ended up having to -- so prior to putting in to Git remote work, I had put in a request to the warden at the other facility to contribute to open source... Because you're technically not supposed to talk to anyone, or have any comment on anything, any kind of -- you're making no post requests. It's like get only.
+
+**Jerod Santo:** \[01:01:57.14\] Gets only. Yeah. Was that a tough pitch, or not so much?
+
+**Preston Thorpe:** Yeah, it's hard when -- I think I contributed to Micro, the text editor... It's written in Go. It's like a Nano replacement. It's actually pretty cool. But that was my first commit. That was my first PR, open source PR... And that was like -- it must've been beginning of '23. Yeah, I think so. It was beginning of '23. And I screenshotted the whole process, and wrote this big proposal, and put the screenshots in there of "This is what it is. This is the code I wrote. This is what it does" and tried to give the administration an idea of what I was going to be doing, or whatever... And yeah, they approved that. That was really cool. That was probably three or four months after that, that I put in to seek remote work... But yeah.
+
+Yeah, when I started at Unlock Labs, I had continued -- I work, like I mentioned... There was a period of time where I would work for Unlock Labs -- like, the 12 or 14 hours a day was all spent on work-work... Because I was determined to prove that they made a good decision in hiring me... And there were certainly no other opportunities for me at that time, so I really wanted to show gratitude for the opportunity with them. And yeah, I had quite a bit to learn, but pretty much always have continued to either build stuff for the DOC, for the education department, like either build tools for them, or my own personal open source projects, or... Yeah, just other contributions.
+
+**Break**: \[01:04:31.01\]
+
+**Adam Stacoviak:** I'd imagine it's like you've got this new superpower that you'd never had before... Because skill-less -- I don't want to say you literally had zero skills, but comparative to what you have now, skillless probably beforehand. You had some experience with LAMP, you've had some experience with obviously internet and different aspects of that, but you've got this new superpower, and you're like "I just want to contribute. I just want to be useful." I'm imagining this is what you're thinking, because of your story that you just shared. "Where can I make a dent? Where can I put some good out in the world? I've been doing so much bad, or so much internally bad thinking, and so much negative self-talk about my identity, and I've had this epiphany and I've got this new dream, and I've got hope", and you've got all this inertia towards it, and you've got this new skill that you're like "Where can I put this skill to use? Where can I be useful?" That's what I'm thinking. Is that a version of what you did?
+
+**Preston Thorpe:** Yeah, most definitely. With all the time... Like, if I had a 20-minute conversation with someone - outside of myself - I would feel guilty that I wasted 20 minutes, because I felt like the three or so years that I had left was such a finite amount of time that any non-trivial amount of time spent doing anything but learning was considered to be like... It was definitely something I felt guilty about.
+
+And I think everyone kind of knows how much enthusiasm newer developers have to rewrite whatever, and take on features, or projects that are real ambitious, that you kind of learn aren't super-realistic... So that ended up with like -- you know, I wrote... The first thing was just an inventory system for the education department to keep track of all the laptops, and the credentials, and who had them, and the condition, and the allow list... Just crud -- that was the first thing that I built for the education department. Yeah, that worked out pretty well. And then when I came here, to this facility, from the -- I was just at a different Main facility. I was at the Main State Prison, and now I'm at, as you mentioned, the Mountain View Correctional Facility. Not Mountain View, California.
+
+**Jerod Santo:** No. All the way on the other side of the country.
+
+**Preston Thorpe:** But yeah, when I came here, they have -- believe it or not, actually, the state of Maine, and from what I hear, most facilities... It's probably something that viewers will be interested to learn. They have quite a bit of internally written software by residents. That's actually a fairly common thing. And I definitely was not aware of this prior, but the Main State Prison had a resident there that wrote pretty much all the software that it runs on to this day. So all the account sheets, the inventory for the woodshop, and inventory for the kitchen, and a system where you scan a barcode in the back of your ID, and use that to access your account to purchase things from the commissary... So there was quite a bit of this handrolled infrastructure at the other facility. And the administration here was aware that that existed. And when I came here, they had been made aware that I was someone that could replicate this. And definitely someone that could maintain some of this stuff here.
+
+\[01:10:04.06\] So the system, the point of sale system that they use for the \[unintelligible 01:10:09.08\] is written in -- it's vb.net. Let's say it's VBA. Yeah, it's definitely VBA, because it runs on Access, and Windows 8. So to this day I still maintain a point of sale written in VBA, on Windows 8, which is the absolute bane of my existence.
+
+But yeah, so the staff here asked me to build a real-time -- something to keep track of what everyone's position in the facility is. So when you come back to the unit, you scan your ID, and it will check you in to the units. Because currently you have to sign out, and there's people whose jobs it is pretty much to maintain these clipboards with everyone's current location, so that if something happens, they have to count everyone, they have a record of where everyone is. So they wanted that built. So yeah, I built that RFID-based system; I made that pretty much production-ready. And that -- they pretty much abandoned wanting to actually get it set up, because I think they were just... The guy who wrote all the previous software at the other facility ended up getting out, and now they're in a position where they don't have anyone to maintain it, and I think they're just trying to avoid that happening here. But I'm actually pretty grateful that they had asked me to do that, because writing a full production front, full stack application definitely I think is being on the list of projects that I completed by myself. I definitely credit quite a bit of my ability today to that.
+
+**Jerod Santo:** Is it a web app, or what kind of an app is it?
+
+**Preston Thorpe:** Yup, it is a web app. It was a \[01:12:31.06\] and SolidJS frontend. It just uses these -- actually, these right here. Of course, this is going to keep alerting me, but yeah... It's RFID HID reader. And yeah, I spent quite a few... I would say probably at least a couple hundred hours on that.
+
+**Jerod Santo:** So do you have to interface with the scanners themselves, or is there a way that you're getting the location data off the RFID?
+
+**Preston Thorpe:** Yeah, so it is just a -- so the client runs in a Tauri window, and these particular RFID readers are keyboard input emulators. I had a client for this on a Raspberry Pi Pico, that actually interfaced with the HID. Because you have administrator privileges to open dev -- like, you input, and actually get the raw bytes from the OS, but the web app frontend just had an invisible form. And because it ran in a Tauri window, it allowed me to just capture all the keyboard input.
+
+**Jerod Santo:** \[01:14:02.00\] Using Tauri... That's one thing I noticed about you, Preston, is you seem to be very on the tip of technology. Like, on your About page: "Big fan of Rust", obviously Neovim, Tmux, Ghostty, Alacrity, you're using Tauri, you're using SolidJS... You just seem like you have a certain group of people who like to live on the edge. Do you know that about yourself? Do you realize that?
+
+**Preston Thorpe:** I would agree. I think some of it probably has to do with the fact of just like the time periods that -- so the last time I programmed anything was in 2009. And then when I kind of came back to it... Like, 2022 is when I started learning again. And I think that was maybe the peak of the Rust hype, if you were to see a chart. And yeah, I don't know, so it just seemed like the technologies that were real popular at that time period.
+
+It was kind of trippy, because you kind of think you already have this pre -- you have this preconceived notion of what people use, and how popular things are, and... I didn't know that PHP is looked down upon now, and it kind of has a reputation now... Like, I was completely unaware of that at the very beginning.
+
+**Jerod Santo:** That's because you're not on social media, Preston. That's where we talk about these things.
+
+**Adam Stacoviak:** The only place, really. Newsletters, podcasts...
+
+**Jerod Santo:** Yeah, newsletters, YouTube... Although PHP is cool again, because Laravel is cool, Taylor Otwell is cool... Making money on the internet is cool...
+
+**Adam Stacoviak:** Facebook is cool...
+
+**Jerod Santo:** Facebook? I don't know about Facebook.
+
+**Adam Stacoviak:** Oh, you know... It's actually kind of cool. I mean, it's kind of cool that Facebook still exists, and that people can connect so well globally. That's pretty wild.
+
+**Jerod Santo:** Hm.
+
+**Adam Stacoviak:** I mean, I can stalk some people I went to high school with. That's a unique attribute in the world. I like that. I'm not doing it, but I can.
+
+**Jerod Santo:** Oh, you like the idea that you could if you wanted to...
+
+**Adam Stacoviak:** If I want to check somebody out before I become one layer friend deeper, I can. I can be like "Well, this is who you associate with, these are your friend groups, this is where we overlap..." That's pretty wild.
+
+**Jerod Santo:** See, I do a different style. I look at someone's blog and I'm like "Okay, you like TMUX, you like Ghostty, Neovim... Hey, we could be friends."
+
+**Adam Stacoviak:** \[laughs\] Because you're talking to developers, Jerod... \[laughter\] I'm like "Preston must be cool, because we're both using Ghostty", you know?
+
+**Adam Stacoviak:** I do like that though a lot.
+
+**Jerod Santo:** So you made a list of things you're not going to do when you get out. What about -- do you have plans? Like, what you are going to do. Because, I mean, it's coming up, man.
+
+**Preston Thorpe:** Yeah, I think the primary thing is definitely make up for lost time with my family. That is definitely -- my dad's retired now, and to be able to hang out with him every day is like -- it's definitely... Family's number one on that list, for sure. And I've been fortunate enough to be able to -- so I purchased a house right across the street from where they live in Michigan... So the thought of fixing up the house, and learning to do all the handy stuff that I never really learned... So yeah, at-home projects, stuff like that is what I look forward to.
+
+**Adam Stacoviak:** \[01:18:09.28\] I'm waiting for the day that somebody writes the screenplay about your life. "From prison to promised land" is a good ring to it, I think. Are you scared at all? Because, I mean, you said you valued this last three years with finite pressure in the fact that if you wasted 20 minutes talking to somebody, you're like "Man, I did not adhere to my goal of total focus." Are you concerned at all? Or do you have anything that you're trying to do to sort of push back against the fact that you will have unfettered access to the world, and distractions everywhere?
+
+**Preston Thorpe:** Not really something I'm super-concerned about. I think I've kind of cemented my mindset and my goals and priorities now so much that I feel really confident. The main reason that I had a lot of that -- you know, the guilt that I would have previously for wasting time that could have been spent learning was because previously I had no future and no hope for a career and money. So I definitely feel a lot more -- things are pretty concrete now, as far as like I know pretty much exactly what I want to do for a career. I've been really fortunate to meet some just awesome people in the industry, I've gotten a lot of support... So yeah, I feel pretty good about the future.
+
+**Adam Stacoviak:** Any particular names you could drop that you can either thank, or just mention, that you've met, that's like stars, or not stars, or just influential to you personally?
+
+**Preston Thorpe:** Yeah, most definitely. I'm super-thankful for the company I work for, for giving me a chance at the time that I met them. Pretty much none of this would have been possible, at least the way that it played out, without them. Great, great group of people, great mission. Definitely everyone check out Unlocked Labs, if you haven't. We build education tools for incarcerated students, and generally try to bring internet access in a meaningful way to correctional facilities.
+
+I want to give -- we're trying to give more people the opportunity that I was able to have, and to see other people given the chance to be able to change their lives. So definitely, big thanks to Unlocked Labs. Yeah, definitely a big thanks as well to yeah, the whole Turso team, Glauber Costa specifically. He's definitely the reason that a lot of this -- my story got picked up by Prime. Prime is absolutely on that list, too.
+
+\[01:21:42.29\] The Primeagen has been the only source of entertainment that I've allowed myself all this time, and he was a huge influence in the previously mentioned discussion about what technology I use, and chose to learn... He's the reason that I use a tiling window manager, and Tmux, and... His story was one of the first that I've found after kind of deciding to go down this route. Watching his "From meth to Netflix" video was really inspirational. So definitely, big thanks to the Primeagen, for sure.
+
+And yeah, I guess I don't know how broad I'm being here, but I would feel horrible if I didn't mention my family, that has been by my side this whole time. And even when I definitely did not deserve it, they continued to support me, and they stuck around for all the years that I was insufferable, just to -- they're great people, and I'm super-fortunate to have such a great family.
+
+**Adam Stacoviak:** That's awesome. We all need people that pour into us. We realize -- there's a lot of us that can pull our life up by our bootstraps kind of thing, and just sort of muster through and get through whatever we have to, and the variations of that scenario can be very wide, obviously... But we definitely need other people in our human journey to get to where we're trying to go. And it's even better when you've got sort of a peer. "From meth to Netflix", for example; something that you can say "Wow, I can emulate that. I can see now there's hope, because there's somebody who's done it, a version of that", or similarity in some way, shape or form that says "This place I'm at is not the end. There is opportunity for a new beginning, a new change, a new identity, a new future." And having those people pour into you, and the Primeagen, being able to watch him, and being entertained by him... I'm entertained by him. He's pretty wild. I think it's awesome having that.
+
+**Preston Thorpe:** Yeah. And I definitely -- I certainly hope that people that are in situations maybe kind of similar as the one I was in can see what I've done and what I've gone through, and think... It's really tough when your life is that far gone; you feel like you're so far away from just a normal life, or the things you see other people have, careers, and family... You feel like that's so unreachable that putting any effort in that direction just feels feeble. And yeah, I definitely hope that people are able to look at my experience and see just how far I was from any of this, and kind of be able to get some hope out of that.
+
+**Adam Stacoviak:** I'm not sure how much further you want to go, Jerod, but I'm kind of curious about your thoughts, Preston, on this reform idea. I think one of the things you said, either in this conversation or in your blog post, how you advocate for -- you're an example of how access to education and betterment, as an example of an inmate or somebody who's incarcerated changing... What does the world not know about your thoughts or advocacy of that, that we should know?
+
+**Preston Thorpe:** \[01:25:48.26\] Yeah, I think my main point there was - there's a lot of people with similar mindsets to myself, in different facilities that don't have any of the opportunities that they have here... That if given the chance, would flourish and would pursue these things, and give themselves a way out. Because without this skill, and without the network... I guess the skill is one thing, but without the ability to network and the ability to make other people aware of your story... To be in a place with no ability to better yourself, and if all you're given is there's like four phones on the wall, and 200 gang members surrounding you... You could be like even me right now, who's so far from having that mindset and being involved in that lifestyle... If you don't have the means to improve yourself, it's really difficult to expect a different behavior.
+
+I think that's what I want people to consider, is if your expectations are that incarcerating people... You know, it's the department of corrections, it's expected that you're going to learn your lesson and you're not going to do it again... But oftentimes the reality is you take someone that maybe has far less exposure to some of the things they're going to be exposed to in prison, and they spend a few years there and just end up absorbing the criminal mindset and the negativity, and they have absolutely no resources to improve themselves. There's no access to meaningful education, or any way to find a way out of the hole that they dug for themselves. To then expect someone to get out and now have a criminal record, and no work history for N years... It's now just orders of magnitude more difficult to actually improve your situation than it was before this punishment.
+
+So when people -- when the recidivism rate is 70% of people return to prison after I think like three years... The reason for that shouldn't be a big surprise. I think making people aware that investing in education and investing in giving people opportunities to improve themselves, and in ways to find a career path, and meaning... That pays off significantly more.
+
+There's all kinds of statistics, I wish I was good at kind of reciting these things... One of my coworkers is. Shout-out, Jess. He can tell you all the statistics about how every dollar spent on education saves N dollars in recidivism costs, and all that. I'm horrible at remembering those things, but... Yeah, I think that's the general idea. I think people need to be given a way out.
+
+**Adam Stacoviak:** Did you ever study the Stanford prison experiment? Either the movie... I don't think you probably have access to that. Maybe. But the study of it; it was based on a real study conducted in 1971. It was a stark realization of normal people going into these scenarios and role-playing, but becoming very real and very abusive, and psychologically jacked up, I would just say, to paraphrase it... Are you familiar with that experiment?
+
+**Preston Thorpe:** \[01:30:19.13\] I am actually familiar with it. I'm curious if the movie shines any light on some of the details that have come out since then. I read that a lot of the participants weren't really aware that -- I think they were encouraged to kind of go farther than... I don't know, I know this is out there, because if you do any amount of research, you'll see that people will come out and say "Yeah, I was a part of that, but I didn't know it was..." They kind of thought it was more of like a game. I don't know, but that is something that I saw. Did the movie play on that at all, or was it kind of still discussed like it was a very serious thing?
+
+**Adam Stacoviak:** I'm not deeply familiar with it. I did watch the film, and I know it's based on real scientific studies. I'm sorry, my child just busted in here... Get out of here, Micah. Trying to bother me... It was a study done, and it was a psychology professor who had done this... And I think he'd canceled it after six days. It was serious. So to paraphrase the scenario, the film was based on this experiment. And the experiment was real. And the fact that they set up a mock prison, they took normal people in the world, with their normal freedoms, probably in college, is my assumption, or university... And they put them in the scenario where there was guards and prisoners, basically. And the prisoners were not truly in prison, they weren't truly prisoners...Not that prisoners should be treated that way. No offense to you, Preston, but the scenario of power, the struggle of power, where you're in a position of power, and then you're in a position of lack of freedom and lack of power. And the utter role reversal -- these people leaned into their roles, they leaned into all these things, and the prisoners literally had become prisoners, and the guards had literally become guards. And it got so real and abusive that even the professor was becoming abusive. And it's just such a stark reality. And the reason why I even bring this up is just to say that you'd mentioned before how you're advocating for more education access, because this is a way to reform mass incarceration. I think it's not the answer, but it may be one of the answers to some folks who are in prison - not so much all of them, but I can be wrong on that front... But it was just this realization that- wow, how in this scenario you are likely abused, you're obviously accessible to people with power, and they can abuse that power... And that what happens when you lose that freedom and how it psychologically changes you. Because these people that were, like I said, non-prisoners, turned prisoners in this experiment, had meltdowns, breakdowns, because they were being treated like people with no value, people with no hope, people that have been in scenarios you've been in, and they were... It was not good. The efficacy of the experiment was true, and it's evident that in scenarios where you are treated like you're worthless and hopeless, you become worthless and hopeless. And you do not act like normal you, you act like somebody who's been abused. And it's just such a -- if you haven't seen the film or looked at the study, I'd recommend just doing it after this podcast, to sort of just have one more layer of understanding of Preston's life. Or, I'm sure, a version of it.
+
+**Preston Thorpe:** \[01:34:23.04\] Yeah, that's something that Maine -- Maine has come very far in quite a few forefronts of corrections... But one thing that they're big on is little things like language. Inmates here are referred to as residents, and it seems like such a -- I apologize, there's people just getting through... Yeah, so Maine has humanized residents quite a bit. And although the power dynamic is still obvious, and there's still -- there's always going to be... There's like a certain kind of person that is more drawn to roles like police officer, or corrections officer, that are more likely to seek out a position of power over someone... And that obviously still happens.
+
+But yeah, I think certainly not a negligible portion of why I had no hope or care about my life was certainly -- I mean, I can absolutely imagine that some of it might have been from spending a third of my life locked up and treated like a number, and like you're not human, and that how you feel, and your opinions just don't matter... And yeah, it's absolutely a real phenomenon, for sure. I think most people would be surprised. And I would say probably 99% of people that have a loved one, that ends up having to do time, are absolutely shocked by some of the experiences that they'll have. And most people's reaction is always "Well, they can't do that." And it's like, but all of -- your whole preconception of your human rights... That's gone. There is none of that. And a lot of people are absolutely shocked by how the average person is treated. Everything in your mind that kind of makes up your psyche and your own self worth is challenged, because at any given point you can be told to... Yeah, I'm not going to go too far into some of the experiences I've had locked up, just because I'm still here, but... Yeah, it certainly would shock most people.
+
+**Adam Stacoviak:** \[01:37:33.10\] Well, I'm sorry that you've had to go through that. I know that you've been down a road that has been troubling, and I think maybe early on you may have been a threat to society, just based on what you've shared about your choices... And who knows the lives that you've impacted as a result of your choices... But I am a believer in redemption, and I'm so thankful that we live in a society where redemption is an opportunity, and that you've been able to see and have that epiphany about your life, that your identity that you thought was accepted by you, pressed upon you etc. was not going to be your long-term reality. And you put it into your mindset that you're going to change, and I'm glad that you were able to do that.
+
+I'm glad that something as simple as a computer and access to the internet, and your previous curiosity into software, and the fact that open source exists... And that in today's age, that if you can produce, and you can be valuable, then that's all that's required to have access to the world is a thing, versus a degree or something else that you can find rehabilitation in your life, and a flourishing future ahead of you, given future choices that will be exemplary of what you've recently done. Keep making these good choices, keep making these choices that are about your future, and the future you want to have versus the ones that you don't ever want to go back to. So I'm thankful for that for you.
+
+**Preston Thorpe:** I appreciate that.
+
+**Jerod Santo:** Well, Preston, is there anything else on your mind, on your heart, or you were hoping us to ask you, which we haven't asked yet, that we could talk about before we let you go?
+
+**Preston Thorpe:** I think we definitely covered quite a bit. Yeah, I don't know... I think that about wraps it up, probably.
+
+**Jerod Santo:** Well, we appreciate you arranging this, being willing to open up, tell your story, share it with us... We're thankful for the freedoms that you do have inside there, to hop on a podcast and turn your video on, and connect to the internet, and talk to the outside world... So thank you, and I guess thank the folks who are in charge there, for giving you this amount of freedom... Until you get your real freedom back. Hopefully less than a year from now, but... What did you say, a year in the worst case?
+
+**Preston Thorpe:** Yeah, a year in the worst case. May 5th of '26 would be the worst case scenario.
+
+**Jerod Santo:** Would be the worst case. So less than a year.
+
+**Adam Stacoviak:** Get that house next to your parents, make up for lost time, do some fixer-upper, make sure you video it and put it on the internet, because you'll be famous when you do that... And all those good things, man. Good luck to you.
+
+**Jerod Santo:** When you sell the rights to your movie, make sure you get good money.
+
+**Adam Stacoviak:** That's right. And come back to me for the title.
+
+**Preston Thorpe:** Gotcha. I appreciate you guys having me.
+
+**Jerod Santo:** You bet.

Securing ecommerce It's complicated (Interview)_transcript.txt

@@ -0,0 +1,283 @@
+**Jerod Santo:** So I'm here with Ilya Grigorik from Shopify, back on the show after years and years. You've been on the show I think four or five times, Ilya. Welcome back.
+
+**Ilya Grigorik:** Thank you. I'm glad to be back.
+
+**Jerod Santo:** What have you been up to, man? I think it was 2021 last time you were on the show, we were talking to Hydrogen. You're still at Shopify, so you've been there a very long time... What have you been up to?
+
+**Ilya Grigorik:** So I think last time we talked about custom storefronts, and a big mission we had at Shopify to enable developers to build customized storefronts using their own application stack... Since then, I've spent a lot of time diving into our APIs and infrastructure, and then also, kind of in a roundabout way, ended up spending a lot of time in checkout, which at the end of the day is kind of the engine of the entire eCommerce operation. Perhaps an apt analogy is kind of like air traffic controller within your commerce operation, because everything, all the planes have to land there. You can think in different pieces in isolation - you have taxes, you have shipping, you have fulfillment concerns, you have inventory, but all of that has to come together during checkout, where you have all the different policies, all the different negotiations, all the UI that needs to be in place... And that has been a really interesting and complex domain to kind of wrap your head around and navigate through. And of course, I think today we're going to dive into one particular aspect of it, which is the compliance aspect, which I admit is not something that I thought I'd be working on, but it turned out to be a really interesting technical challenge.
+
+**Jerod Santo:** So you've been at Shopify for how long now?
+
+**Ilya Grigorik:** In dog years, it feels like forever. In chronological time, I think it's been four years. But it's been a pressure cooker.
+
+**Jerod Santo:** Yeah. And you before that, I think -- was it GitHub, and Google, or you were at PostRank? You started PostRank? Can you tell us just briefly your travels?
+
+**Ilya Grigorik:** Sure. Let's see how far do we want to rewind. I started my professional career as a founder of a startup. This was back in the 2011 era. Our insight at the time was on the heels of Web 2.0 and all of the social things that are happening, blogs at their heyday, and all the rest... We figured that we could create a better search algorithm. So if you think of PageRank, the original PageRank, of treating links to perform the ranking, effectively that's a thumbs up... Except that when we approached this problem - and actually it wasn't 2011, it was 2008; we observed that there was a lot of extra signals available. There was literal thumbs up from different social platforms, you could leave comments, you can share them on different surfaces... So if we could aggregate all of those signals, we could build a better, kind of human-driven algorithm for identifying what are the interesting topics.
+
+So that was kind of the technical underpinning. And then we went on to build a number of products around it, which were analytics for publishers, to help them understand where their audience is, where the content is being discussed, where people are engaging... There was a product for marketing agencies, which kind of worked in reverse, which is, "Hey, if I have a thing that I'd like to seed, who are the folks that I should be engaging? What are the communities, and all the rest?" And through that work, that led us to Google, which acquired the company, and I ended up working on Google Analytics at the time, integrating a lot of this kind of social analytics know-how that we acquired into the product... And later took a hard pivot into infrastructure, technical infrastructure within Google, where we did a lot of fun things, like building radio towers to figure out if we could build a faster and better radio network, and then learning that that's a hard problem... \[laughter\] But then later, that actually became \[unintelligible 00:07:55.20\] which is an overlay network.
+
+**Jerod Santo:** Oh, yeah.
+
+**Ilya Grigorik:** \[08:00\] And in the process, I picked up the kind of ambiguous problem of "Hey, we keep talking about performance and measuring performance, we want to make it better, but how do you objectively quantify it?" It's one of those things where you kind of know it when you see it; it's like, that felt slow. But if I just asked you to put a technical metric on it, how do you actually measure that?
+
+So we spent a lot of time with browser developers in W3C, I was the co-chair of the W3C Work Performance Working Group, just wrestling with that problem of "How do you measure fast? Is it the onload time? No, not really. Okay, if it's when did pixels paint on the screen - how do you measure that? And which pixels are most important?" So this leads you down those kind of interesting cascade of questions.
+
+So that took a while; that was a good five or six years of my life of working on standards and working with browser developers, which was a lot of fun... And later, I decided to join Shopify, because commerce was clearly an interesting area and a deep domain area... And that's been the last four years of my life here, where I got to work on building custom storefronts, which I think we covered in our last show...
+
+**Jerod Santo:** Yup.
+
+**Ilya Grigorik:** ...of like what is the Shopify opinionated toolkit for building custom experiences. So today, this is actually Hydrogen; that has evolved quite a bit since we last talked. It is a Remix-based stack with a lot of built-ins for building beautiful, customized experiences, powered by Shopify APIs.
+
+From there, that work also led me into API infrastructure, so looking at our GraphQL APIs and trying to understand, first of all, do we expose the right capabilities there? But second, also, once again, performance capabilities and all the rest. We have buyers all around the world, we want to deliver great user experience to all the buyers... So how do you deploy a global cart? And how do you deliver the right experience reliably? And then finally, that led me into the guts of like technical infrastructure, like how do we actually stand up app servers in our Ruby stack? Shopify is a Ruby primarily company, so rebooting our application stack... And also working on checkout, which brings us back to the earlier part of the conversation.
+
+**Jerod Santo:** Yeah, full circle. So we're definitely gonna talk checkouts. Since you've somewhat moved on from your at least web performance years, I'm curious to get your take on recent work, specifically Core Web Vitals. Is that something you've been tracking? And do you have a hot take? Do you like that? Do you think it hits the mark, misses the mark? What are your thoughts on that as a metric?
+
+**Ilya Grigorik:** Core Web Vitals was one of my key projects when I worked at Google.
+
+**Jerod Santo:** Oh, that was you?
+
+**Ilya Grigorik:** So at least to me, this is definitely -- yeah, well, it wasn't purely me, but it was one of the key things that we incubated.
+
+And part of it was -- it's actually the same question, "What is the definition of a vital?" The incentive behind the vital was like a vital signal, just like you have a heartbeat in a human body. Like, what are those things that you measure about a website that tells you that it's a good experience? And the key problem that we wanted to solve was first come up with some shared agreement across browsers of how we can measure that reliably, and not just in the lab environment... Because the thing that we keep learning time and time again is that the outside world is just so unpredictable that you have to measure what happens in the real world. You can bake in all kinds of assumptions into your model, and then you get consistently surprised when you release your application or website into the public, and you're like "Wow, I never expected this amount of traffic to come from this particular region, which happens to have this routing topology, and my CDN just doesn't account for it." Or like "My API is located in North America, but I have this tidal wave of users coming from, I don't know, Europe, or somewhere else, and for them the experience is just that much worse."
+
+\[11:58\] So RUM, or real user measurement metrics are critical, and Web Vitals was our attempt to first of all define what those are, what is that subset, and second, what are the recommended thresholds? Because everyone has a different definition. Is fast 100 milliseconds? Is fast one second? We try to align on that.
+
+So I'm really glad to see that the Web Vitals has continued to evolve. The initial set when we first published - I believe it was back in 2020 or so - it focused on loading metrics. But we knew even when we were walking into that announcement that we really need to also talk about interactivity. Like, it's not just the pixels rendered fast, it's also "Hey, is it responsive?"
+
+**Jerod Santo:** You've gotta use the thing.
+
+**Ilya Grigorik:** Is the page locking up when I'm trying to interact with it? How about scrolling? How smooth is that? So Web Vitals continues to evolve and add those metrics, and I think that's great, and it's really important for us as an industry to have that shared definition of what good looks like.
+
+**Jerod Santo:** It shows how fast the internet moves, because I thought Core Web Vitals was still relatively new, and it turns out it's five years ago, and you were the one working on it, so... Crazy.
+
+**Ilya Grigorik:** I think it highlights the complexity of the problem. It takes a long time to propagate conducts, practices and metrics. But yes, it's been a long, long journey, and it was a great capstone of all the work that I did at Google and web performance. Shipping Web Vitals felt like a good milestone, and that allowed me to kind of give myself permission to shift attention to other things.
+
+**Jerod Santo:** Right. Well, let's do that. Let's shift attention to checkouts, compliance, security, PCI... Some of these things that honestly scare away or maybe lull to sleep many of us. We start talking about compliance matters... PCI version four is burgeoning, or maybe it's out now... I don't know. Tell us the skinny. What is PCI, why does it matter, and then we'll get to what's new in the latest versions.
+
+**Ilya Grigorik:** Sounds good. So first, let's unpack the acronyms, because those are always not helpful. PCI stands for Payment Card Industry, and it defines a set of security requirements that you have to comply with as someone who processes sensitive credentials.
+
+So an example would be your personal access number, your credit card number, plus the CVV and all the other data. There's a set of protections put in place around that for if you're handling that data, then how you should be treated, what kind of security precautions those services must comply by, and all the rest. And it is a fairly burdensome set of requirements to comply with, then you have to get periodically audited and show that you're in good standing, and all the rest. So as a consumer, this is great, because fraud on the internet is definitely a big thing; it's still very much an unsolved problem. Entering a credit card number into a random web form is not a secure undertaking. But we've managed to build a relatively reliable experience for consumers, thankfully.
+
+Now, what's different about PCI before? In PCIv3 a key requirement was that you had to protect the service or the surface area where you you're entering your credentials. So technically, how we've solved that as an industry is we said, "Okay, well, if I want to accept payments on the web, I'm not going to do the obvious thing, which is just put a random form on my page and start accepting credit cards, because then I'm accepting that data, like my service will get a POST request, and I'm going to have the unencrypted payment credentials, and now I'm liable for all the compliance." Instead, why don't we outsource this problem? And hey, we actually have a great tool in the web platform. It's called iFrames. iFrames can give us an ability to embed an external service that can basically do this. And we can skin it in a way that it looks seamless.
+
+\[16:11\] Most pages that you visit on the internet, the payment form, if you actually open up your dev tools, you'll see that it's iFramed for this specific reason, but it doesn't look janky; it looks integrated into the website. And the nice property of this solution is you can then just -- basically all of the inputs, all of the mouse events are obfuscated from the parent page. So that means you can wholly delegate the responsibility for PCI, v3 at least, to your provider.
+
+A common example of that would be someone like Stripe. If you want to accept payments on your website, they provide Stripe elements, you import a web component, pass it a few props, and boom, you have a checkout form under the hood, it'll inject an iFrame, and do all the things on your behalf. So that's great. And that's been effectively where we've settled. I think PCIv3 came into existence around 2013 or 2014. So the last, let's say 10 years or so, that's how we've solved that problem as an industry.
+
+Now, that is really good, but it's not sufficient. So in the interim, what we've observed is "Hey, sure, you've isolated this particular input into a secure sandbox... But what happens if your top-level page gets compromised?" Let's say you have a supply chain attack, or you have an XSS hole in your checkout page, and someone injects a malicious script. What could they do with that? Well, what stops them from removing your secure input form and replacing it with a fake one? Or maybe providing an overlay and then tricking the user effectively into entering information into an insecure form that then exfiltrates the data, and then swaps in the original. This class of attack is called skimming attacks, also known as mage card attacks, which is a nod to Magento. Not to cast shade on Magento, but I think one of the first published large-scale instances of this attack was against Magento, which had some flaw. Magento is an e-commerce platform, open source... And for better or for worse, that mage card attack name has stuck. Now, to be clear, this is a problem that spans all platforms regardless, as long as you have some sort of vector for attack.
+
+So PCIv4 tries to solve for this particular problem. It tries to tighten the perimeter to say "It's no longer sufficient to protect the payment page. You also have to protect or provide some guarantees around the parent page, the thing that is embedding this payment form."
+
+**Jerod Santo:** Okay...
+
+**Ilya Grigorik:** And specifically, there's a set of provisions, I think, in the spec itself - which is very long; we'll zoom in on one particular aspect of this whole conversation, which is section 643. It's one of these random numbers that you just remember once you've been long enough in the PCI game.
+
+**Jerod Santo:** One of my favorite sections...
+
+**Ilya Grigorik:** Yup. 643 defines, in high-level terms, three requirements. It says "Hey, for the parent page, you have to maintain an inventory of all scripts that are being executed on the page. And also, please document why they're necessary and how they're being used." So just like "Give me an inventory." One. Two, once you have that inventory, have some mechanism to ensure that only those scripts, the authorized scripts are being loaded. And then finally, have some way to guarantee or check the integrity of each loaded script. Because you could say, "Hey, this is my inventory, these are the scripts I've authorized, but what if that thing got compromised, as an example?" Somebody replaced it with a malicious script because of a supply chain attack or otherwise.
+
+\[19:59\] So those three things combined give you strong assurances about what's executing in a top level page, which is great. Now, the practical reality of how you go about implementing that, as you can imagine, is complicated.
+
+**Jerod Santo:** It sounds like a lot of work.
+
+**Ilya Grigorik:** Yeah, exactly. And it's complicated for two reasons. First, we should partition this problem into -- like, if the two of us had had a checkout page and we were to sit down and try to think through "Okay, we need to meet these compliance requirements", I would partition the problem into first party and third party scripts, first of all. Like, okay, for the first party content - yes, we can define a process; we audit which scripts we include, we audit their dependency, we have some security review, we have release process, we have CI checks... Sure, I can give you inventory of those scripts. Also, because they're first party, I can put a content security policy, and maybe even put a sub-resource integrity, which are hashes that effectively fingerprint the specific version of that page. So maybe during my build, I can just enforce a CSP snapshot, the hashes, put those on a thing, and great. Now we have strong assurances that the scripts that are being executed on the page are mine, and tied to a specific version. So far, so good. Now, what about the third parties? One of the challenges with checkout is they're one of the most important pages for the entire e-commerce operation.
+
+**Jerod Santo:** For sure.
+
+**Ilya Grigorik:** Like, this is where instrumentation is critical. You want to know performance telemetry. You want to know which elements the user is interacting with, because that affects conversion. You are likely running A/B tests. So you have either first-party or a set of third-party vendors doing that. Of course, you have all the conversion pixels that need to be executed because of all the ad campaigns and all of the analytics that you need to drive that entire up-funnel, down-funnel loop. And what about all of the other marketing pixels that you may need?
+
+It's not uncommon -- if you just take an average checkout page and you open up the network tab, you'll probably find hundreds of scripts on many of them. And oftentimes it'll be "Hey, we load a tag manager that then allows our marketing and other teams to inject whatever they need, to drive the whole process." But now we're staring at that problem and we're saying, "So how exactly do I apply an inventory, and all these things, because first of all, my partner asked me to put a tag manager so they can load things?" Well, I need to unwind that decision, right? Because now I need to know exactly everything that's being executed, and I need to have kind of a full transitive chain of all the dependencies. I need to be able to account for that. Second, you need to provide -- how do I know what CSP policy should I define? Can I just say "Only load from partner.com, or is the partner also loading from some other CDNs?" Well, I need to go ask the partner for what those assurances are.
+
+And then lastly, if I want to ensure integrity, that's not my content. How do I obtain the hash of the thing? And then if that partner wants to rev the version of their script, how do I get the hash so I can put the thing inside, and then also I'm not the one injecting the content into the page. So it becomes this really complicated rigmarole of "Actually, I just cannot do this."
+
+**Jerod Santo:** Yeah, it sounds not possible.
+
+**Ilya Grigorik:** Precisely, which is one of those things where the standard was written with good intent, and they in passing mentioned "Hey, you have these tools; you have content security policy, you have some research integrity. In principle, in theory, you have the right things to do this job. In practice, if you unpack your average checkout page on the web, it's like, I don't know how I would achieve this. I could guarantee maybe a slice of it for first party, but how do I solve this for third parties?"
+
+**Jerod Santo:** Right...
+
+**Ilya Grigorik:** So it turns out it's complicated, right? \[laughter\]
+
+**Jerod Santo:** I thought you might say that.
+
+**Break**: \[24:10\]
+
+**Ilya Grigorik:** So how do we approach this at Shopify - let me take first a branch into Shopify, and then we can talk about kind of the broader landscape.
+
+**Jerod Santo:** Sure.
+
+**Ilya Grigorik:** We've been on a mission to provide stronger control and behavior over checkout. Not just because of compliance, but because we want upgrade safety, reliability, performance and security in checkout. And our observation is -- first of all, for those not familiar, Shopify provides a hosted checkout experience, where you don't get access to the underlying HTML. We provide the base UI, and we allow you to configure it. And it's a very flexible system. You can customize the branding, you can introduce custom components, you can install apps that introduce components... You can do a lot of customizations to make it feel like your own. But a key principle that we've been operating on is that we want a set of predefined -- first of all, we're going to define the UI elements, because we want to preserve consistency in the experience, and we want to optimize for performance, security and all the rest. What that allows us to do is to say, "Actually, we're not going to allow any third-party scripts in our top-level page." And that is a very consequential and big decision. This has been work that we've been on an arc for about three years, if not more, to achieve. We're finally there, and now we're reaping the benefits of that.
+
+So then the question is, "Wait a second, so you excluded all third-party scripts... But what about all those shiny things that you just mentioned earlier?" The analytics, the customizations, the everything else. And this is where sandboxing comes in. So our decision was to say -- effectively, the moment you introduce a third-party script into a top-level page, you have untrusted content and you've compromised all integrity of the top-level page. We cannot provide any assurances on integrity of the top-level page... Because in the past, when we did allow merchants to bring their own JavaScript into top-level page, they just ended up doing things that break compatibility. They'll hook in a specific selector to inject an element, knowing full well that we've never defined a contract for it, and that if we change that, we will break them. And then security is compromised as well, because they're introducing their own scripts, and we can't provide any assurance. So we took away that capability and said, "Instead, we're going to give you a sandbox. So we're going to spin up a set of web workers and give you a bridge." So for example, we've built a library in open source called RemoteDOM, which allows you to construct an element tree in an isolated worker that operates off the main thread... And then that UI is reflected back for you in the parent page. So it feels -- like, ergonomically, DX-wise, it feels still very straightforward, because you're just manipulating elements. And we provide a predefined set of UI elements that fit into the checkout UI, and work with all the branding primitives... But we do that work on your behalf. And the critical part is because we control the bridge between the web worker and the top-level page, we have tight control over what kind of mutations can be pushed between the parent and the isolated worker.
+
+So you can't just arbitrarily inject JavaScript or perform unsafe operations on the parent page. So in that way, we can take any third party script, put it into a sandbox and say, "You know what, you can do whatever the heck you want in that environment." You can load a transitive chain of other dependencies. We don't particularly care, because all we know is that the operations that you can pass back to the parent page are a safe and approved set that we will allow. And we also control what data is exposed to you. So for example, if you have an extension that wants access to some sensitive buyer data, first that application and then the worker itself needs to have the right consent. So a worker that has not been granted the right consent by the merchant or the buyer will just not have access to that data.
+
+\[32:00\] So that is our solution for extensibility, and it allows us to partition the problem of first party and third party content. It's based on remote DOM, and then we use the same technology for pixels or analytics as well, where we define an event bus, we emit all the events, and analytics providers are executed in the sandbox as well.
+
+**Jerod Santo:** Is that a compromise in terms of functionality? Do you get 100% of what you could do before in terms of what you are providing? Or are you constraining people through and losing features along the way?
+
+**Ilya Grigorik:** Yeah, you asked exactly the right question. So the answer is we've had to rebuild a lot of stuff, because a web worker, if you're familiar, is not the same thing as working on the top-level page, right? It doesn't give you access to the DOM, it doesn't expose all the same events. So the reason it took us as long as it did to layer all of this infrastructure is because we had to work with partners and replicate, "So what do you actually need? Instead of raw access to the DOM tree, what are you looking for?"
+
+For example, if you were building a heat map solution, as an example... Some of our merchants are really keen on having a very visual, clear understanding of how users are behaving on their checkout page. You need access to a lot of different events and elements. Okay, well, let's work through that and figure out what is the right subset that we can expose via this bridge. So over time we've built up a collection of these APIs and primitives, some of which effectively replicate what is available on the parent page.
+
+One of the challenges here, by the way, is if you've ever worked with web workers, is they use asynchronous communication. So you have to post a message between a web worker and the top-level page, whereas a lot of the DOM APIs are synchronous APIs. So if you're just naively writing code, expecting it to be executed on a top level page, you would use synchronous APIs. So we've had to shim some of that in places -- we try to keep it as close as we can to what you would expect as a developer, because we don't want to impose additional friction... But in certain places we had to provide replacement APIs where we said "Look, you're building for Shopify. It will operate across a scale of millions of merchants. If you're building an application, it is worth for you to do this extra step, because then you have all of these insurances in place."
+
+**Jerod Santo:** Right.
+
+**Ilya Grigorik:** So a lot of hand-holding with partners and getting the developers to adopt all of those APIs. But the benefit of all of that work today is - I'm not going to say we're done, because there's still more things to build... But we're in a really good place, because now all of our merchants are running on the sandboxed primitive that I've described, and what we can provide is, first of all, upgrade safety. We can safely roll forward our capabilities in checkout, knowing that customizations that we've deployed will not break as you move forward, because we control the bridge, we control the API interface... So if we change the underlying API on our side, we can still provide guarantees about that.
+
+We have reliability. We know that -- for example, we saw examples where merchants would inject scripts where a partner would just timeout. So they would have some logic, and for some reason their service goes down, and then the checkout's broken, because - well, it's just waiting to render. Like "Shopify, you've broken the checkout \[unintelligible 00:35:16.24\] Actually, it's your script that you injected of a partner that failed to scale to your flash sale. So now we have assurances about that.
+
+And then finally, performance and security. Another benefit of putting work into the sandbox is it moves all the work off the main thread. So you can't have code that monopolizes the main thread and renders the UI unresponsive... Which gets back to our web vitals conversation. We \[unintelligible 00:35:45.01\] better performance guarantees about how the page is loaded, how responsive it is, and the rest.
+
+And finally, there's a security bit, which is we know that you can't inject arbitrary content in a top-level page and exfiltrate data... And then finally, you have PCI compliance, because now we have a clean partition where we say "As Shopify, as a platform, we will provide all of the inventory authorization and integrity checks for the first-party scripts that are executed on a top-level page. And oh, by the way, you can totally bring third-party content, but we will execute it in this isolated context, that allows us to punt that problem and not have to worry about all of the integrity problems that happen when you just include it in a top-level page."
+
+**Jerod Santo:** \[36:30\] Hm... So did I hear you right that you said all your merchants are already using this? You're able to deploy that without -- or did you not say all?
+
+**Ilya Grigorik:** Yes, all. So this has been a long journey, to move all of our merchants onto this new platform. But as of earlier this year, 99.9% of our merchants are on this platform. There may be like one or two exceptions, but effectively, any Shopify Power checkout that you visit today as a consumer is running on this infrastructure.
+
+**Jerod Santo:** Hm... And that was something that they had to opt into, or that you just did on a -- like, how'd you all roll that out? You said it took a long time, but what did it look like?
+
+**Ilya Grigorik:** Well, it took a long time because of the right question that you asked, which is "Hey, what did you have to take away?" And the answer is we had to rebuild a lot of the capabilities, because we've created this isolated environment, we've had to recreate a lot of APIs... So a lot of our work was working with other developers, partners who provide capabilities that merchants want in checkout, to make sure that they can bring the same capabilities into this new world of sandbox execution. That was the long haul. And then for some merchants that had ability to manipulate content in a top level page, it was a combination of documentation, handholding, consulting, and just getting them to move to the new world, so they can benefit from all of these capabilities.
+
+**Jerod Santo:** Right.
+
+**Ilya Grigorik:** But we're there, and the time is right, because now you have PCIv4 compliance effectively taking care for you.
+
+**Jerod Santo:** And do you think that PCIv4 compliance means you cannot be skimmed in the way that you could prior? Or do you think it could still happen in new and exciting ways?
+
+**Ilya Grigorik:** Right, right... So I think this actually is another layer that we should add here. What I've described is runtime compliance, or runtime guarantees. So the thing that we've built actually allows us to provide assurance, or like extend some guarantees over - we just know that it's not possible to inject third party content. So if you have a supply chain attack on that, it's isolated into a thing that doesn't matter.
+
+**Jerod Santo:** Right.
+
+**Ilya Grigorik:** In practice, I think what a lot of other players and e-commerce providers will end up using or how they will provide compliance is retroactive monitoring. So PCI does not enforce a requirement that you have to have runtime guarantees. What it says is "Hey, you should have a process that provides an inventory, make sure that scripts are authorized, and you have the integrity." It doesn't specify that it needs to be guaranteed.
+
+So practically, how could you implement this? And how do most -- like, if you go and search for PCI compliance security products, you will find plenty that will basically say "Hey, I know a great solution for your PCI problem. You know what it is? Deploy my JavaScript into your page, because more JavaScript is always a solution... And I will instrument the page and listen for all the things that are happening. I will observe all the other scripts, I will build an inventory, I will monitor if it changes, I will try to provide hashes, and effectively, you can delegate this problem to me."
+
+\[40:00\] Now, you can see a flaw in that reasoning, right? It's like, how do you know that your script is not going to get compromised?
+
+**Jerod Santo:** Right. Who watches the watchers...?
+
+**Ilya Grigorik:** Well, there's that. And how do you know that the malicious thing doesn't come up with a clever way to obfuscate itself from you? It's the antivirus problem.
+
+**Jerod Santo:** Right. Cat and mouse.
+
+**Ilya Grigorik:** Right. A virus hiding from the antivirus problem. But that is likely a solution that many will adopt as a retroactive solution. So effectively, you observe if anything has changed... It's like "Oh, well, that's odd. I'm seeing a set of reports for a script that I did not expect relative to my inventory as I defined. Does that indicate that I have a problem on my site? Probably."
+
+So there's some guardrails that PCI sets for how long that period can be, and how you need to react to it... But it is strictly lesser and a less secure experience. Which gets back to your question. If you have these assurances, does it mean that the class of attacks is eliminated? The answer is "It depends on how you implement it." So in our case, as Shopify, I would feel pretty strongly about extending a promise of like "Yeah, unless our content, first party content is compromised, you'll be very hard to compromise this page."
+
+**Jerod Santo:** Right.
+
+**Ilya Grigorik:** Now, what we can control as Shopify is the buyer has installed a browser extension that injects arbitrary scripts into the page. That is outside of our control, because that operates at a higher layer. Or maybe you even have malware on your computer that does things, and injects content into the page, or otherwise intercepts when you're typing. Those things are still possible. It is not a "We've completely eliminated this type of attack", but it certainly makes it a heck of a lot harder... Because now it means that at least at a minimum the merchants are required to detect these changes, or these attacks, and remediate; so they can't just go unnoticed.
+
+**Jerod Santo:** Mm-hm. So this all sounds like a lot of really good work you all have done at Shopify, for Shopify and Shopify's customers. Thinking bigger, it would be great if your hard work and years of rethinking this runtime, and sandboxing, and actually providing the security that PCIv4 wants everybody to have, whether or not they do or not to be compliant... Can't some of that get into the browser? Could your work extend beyond Shopify's borders and help other people, too?
+
+**Ilya Grigorik:** Yeah, absolutely. And I think this is key. This is not just about Shopify, it's about improving the buyer experience on the on the web, holistically. So two things to answer that. First of all, the RemoteDOM library that I mentioned, it's an open source project that we've built in open source. So if you go to github.com/shopify/remotedom, you'll find that there. Take a look at it, use it. That's the technology that powers Shopify Checkout. Other large companies have already adopted it. I believe Stripe is using it for their apps. Actually, fun story... When we built the project, I think Stripe beat us to using it in a production product.
+
+**Jerod Santo:** Really?
+
+**Ilya Grigorik:** Even though we were the ones developing it for checkout.
+
+**Jerod Santo:** Good for them.
+
+**Ilya Grigorik:** But it is used at Shopify and by other big big players to provide this form of isolation. And the general pattern is "Hey, I have a trusted first party surface into which I want to bring in third party content. And I don't want to compromise the integrity of my first party top level service." Well, RemoteDOM is one of the like technical solutions for that, so please take a look at that. That's answer number one.
+
+\[43:53\] Second, though, and coming back to the browser conversation... Absolutely, the primitives that we have in browsers today - content security policy and SRI - we can make better. And we've actually done a bunch of work on exactly that at Shopify. We don't want to do work in JavaScript that we could push into the browser, because the browser is just much more efficient, and it has capabilities that otherwise would be very hard for us to replicate.
+
+So first, let's enumerate some trivial examples of gaps... Script integrity. Sub resource integrity - for those not familiar, on your script tag you can pass in effectively a hash. So when you inject the tag into your HTML, you can pass in a hash that is a fingerprint, and when the browser loads the script before it executes it, it can compare the hash of the thing that it fetched, versus what you've defined, and say "Hey, if those two things match, great. I will execute the script. Otherwise I'm going to raise a violation and not report this." That's a big capability that exists in browsers today. It's not simple to deploy, but it is doable, because you need to figure out "How do I get these hashes and how do I inject them at the right place?" But then one of the gaps that existed for a long time was module imports. So SRI worked for top level scripts, but if you're building a JavaScript application and you're using an import, you just could not pass in an integrity hash. Why? Well, because module imports came after the sub resource integrity spec was designed. So it was just never a thing.
+
+**Jerod Santo:** Gotcha.
+
+**Ilya Grigorik:** That was a pain point for us, because we use module imports at Shopify, so we worked with Chrome and Safari to upstream some patches to get that supported for module imports. So the good news is that's now baked in, I believe as of May of 2024; that's I think when Safari shipped it in their release. Both Chrome and Safari support SRI for module imports. So that's one.
+
+Another thing that came up in our thinking when we were exploring CSP compliance and how do we make our own life simpler is this idea of require SRI for. So what if you could express a content security policy that says "Hey, all scripts must have an SRI, or integrity hash."
+
+**Jerod Santo:** Gotcha.
+
+**Ilya Grigorik:** And why is that useful? Well, then you can make a strong claim that if you have that policy being enforced by the browser, then if for some reason you sneak through, by accident or malicious act, a script that doesn't have it, they would just be rejected. Which today would just execute normally, without any questions. And even though that might be hard to deploy in an enforcement mode, it could totally work and be really useful in a report-only mode. So for those not familiar with content security policy, you have an enforcement mode and a report-only mode, where you can get violations... Which is incredibly useful, because you could say "Hey, this is a policy I would like to enforce. What are the violations?" So with the require SRI for you could deploy this in report only mode and say "Great, now I'm going to get reliable reports from the browser, from the wild, for any time a browser detects that a script is missing an SRI capability." This is great, because sophisticated attackers would not emit these scripts on every single page load. They might target specific users, or a class of user, or maybe they target a specific browser, or maybe if it's an extension, it'll apply some sort of other heuristic. It's very hard to -- this kind of mirrors our conversation on why RUM is important, real user measurement. Gathering violation reports from real users gives you a much better and reliable signal for where the problems are.
+
+\[47:49\] So require SRI for is another capability that we've shipped into Chrome, and that allows you to get violations on missing SRI attributes, which allows you to build an inventory of like "This is the list for me to burn down and figure out why. And if anything changed, how should I react to it?"
+
+Another example is - okay, great, now we have these reports coming in... Wouldn't it be nice if we could also get the hash of the content? Today you would just get a report saying, "Hey, I detected script from example.com/xyz.gs", but what was the content of that? You don't know, right? Wouldn't it be nice if you could also get a hash in the report such that you could audit it and say "Oh, well, maybe that's totally okay, because the partner revved their version, and it just happens to be the V2. I put that into my approved list and everything's fine." Versus "I have no idea if that was a compromised version or a legitimate version of the script."
+
+**Jerod Santo:** Interesting. So pardon my ignorance for a moment, but where does the reporting take place, or post to? The browsers doing the reporting? Is it --
+
+**Ilya Grigorik:** Correct.
+
+**Jerod Santo:** Who gets the report and how? Is it the browser sends it off somewhere, or...?
+
+**Ilya Grigorik:** Yup. So on the wire, when you emit a page, you can define a content security policy, or CSP policy in a header. And you would define, for a script source, for example a list of origins from which you're allowed to fetch for images and all the rest. You also have a report to target, and a separate report to header, that provides a specification for -- you specify the endpoint to where you want the violation report to be reported.
+
+**Jerod Santo:** Gotcha.
+
+**Ilya Grigorik:** And as good hygiene, that reporting endpoint should ideally be a distinct origin, and all the rest. But you provide a destination. You can find services that will do this for you. They'll just say, "Point your report to us, and we will provide a dashboard in which you can drill down reports, we will aggregate, we'll give you metrics, and all the rest." That's something that we do in-house at Shopify, and I think many other large providers will do on their own... But you could outsource that problem. But just having the ability to even get the report with "Hey, a report has been emitted, because the script is missing an integrity hash" is by itself a really useful capability... Because otherwise you'd probably have to set up some sort of crawling infrastructure that periodically checks your page and says "I accessed this page from five different points on the globe, every 24 hours, and I observed that nothing has changed." Well, that's good, but we could do much better by just actually observing what the real users are seeing, and getting the actual reports of violations.
+
+**Jerod Santo:** Gotcha. So this new one, require SRI for, would work in like manner, as the CSP violations, in terms of reporting and everything else.
+
+**Ilya Grigorik:** Right. The CSP policy is "require SRI for" scripts. So you're saying "All of my script resources must have a hash", and then you can configure that to be a report-only policy, such that it would still execute if the script is missing the hash, but you would get the violation fired in the background. And the browser has its own logic for prioritizing batching delivery and doing all of that, to get you the report.
+
+**Jerod Santo:** Now, do you deploy this one in Shopify?
+
+**Ilya Grigorik:** Yup.
+
+**Jerod Santo:** And do you use it in report mode, or do you have a lockdown mode, or how do you use it?
+
+**Ilya Grigorik:** So for this one it would be a report mode. But it depends on how -- it depends on the shape of your checkout and how much control you have over your first party or third party content. Just to double back on that, for Shopify, for our checkout, we enforce a CSP policy. Actually, let me run through the whole list. For our first party content, we have a well-defined process for vetting all the dependencies, and a process for updates, auditing, to make sure that we provide some guarantees over "If the library that we depend on has been compromised, how can we detect that?"
+
+\[52:17\] We have change management process for it. So this is the reviews, testing, CI, all the things that you would expect. That allows us to create the inventory. We know from where it's served, which means that we can define a strict CSP policy that says "You should only fetch from these sub-origins that we trust." In our build step, we can inject the hashes, the SRI hashes for known content, and we can also emit the "require SRI for" policy, to ensure that if anything else, for some reason if we missed some script, that we would get a violation on that... Because we don't want to break checkout, but we want to be notified immediately if those things are detected; then we can react to it. And we have our own reporting endpoint, which we aggregate, we look at the reports... This is a thing that merchants don't have to worry about, because we do this work on their behalf, and we can provide this guarantee over overall integrity.
+
+And then finally, we've protected the parent page, but the payment credentials page or the payment form itself is also isolated into its own iFrame, just as it was before. So this is a defense in depth; we've protected the parent, but we also have our own implementation of the iFrame, and the full PCI compliance behind that particular form.
+
+**Jerod Santo:** Well, that's a lot of stuff for PCI compliance, Ilya. What happens with v5? How many years are you going to put into that one? \[laughter\]
+
+**Ilya Grigorik:** I don't know. That's a good question. I'm pretty sure that v4 will keep us busy for a long while...
+
+**Jerod Santo:** Yeah, because this is only section 6.4.3, right? That's all we're talking about right here.
+
+**Ilya Grigorik:** That's right. There's all the others.
+
+**Jerod Santo:** Okay, so interesting stuff... It sounds like you've solved some really difficult technical challenges in order to do this in a way that's not just compliant, but actually in the spirit of the compliance as well. Trying to actually make it more secure. What are some takeaways for our listeners out there? Maybe they're doing their own checkout, maybe they have compliance they need to do... Maybe they just want some more secure websites. What do you think they could be thinking walking away from this? If they're not in the actual situation that Shopify is in and having to implement this stuff? What could they learn from this conversation?
+
+**Ilya Grigorik:** Yeah, I think the meta pattern and message to take away here is broadly the integrity and security of first-party versus third-party content. We mix first-party and third-party in most contexts, but even outside of checkout, there are many surfaces. Let's say you have an admin surface or a privileged surface that you only want certain users to access, and you want some sort of extensibility in there. So you want to bring in third-party content, or customization, and all the rest. The pattern that we're describing with isolating third-party content is a generic pattern that you can deploy there. So we use the same sandboxing technology in checkout, we use the same technology in our admin. So for merchants, we allow customizations and third-party developers to bring in their custom UI and other aspects. As you can imagine, that's a very sensitive surface. Order data is there, customer data is there... You don't want to just open up a Pandora's box of "inject your arbitrary JavaScript" because who knows where that data might travel. So the isolation primitive - it may be a remote DOM, it may be something else, but this way of thinking, of isolating into either an iFrame or a worker I think is a pattern that we should be adopting more widely... And it has these additional benefits.
+
+\[56:09\] You have better assurances about security, yes, performance as well, because you're isolating content and moving it off the main thread... You get to define the API contract, so you have better upgradability if you need to maintain that... And I think that's just something that we need to get better at on the web.
+
+Now, the challenge, I think, for all of us industry practitioners is to think through "Boy, the worker is kind of like this naked environment..." We should probably think about "How do we figure out some better set of APIs where we don't have to reinvent the entire wheel?", just as we did at Shopify, for "Great, now I want to build a heat map thing. What does that mean? How do I mirror the entire stream of events, from top-level page into this isolated environment?" I think we can do some thinking and innovation there.
+
+**Jerod Santo:** Very cool. Anything else that's on your mind, that we haven't discussed in this context, or honestly, in any developer context? I always love to hear your opinions on stuff. Anything else on your mind?
+
+**Ilya Grigorik:** I think one really interesting topic, coming back to the world of checkout and commerce, is of course agents, and how agents will interact or how they might affect any of these behaviors.
+
+**Jerod Santo:** Oh, yes... MCP - are you down with MCP? That's the newest acronym, Model Context Protocol. It's burgeoning.
+
+**Ilya Grigorik:** Yup, MCP is definitely top of mind and we're looking at it intently. We're using it for a number of tools and internal services at Shopify. We're also considering if and how we should be exposing MCP as a protocol and endpoint, as a service on behalf of merchants. So imagine you could have a merchant storefront as a remote MCP endpoint... But more broadly, if you think of -- let's imagine you're interacting with an agent, asking it "Hey, I'd like a pair of white sneakers, size 10, $50 to $100 range. Please go find me a pair and checkout." Under the hood, the agent might crawl the web, find the storefront, add to cart, head to checkout, and what does it do then as it's looking at a payment form? Is it the responsibility of the agent to hold on to your payment credentials? And what are the implications of that for entering? How does it enter those credentials? Are there any security and compliance problems or challenges in that? I think that's a wide open question that we as an industry are yet to figure out an answer. Is the human required in that loop? What if it's an accelerated checkout where maybe information is vaulted, right? I think there's a range of questions and answers that we need to figure out in this space.
+
+**Jerod Santo:** Hm... What's your personal thought on "Is the human required in the loop?" How do you feel confidence-wise on removing the human from that loop?
+
+**Ilya Grigorik:** I think it's context-dependent. I think there's definitely a class of commerce in certain types of transactions where I know exactly what I want, there's very low risk, and it's kind of a predefined flow where I just say "Look, I need another carton of milk. You know exactly what I'm looking for, you know where to shop, and please go finish it. I just want it on my front door." And then there's other types of experiences where maybe this is your first time engaging with a merchant, maybe merchant has a set of rules where they actually require the agent to decelerate, because "Hey, for compliance reasons, I may need to verify your age" or "I need you to read this disclaimer on this product before you purchase it." You can't just have the agent blindly ignore that context, or click approve, and then proceed with the transaction.
+
+\[59:56\] So I think we'll need to define some protocol or shared mechanism to signal to agents that "Hey, in this particular case, I need you to pause and ask for a human to either confirm or take over control and complete the transaction."
+
+**Jerod Santo:** There's so many questions there... I just don't feel like I even have the brain right now to analyze all the things that have to be considered. I'm glad that you're -- are you going to be working on this for Shopify? Are you going to stay in PCI land? What's next for you inside? Is this an active thing that you're thinking about for Shopify?
+
+**Ilya Grigorik:** It is definitely an active area of exploration for us. That is one of the things I'm looking with our team and many of our partners who are building these agents, we're trying to figure out "What is the future of checkout, where agents drive some meaningful portion of that experience? What does a good experience even look like in that context?" So I think those are all very interesting and pertinent questions given where we are today.
+
+**Jerod Santo:** Well, we'll have to have you come back in a year or two and let us know what you end up building as you've figured it all out. You seem to have figured it out. At least this hairy technical problem, this new PCI stuff. I'm sure you'll figure out something.
+
+**Ilya Grigorik:** Yeah. We'd love to be back. And at the rate that we're moving in the AI world, in a year or two for now, who knows what will be there...? So we'll see.
+
+**Jerod Santo:** Yes, I'm trying to think of the most recent person who said "Six to nine months and LLMs will be writing a hundred percent of code." So, I mean, who knows, man...? Maybe you and I'll be out on the street corner, talking about this stuff.
+
+**Ilya Grigorik:** I doubt that is the case, but who knows.
+
+**Jerod Santo:** \[laughs\] Yeah, me too. But not a week goes by that somebody doesn't declare "Software engineering is dead", or dying. So I had to squeeze that one in there.
+
+**Ilya Grigorik:** Yes, I think what we're actually saying is the definition of what software engineering is is changing.
+
+**Jerod Santo:** It's gonna have to, isn't it?
+
+**Ilya Grigorik:** I am constantly amazed by what AI is capable of doing in terms of software development, but I'm also constantly surprised by the silly and stupid mistakes that it makes. And oftentimes those mistakes are actually due to misunderstanding or lack of poor definition of the problem that is being solved. It's kind of putting the mirror back to yourself, right? ...because oftentimes you'll find that "Actually, you know what? You did exactly the right thing, the way I expressed it, but that's not what I meant. And I didn't even know what I meant when I typed it, because now that I've seen the mistake, I understand what I was actually trying to get to." So it is this art of defining the problem, and rubber duck programming, and I think we're heading more and more towards a world where we're actively collaborating, instead of hands-on keyboard typing if statements.
+
+**Jerod Santo:** Yeah. The best rubber duck programmers might be the best programmers of the future. The ones who can just talk it out the best, figure it out as you go. Alright, Ilya, appreciate you coming on the show and chatting with us, and looking forward to having you back soon.
+
+**Ilya Grigorik:** Thank you, Jerod.

Securing ecommerce： ＂It's complicated＂_transcript.txt

@@ -0,0 +1,669 @@
+[0.00 --> 11.72]  Welcome to the ChangeLog, where each and every week we sit down with the hackers, the leaders,
+[12.18 --> 17.80]  and the innovators of the software world to pick their brain, to learn from their mistakes,
+[18.10 --> 22.84]  to get inspired by their accomplishments, and to have a lot of fun along the way.
+[22.84 --> 29.50]  On this episode, I'm joined by Ilya Gregorik, distinguished engineer and technical advisor
+[29.50 --> 36.24]  to the CEO at Shopify. Ilya has been hard at work securing e-commerce checkouts from sophisticated
+[36.24 --> 42.48]  new attacks, such as digital skimming, and he's here to share all the technical intricacies and
+[42.48 --> 48.84]  far-reaching implications of this work. But first, a mention of our partners at Fly.io,
+[48.84 --> 55.42]  the public cloud built for developers who ship. You know we love Fly, and you might too. Check
+[55.42 --> 61.66]  them out at Fly.io. Okay, Ilya Gregorik on the ChangeLog. Here we go.
+[67.60 --> 74.44]  Well, friends, I'm here with a good friend of mine, David Hsu, the founder and CEO of Retool.
+[74.44 --> 80.06]  So David, I know so many developers who use Retool to solve problems, but I'm curious. Help me to
+[80.06 --> 87.70]  understand the specific user, the particular developer who is just loving Retool. Who's your
+[87.70 --> 96.84]  ideal user? Yeah, so for us, the ideal user of Retool is someone whose goal, first and foremost,
+[96.84 --> 102.50]  is to either deliver value to the business or to be effective. Where we candidly have a little bit
+[102.50 --> 107.00]  less success is with people that are extremely opinionated about their tools. If, for example,
+[107.00 --> 110.88]  when you're like, hey, I need to go use WebAssembly, and if I'm not using WebAssembly,
+[111.00 --> 115.08]  I'm quitting my job, you're probably not the best Retool user, honestly. However, if you're like,
+[115.08 --> 119.02]  hey, I see problems in the business, and I want to have an impact, and I want to solve those problems,
+[119.34 --> 123.76]  Retool is right up your alley. And the reason for that is Retool allows you to have an impact
+[123.76 --> 128.00]  so quickly. You could go from an idea, you could go from a meeting, like, hey, you know what,
+[128.04 --> 132.84]  this is an app that we need, to literally having the app built in 30 minutes, which is super,
+[132.84 --> 136.98]  super impactful on the business. So I think that's the kind of partnership, or that's the
+[136.98 --> 140.68]  kind of impact that we'd like to see with our customers. You know, from my perspective,
+[140.94 --> 147.36]  my thought is that, well, Retool is well known. Retool is somewhat even saturated. I know a lot
+[147.36 --> 150.86]  of people who know Retool, but you've said this before. What makes you think that Retool is not
+[150.86 --> 157.56]  that well known? Retool today is really quite well known amongst a certain crowd. Like, I think if you
+[157.56 --> 162.82]  had a poll, like, engineers in San Francisco, or engineers in Silicon Valley, even, I think it'd probably get,
+[162.82 --> 169.26]  like, a 50, 60, 70% recognition of Retool. I think where you're less likely to have heard of Retool is
+[169.26 --> 175.16]  if you're a random developer at a random company in a random location, like the Midwest, for example,
+[175.44 --> 180.24]  or like a developer in Argentina, for example, you're probably less likely. And the reason is,
+[180.30 --> 184.32]  I think we have a lot of really strong word of mouth from a lot of Silicon Valley companies,
+[184.32 --> 188.60]  like the Brex's, Coinbase's, DoorDash's, Stripes, et cetera, of the world. There's a lot of chat,
+[188.68 --> 192.34]  Airbnb is another customer, NVIDIA is another customer, so there's a lot of chatter about Retool
+[192.34 --> 197.44]  in the Valley. But I think outside of the Valley, I think we're not as well known. And that's one
+[197.44 --> 203.34]  goal of ours, to go change that. Well, friends, now you know what Retool is, you know who they are,
+[203.52 --> 207.26]  you're aware that Retool exists. And if you're trying to solve problems for your company,
+[207.64 --> 213.58]  you're in a meeting, as David mentioned, and someone mentions something where a problem exists,
+[213.58 --> 221.22]  and you can easily go and solve that problem in 30 minutes, an hour, or some margin of time,
+[221.22 --> 226.24]  that is basically a nominal amount of time. And you go and use Retool to solve that problem.
+[226.46 --> 233.54]  That's amazing. Go to retool.com and get started for free or book a demo. It is too easy to use
+[233.54 --> 238.54]  Retool. And now you know, so go and try it. Once again, retool.com.
+[252.96 --> 261.26]  So I'm here with Ilya Gregorik from Shopify, back on the show after years and years. You've been on the
+[261.26 --> 265.30]  show, I think four or five times, Ilya. Welcome back. Thank you. I'm glad to be back.
+[266.32 --> 269.62]  What have you been up to, man? I think it was 2021. Last time you were on the show,
+[269.66 --> 273.22]  we were talking to Hydrogen. You're still at Shopify. So you've been there a very long time.
+[273.32 --> 277.34]  What have you been up to? So I think, yeah, last time we talked about custom storefronts and
+[277.34 --> 283.20]  a big mission we had at Shopify to enable developers to build customized storefronts using their own
+[283.20 --> 288.68]  application stack. Since then, I've spent a lot of time diving into our APIs and infrastructure.
+[288.68 --> 293.30]  And then also kind of in a roundabout way, ended up spending a lot of time in checkout,
+[293.50 --> 298.34]  which at the end of the day is kind of the engine of the entire e-commerce operation, right? Like
+[298.34 --> 303.26]  an analogy, perhaps an apt analogy is kind of like air traffic controller within your commerce operation,
+[303.26 --> 308.28]  because everything, all the planes have to land there. You can think in different pieces in
+[308.28 --> 313.70]  isolation. You have taxes, you have shipping, you have fulfillment concerns, you have inventory,
+[313.70 --> 318.12]  inventory. But all of that has to come together during checkout, where you have all the different
+[318.12 --> 325.80]  policies, all the different negotiations, all the UI that needs to be in place. And that has been a
+[325.80 --> 331.92]  really interesting and complex domain to kind of wrap your head around and navigate through. And of
+[331.92 --> 335.94]  course, I think today we're going to dive into one particular aspect of it, which is the compliance
+[335.94 --> 341.56]  aspect, which I admit is not something that I thought I'd be working on, but it turned out to be at a
+[341.56 --> 344.96]  really interesting technical challenge. So you've been at Shopify for how long now?
+[345.40 --> 350.60]  In dog years, it feels like forever. In chronological time, I think it's been for four years,
+[350.60 --> 353.54]  but it's been a pressure cooker.
+[354.70 --> 359.46]  Yeah. And you before that, I think, was it GitHub and Google or you're at PostRank?
+[359.82 --> 362.84]  You started PostRank. Can you tell us just briefly your travels?
+[363.32 --> 369.04]  Sure. Let's see how far do we want to rewind. I started my professional career as a founder of a
+[369.04 --> 377.24]  startup. This was back in the 2011 era. And our insight at the time was on the heels of
+[377.24 --> 384.14]  Web2 and all of the social things that are happening, blogs at their heyday and all the rest.
+[385.22 --> 390.96]  We figured that we could create a better search algorithm. So if you think of PageRank as the
+[390.96 --> 397.02]  original PageRank of treating links to perform the ranking, effectively, that's a thumbs up,
+[397.02 --> 401.78]  right? Except that when we approached this problem, and actually it was not, it wasn't 2011,
+[401.88 --> 407.28]  it was 2008. We observed that there was a lot of extra signals available. Like there was literal
+[407.28 --> 412.34]  thumbs up from different social platforms. You could leave comments, you can share them on
+[412.34 --> 415.94]  different surfaces. So if we could aggregate all of those signals, we could build a better kind of
+[415.94 --> 421.84]  human-driven algorithm for identifying what are the interesting topics. So that was the kind of the
+[421.84 --> 426.46]  technical underpinning. And then we went on to build a number of products around it, which were
+[426.46 --> 432.06]  analytics for publishers to help them understand where their audience is, where the content is being
+[432.06 --> 437.74]  discussed, where people are engaging. There was a product for marketing agencies, which kind of worked
+[437.74 --> 442.70]  in reverse, which is, hey, if I have a thing that I'd like to seed, who are the folks that I should be
+[442.70 --> 448.48]  engaging? What are the communities and all the rest? And through that work, that led us to Google,
+[448.48 --> 453.82]  which acquired the company. And I ended up working on Google Analytics at the time, integrating a lot
+[453.82 --> 459.28]  of this kind of social analytics know-how that we acquired into the product, and later took a hard
+[459.28 --> 465.60]  pivot into infrastructure, technical infrastructure within Google, where we did a lot of fun things,
+[465.84 --> 469.84]  like building radio towers to figure out if we could build a faster and better radio network,
+[470.10 --> 477.14]  and then learning that that's a hard problem. But then later that actually became Google Fi,
+[477.14 --> 483.84]  which is an overlay network. And in the process, I picked up the kind of ambiguous problem of,
+[484.52 --> 488.10]  hey, we keep talking about performance and measuring performance, like we want to make it better,
+[488.20 --> 492.40]  but how do you objectively quantify it? It's one of those things where you kind of know it when you
+[492.40 --> 496.72]  see it. It's like that felt slow. But if I just asked you to put a technical metric on it,
+[497.06 --> 503.96]  how do you actually measure that? So we spent a lot of time with browser developers in a W3C. I was a
+[503.96 --> 508.02]  co-chair of the W3C, War Performance Working Group, just wrestling with that problem of,
+[508.34 --> 514.72]  like, how do you measure fast? Like, is it the unload time? No, not really. Okay, if it's like,
+[514.80 --> 519.02]  when did pixels paint on the screen? But how do you measure that? And which pixels are most important?
+[519.78 --> 525.38]  So this leads you down to this kind of interesting cascade of questions. So that took a while.
+[525.92 --> 531.22]  That was a good, you know, five or six years of my life of working on standards and working with
+[531.22 --> 537.18]  browser developers, which was a lot of fun. And later I decided to join Shopify because commerce was
+[537.18 --> 545.46]  clearly an interesting area and a deep domain area. And that's been the last four years of my life here,
+[545.46 --> 551.12]  where I got to work on building custom storefronts, which I think we covered in our last show.
+[551.58 --> 552.02]  Yep.
+[552.22 --> 560.06]  Like, what is the Shopify opinionated toolkit for building custom experiences? So today, this is
+[560.06 --> 566.88]  actually Hydrogen. That has evolved quite a bit since we last talked. It is a remix-based stack
+[566.88 --> 573.02]  for, with a lot of built-ins for building, like, beautiful customized experiences powered by Shopify
+[573.02 --> 579.84]  APIs. From there, that work also led me into API infrastructure. So looking at our GraphQL APIs
+[579.84 --> 585.56]  and trying to understand, first of all, do we expose the right capabilities there? But second,
+[585.56 --> 590.46]  also, once again, performance capabilities and all the rest, we have buyers all around the world.
+[590.78 --> 597.22]  We want to deliver great user experience to all the buyers. So like, how do you deploy a global cart?
+[597.70 --> 604.56]  And how do you deliver the right experience reliably? And then finally, that led me into the
+[604.56 --> 608.96]  guts of like technical infrastructure. Like how do we actually stand up app servers in our, like,
+[609.04 --> 613.74]  Ruby stack? Shopify is a Ruby primarily company, right? So rebooting our application stack,
+[613.74 --> 618.54]  and also working on checkouts, which kind of brings us back to the earlier part of the conversation.
+[619.14 --> 623.22]  Yeah, full circle. So we're definitely going to talk checkouts. Since you somewhat moved on from
+[623.22 --> 630.58]  your, at least web performance years, I'm curious to get your take on recent work, specifically Core Web
+[630.58 --> 635.82]  Vitals. Is that something you've been tracking? And do you have a hot take? Do you like that? Do you
+[635.82 --> 638.94]  think it's hits the mark, misses the mark? What are your thoughts on that as a metric?
+[639.56 --> 641.96]  Core Vitals was one of my key projects when I worked at Google.
+[641.96 --> 646.60]  So it's, it's, it's, it's, at least to me, this is definitely, yeah, well, it wasn't purely me,
+[646.68 --> 652.02]  but it was one of the key things that we incubated. And part of it was, it's actually the same
+[652.02 --> 656.68]  question. Like it, what is the definition of a vital, right? Like a vital is like a vital,
+[656.80 --> 661.24]  then the incentive behind the vital was like a vital signal, just like you have a heartbeat in a human
+[661.24 --> 666.28]  body. Like what are those things that you measure about a website that tells you that it's a good
+[666.28 --> 672.98]  experience. And the key problem that we wanted to solve was first come up with some shared agreement
+[672.98 --> 679.54]  across browsers of how we can measure that reliably and not just in a lab environment, because the
+[679.54 --> 685.50]  thing that we keep learning time and time again is that the outside world is just so unpredictable
+[685.50 --> 692.60]  that you have to measure what happens in the real world. Like you can, you can bake in all kinds of
+[692.60 --> 697.52]  assumptions into your model. And then you get consistently surprised when you release your
+[697.52 --> 702.12]  application or website into the public and you're like, wow, you know, I never expected this amount
+[702.12 --> 707.34]  of traffic to come from this particular region, which happens to have this routing topology and my CDN
+[707.34 --> 712.30]  just doesn't account for it. Or like my API is located in North America, but I have this tidal wave of
+[712.30 --> 718.12]  users coming from, I don't know, Europe or somewhere else. And for them, the experience is just that much
+[718.12 --> 724.08]  worse. So ROM or real user measurement metrics are critical. And Web Vitals was our attempt to,
+[724.22 --> 730.04]  first of all, define what those are, like what does that subset? And second, what are the recommended
+[730.04 --> 735.78]  thresholds? Right? Because everyone has a different definition, like is fast 100 milliseconds, is fast
+[735.78 --> 744.08]  one second? We try to align on that. So I'm really glad to see that the Web Vitals has continued to evolve.
+[744.08 --> 750.56]  Like the initial set, when we first published, I believe it was back in 2020 or so, it focused on
+[750.56 --> 756.02]  loading metrics. But we knew even when we were walking into that announcement that we really need
+[756.02 --> 763.12]  to also talk about interactivity. Like it's not just that the pixels rendered fast, right? It's also,
+[763.48 --> 771.14]  hey, is it responsive? Is the page locking up when I'm trying to interact with it? How about scrolling?
+[771.14 --> 777.68]  Like how smooth is that? So Web Vitals continues to evolve and add those metrics. And I think that's great.
+[777.74 --> 781.76]  And it's really important for us as an industry to have that shared definition of what good looks like.
+[782.26 --> 786.74]  Shows how fast the internet moves. Because I thought core Web Vitals was still relatively new.
+[786.96 --> 791.64]  And it turns out it's like, you know, five years ago and you're the one working on it. So crazy.
+[792.18 --> 797.38]  I think it highlights the complexity of the problem. It takes a long time to propagate
+[797.38 --> 804.04]  kind of those practices and metrics. But yes, it's been a long journey and it was a great capstone
+[804.04 --> 809.36]  of like all the work that I did at Google and web performance. What shipping with Vitals felt like a
+[809.36 --> 815.08]  good milestone and that allowed me to kind of give myself permission to shift attention to other things.
+[815.08 --> 823.54]  Right. Well, let's do that. Let's shift attention to checkouts, compliance, security, PCI, some of these
+[823.54 --> 830.00]  things that honestly scare away or maybe lull to sleep. Many of us, we start talking about compliance
+[830.00 --> 837.52]  matters. PCI version four is burgeoning or maybe it's out now. I don't know. Tell us the skinny. What is
+[837.52 --> 842.46]  PCI? Why does it matter? And then what's, we'll get to what's new in the latest versions.
+[842.46 --> 849.56]  Sounds good. So first, let's unpack the acronyms because those are always not helpful. PCI stands
+[849.56 --> 856.28]  for payment card industry and it provides us, it defines a set of security requirements that you
+[856.28 --> 864.38]  have to comply with as someone who processes sensitive credentials. So an example would be your
+[864.38 --> 870.22]  personal access number or your credit card number plus the CVV and all the other data, right? There's a
+[870.22 --> 875.02]  set of protections put in place around that for if you're handling that data, then how you should be
+[875.02 --> 881.26]  treated, what kind of security precautions those services must comply by and all the rest. And it is
+[881.26 --> 886.98]  a fairly burdensome set of requirements to comply with. Then you have to get periodically audited and
+[886.98 --> 892.60]  show that you're in good standing and all the rest. So as a consumer, this is great because
+[892.60 --> 899.50]  fraud on the internet is definitely a big thing isn't still very much an unsolved problem, right? Like
+[899.50 --> 907.06]  it is entirely entering a credit card number into a web or random web form is not a secure
+[907.06 --> 914.42]  undertaking, right? But we've managed to build a relatively reliable experience for consumers,
+[914.42 --> 922.34]  thankfully. Now what's different about PCI v4? In PCI v3, a key requirement was that
+[922.34 --> 929.70]  you had to protect the service or the surface area where you're entering your credentials. So
+[929.70 --> 934.68]  technically how we've solved that as an industry is we said, okay, well, if I want to accept payments
+[934.68 --> 940.44]  on the web, I'm not going to do the obvious thing, which is going to put a random form on my page and
+[940.44 --> 946.44]  start accepting credit cards because then I'm accepting that data. Like my service will get a
+[946.44 --> 953.12]  post request and I'm going to have the unencrypted payment credentials. And now I'm liable for all of
+[953.12 --> 958.96]  the compliance. Instead, why don't we outsource this problem? And hey, we actually have a great tool
+[958.96 --> 964.88]  in the web platform. It's called iFrames. iFrames can provide us to, can give us an ability to embed
+[964.88 --> 970.38]  an external service that can basically do this. And we can skin it in a way that it looks seamless.
+[970.44 --> 975.88]  Right. Most pages that you visit on the internet, the payment form, if you actually open up your
+[975.88 --> 982.04]  dev tools, you'll see that it's iFramed for the specific reason, but it doesn't look janky. It
+[982.04 --> 987.00]  looks integrated into the website. And the nice property of the solution is you can then just
+[987.00 --> 994.96]  basically all of the inputs, all of the, and like mouse events are obfuscated from the parent page.
+[994.96 --> 1002.66]  So that means you can wholly like delegate the responsibility for PCI v3, at least to your
+[1002.66 --> 1007.16]  provider. A common example of that would be someone like Stripe. If you want to accept payments on your
+[1007.16 --> 1013.66]  website, they provide a like Stripe elements. You import a web component, pass out a few props and
+[1013.66 --> 1019.86]  boom, you have a checkout form under the hood. It'll inject a knife frame and do all the things on your
+[1019.86 --> 1024.66]  behalf. So that's great. And that's been effectively where we've settled. You know, I think PCI v3 came
+[1024.66 --> 1032.62]  into existence around 23 or sorry, 2013 or 2014. So the last, let's say 10 years or so, that's how we've
+[1032.62 --> 1042.06]  solved that problem as an industry. Now that is really good, but it's not sufficient. So in the interim,
+[1042.06 --> 1049.42]  what we've observed is, Hey, sure, you've isolated this particular input into a secure sandbox.
+[1050.20 --> 1054.74]  But what happens if your top level page gets compromised? Let's say you have a supply chain
+[1054.74 --> 1060.26]  attack or you have an XSS hole in your checkout page and someone injects a malicious script.
+[1060.54 --> 1068.04]  What could they do with that? Well, what stops them from removing your secure input form and replacing
+[1068.04 --> 1073.40]  with a fake one or maybe providing an overlay and then tricking the user effectively into entering
+[1073.40 --> 1080.64]  information into an insecure form that then exfiltrates the data and then swaps in the original,
+[1081.16 --> 1088.10]  right? This class of attack is called like skimming attacks, also known as mage card attacks,
+[1088.50 --> 1093.88]  which is a nod to Magento, not to cast shade on Magento, but I think one of the first like published
+[1093.88 --> 1100.04]  large scale instances of this attack was against Magento, which had some flaw. Magento is an
+[1100.04 --> 1105.94]  e-commerce platform, open source. And for better or for worse, that like mage card attack name has stuck.
+[1106.34 --> 1112.74]  Now, to be clear that this is a problem that spans all platforms regardless, right? As long as you have
+[1112.74 --> 1119.78]  some sort of vector for attack. So PCI before tries to solve for this particular problem. Like it tries to
+[1119.78 --> 1126.34]  tighten the perimeter to say, it's no longer sufficient to protect the payment page. You also
+[1126.34 --> 1132.00]  have to protect or provide some guarantees around the parent page, the thing that is embedding this
+[1132.00 --> 1139.92]  payment form. Okay. And specifically, there's a set of set of provisions. I think that in the spec
+[1139.92 --> 1144.42]  itself, which is very long, like we'll zoom in on one particular aspect of this whole conversation,
+[1144.42 --> 1149.98]  which is like section 643. It's one of these random numbers that you just remember once you've
+[1149.98 --> 1159.22]  been long enough in the PCI game. Yep. 643 defines in high level terms, three requirements. It says,
+[1159.38 --> 1164.56]  hey, for the parent page, you have to maintain an inventory of all scripts that are being executed
+[1164.56 --> 1169.62]  on the page. And also please document why they're necessary and how they're being used.
+[1170.62 --> 1177.32]  Right. So just like, give me an inventory one. Two, once you have that inventory, have some mechanism
+[1177.32 --> 1182.26]  to ensure that only those scripts, the authorized scripts are being loaded. And then finally,
+[1182.80 --> 1188.72]  have some way to guarantee or check the integrity of each loaded script, right? Because you could say,
+[1188.84 --> 1193.34]  hey, this is my inventory. These are the scripts I've authorized, but what if that thing got compromised
+[1193.34 --> 1198.30]  as an example, right? Somebody replaced it with a malicious script because of supply chain attack
+[1198.30 --> 1204.36]  or otherwise. So those three things combined give you strong assurances about what's executing a
+[1204.36 --> 1209.50]  top level page, which is great. Now, the practical reality of how you go about implementing that,
+[1209.74 --> 1213.22]  as you can imagine, is complicated. Sounds like a lot of work.
+[1213.74 --> 1217.88]  Yeah, it's good. Exactly. And it's complicated for two reasons. Like first, we should partition this
+[1217.88 --> 1222.48]  problem into like if the two of us had had a checkout page and we were to sit down and try to
+[1222.48 --> 1228.58]  think through like, okay, we need to meet these compliance requirements. I would partition the
+[1228.58 --> 1233.06]  problem into first party and third party scripts, first of all, right? Like, okay, for the first party
+[1233.06 --> 1240.42]  content, yes, like we can define a process for we audit which scripts we include, we audit their
+[1240.42 --> 1246.28]  dependency, we have some security review, we have release process, we have CI checks. Sure, I can give you
+[1246.28 --> 1252.24]  inventory of those scripts, right? Also, because they're first party, I can put a content security
+[1252.24 --> 1258.00]  policy and maybe even put a sub resource integrity, which are hashes that effectively fingerprint the
+[1258.00 --> 1265.74]  specific version of that page. So maybe during my build, I can just enforce a CSP, snapshot the hashes,
+[1265.92 --> 1272.10]  put those on the thing and like, great. Now I have strong assurances that the scripts that are being
+[1272.10 --> 1280.32]  executed on the page are mine and tied to a specific version. So far, so good. Now, what about the third
+[1280.32 --> 1287.76]  parties? One of the challenges with checkout is they're one of the most important pages for the
+[1287.76 --> 1294.66]  entire e-commerce operation. Like this is where instrumentation is critical, right? You want to know
+[1294.66 --> 1299.60]  performance telemetry. You want to know which elements user is interacting with because that
+[1299.60 --> 1306.96]  affects conversion. You are likely running A-B tests. So you have either first party or a set of third
+[1306.96 --> 1312.24]  party vendors doing that. Of course, you have all of the conversion pixels that need to be executed
+[1312.24 --> 1316.48]  because all the ad campaigns and all of the analytics that you need to drive that entire up
+[1316.48 --> 1323.14]  funnel, down funnel loop. And what about all of the other marketing pixels that you may need?
+[1324.42 --> 1330.34]  It's not uncommon. If you just, if you take an average checkout page and you open up the network tab,
+[1330.34 --> 1337.30]  you'll probably find hundreds of scripts on many of them, right? And oftentimes it'll be,
+[1337.94 --> 1344.94]  hey, we load a tag manager that then allows our marketing and other teams to inject whatever they
+[1344.94 --> 1353.92]  need to drive the whole process. But now we're staring at that problem and we're saying, so how exactly do I
+[1353.92 --> 1359.80]  apply an inventory and all these things? Because first of all, like my partner asked me to put a tag
+[1359.80 --> 1365.52]  manager so they can load things. Well, I need to unwind that decision, right? Because now I need to
+[1365.52 --> 1370.12]  know exactly what everything that's being executed and I need to have kind of a full transit of chain
+[1370.12 --> 1377.90]  of all dependencies. I need to be able to account for that. Second, you need to provide, how do I know
+[1377.90 --> 1384.34]  what CSP policy should I define? Can I just say only load from partner.com or is the partner also
+[1384.34 --> 1389.98]  loading from some other CDNs? Well, you know, that's, I need to go ask the partner for what those
+[1389.98 --> 1396.32]  assurances are. And then lastly, if I want to ensure integrity, that's not my content. How do I obtain
+[1396.32 --> 1401.26]  the hash of the thing? And then if that partner wants to rev the version of their script, how do I
+[1401.26 --> 1407.06]  get the hash so I can put the thing inside? And then also I'm not the one injecting the content
+[1407.06 --> 1412.00]  into the page. So it becomes this like a really complicated rigmarole of like, actually, I just cannot do this.
+[1412.00 --> 1418.80]  Yeah. Sounds not possible. Precisely. Which is one of those things where the standard was written
+[1418.80 --> 1423.36]  with good intent, right? And they in passing mentioned, Hey, you have these tools, you have
+[1423.36 --> 1428.84]  content security policy, you have sub research integrity in principle, in theory, you have the
+[1428.84 --> 1435.08]  right things to do this job. In practice, if you unpack your average checkout page on the web, it's like,
+[1435.38 --> 1439.52]  I don't know how I would achieve this. I could guarantee maybe a slice of it for first party,
+[1439.52 --> 1446.00]  but how do I solve this for third parties? Right. So turns out, it's complicated, right?
+[1446.64 --> 1449.62]  So I thought you might say that.
+[1469.52 --> 1481.46]  Well, friends, I'm here with Scott Dietzen, CEO of Augment Code. Augment is the first AI coding
+[1481.46 --> 1486.86]  assistant that is built for professional software engineers and large code bases. That means context
+[1486.86 --> 1492.66]  aware, not novice, but senior level engineering abilities. Scott Flex for me, who are you working
+[1492.66 --> 1495.04]  with? Who's getting real value from using Augment Code?
+[1495.04 --> 1501.22]  So we've had the opportunity to go into hundreds of customers over the course of the past year,
+[1501.36 --> 1506.84]  and show them how much more AI could do for them. Companies like Lemonade, companies like
+[1506.84 --> 1512.96]  Kodem, companies like Lineage and Webflow. All of these companies have complex code bases. If I take
+[1512.96 --> 1518.24]  Kodem, for example, they help their customers modernize their e-commerce infrastructure. They're
+[1518.24 --> 1523.18]  showing up and having to digest code they've never seen before in order to go through and make these
+[1523.18 --> 1528.36]  essential changes to it. We cut their migration time in half because they're able to much more
+[1528.36 --> 1534.04]  rapidly ramp, find the areas of the code base, the customer code base that they need to perfect and
+[1534.04 --> 1538.88]  update in order to take advantage of their new features. And that work gets done dramatically more
+[1538.88 --> 1540.82]  quickly and predictably as a result.
+[1541.48 --> 1546.44]  Okay, that sounds like not novice, right? Sounds like senior level engineering abilities.
+[1547.14 --> 1552.14]  Sounds like serious coding ability required from this type of AI to be that effective.
+[1552.14 --> 1557.68]  100%. You know, these large code bases, when you've got tens of millions of lines in a code base,
+[1557.74 --> 1562.88]  you're not going to pass that along as context to a model, right? That would be so horrifically
+[1562.88 --> 1568.78]  inefficient. Being able to mine the correct subsets of that code base in order to deliver AI insight
+[1568.78 --> 1576.18]  to help tackle the problems at hand. How much better can we make software? How much wealth can we release
+[1576.18 --> 1582.50]  and productivity can we improve if we can deliver on the promise of all these feature gaps and tech debt?
+[1582.74 --> 1588.04]  AIs love to add code into existing software. You know, our dream is an AI that wants to delete code,
+[1588.18 --> 1592.80]  make the software more reliable rather than bigger. I think we can improve software quality,
+[1593.30 --> 1599.06]  liberate ourselves from tech debt and security gaps and software being hacked and software being fragile
+[1599.06 --> 1604.62]  and brittle. But there's a huge opportunity to make software dramatically better. But it's going to take an AI
+[1604.62 --> 1608.44]  that understands your software, not one that's a novice.
+[1609.02 --> 1613.88]  Well, friends, Augment taps into your team's collective knowledge, your code base, your documentation,
+[1614.12 --> 1619.24]  dependencies, the full context. You don't have to prompt it with context. It just knows.
+[1619.70 --> 1622.80]  Ask it the unknown unknowns and be surprised.
+[1622.80 --> 1629.18]  It is the most context aware developer AI that you can even tap into today. So you won't just write code
+[1629.18 --> 1635.04]  faster. You'll build smarter. It is truly an ask me anything for your code. It's your deep thinking buddy.
+[1635.52 --> 1645.98]  It is your stay in flow antidote. And the first step is to go to augmentcode.com. That's A-U-G-M-E-N-T-C-O-D-E.com.
+[1646.32 --> 1651.52]  Create your account today. Start your free 30 day trial. No credit card required.
+[1651.52 --> 1654.22]  Once again, augmentcode.com.
+[1658.06 --> 1662.76]  So how do we approach this at Shopify? I think there's, let me take first a branch into Shopify
+[1662.76 --> 1668.00]  and then we can talk about kind of the broader, broader landscape. We've been on a mission to
+[1668.00 --> 1674.74]  provide stronger control and behavior over checkout, not just because of compliance, but because we want
+[1674.74 --> 1680.86]  upgrade safety, reliability, performance and security in checkout. And our observation is
+[1680.86 --> 1686.54]  first of all, for those not familiar, Shopify provides a hosted checkout experience where
+[1686.54 --> 1695.42]  you don't get access to the underlying HTML, right? We provide the base UI and we allow you to
+[1695.42 --> 1699.38]  configure it. And it's a very flexible system. You can customize the branding. You can custom,
+[1699.48 --> 1703.32]  you can introduce custom components. You can install apps that introduce components.
+[1703.32 --> 1709.32]  You can do a lot of customizations to make it feel like your own. But a key principle that we've
+[1709.32 --> 1714.94]  been operating on is that we want a set of predefined, first of all, we're going to define the UI
+[1714.94 --> 1721.84]  elements because we want to preserve consistency and experience. And we want to optimize for performance,
+[1722.10 --> 1729.70]  security and all the rest. What that allows us to do is to say, actually, we're not going to allow
+[1729.70 --> 1737.12]  any third-party scripts in our top-level page. And that is a very consequential and big decision.
+[1737.28 --> 1744.04]  This has been work that we've been on an arc for about three years, if not more, to achieve. And
+[1744.04 --> 1749.14]  we're finally there. And now we're reaping the benefits of that. So then the question is,
+[1749.70 --> 1753.34]  wait a second. So you excluded all third-party scripts, but what about all those shiny things that
+[1753.34 --> 1757.40]  you just mentioned earlier, right? The analytics, the customizations, the everything else.
+[1757.40 --> 1764.24]  And this is where sandboxing comes in. So our decision was to say, only effectively, the moment
+[1764.24 --> 1769.50]  you introduce a third-party script into top-level page, you have untrusted content and you've
+[1769.50 --> 1775.00]  compromised all integrity of the top-level page. We cannot provide any assurances on integrity of
+[1775.00 --> 1782.72]  the top-level page, right? Because in the past, when we did allow folks or merchants to bring their
+[1782.72 --> 1789.34]  own JavaScript into top-level page, they just end up doing things that break compatibility.
+[1789.34 --> 1793.98]  Like they'll hook in a specific selector, right, to inject an element,
+[1794.56 --> 1799.62]  knowing full well that we've never defined a contract for it. And then if we change that,
+[1799.70 --> 1804.56]  we will break them. And then security is compromised as well, because they're introducing
+[1804.56 --> 1806.42]  their own scripts and we can't provide any assurance.
+[1806.42 --> 1811.74]  So we took away that capability and said, instead, we're going to give you a sandbox.
+[1812.20 --> 1818.06]  So we're going to spin up a set of web workers and give you a bridge. So for example, we've built a
+[1818.06 --> 1824.96]  library and open source called Remote Dumb, which allows you to construct an element tree in an
+[1824.96 --> 1832.42]  isolated worker that operates off the main thread. And then that UI is reflected back for you in the
+[1832.42 --> 1839.64]  parent page. So it feels like ergonomically, DX-wise, it feels still very straightforward
+[1839.64 --> 1844.82]  because you're just manipulating elements. And we provide a predefined set of UI elements that
+[1844.82 --> 1851.94]  fit into the checkout UI and work with all the branding primitives. But we do that work on your
+[1851.94 --> 1856.30]  behalf. And the critical part is because we control the bridge between the web worker and the top-level
+[1856.30 --> 1862.04]  page, we have tight control over what kind of mutations can be pushed between the parent,
+[1862.04 --> 1868.82]  and the isolator worker. So you can't just arbitrarily inject JavaScript or perform unsafe operations
+[1868.82 --> 1876.18]  on the parent page. So in that way, we can take any third-party script, put it into a sandbox and say,
+[1876.26 --> 1881.58]  you know what, you can do whatever the heck you want in that environment, because you can load
+[1881.58 --> 1888.12]  a transit of chain of other dependencies. We don't particularly care because all we know that
+[1888.12 --> 1894.96]  the operations that you can pass back to the parent page are safe and approved set that we will allow.
+[1895.72 --> 1901.76]  And we also control what data is exposed to you. So for example, if you have an extension that wants
+[1901.76 --> 1911.00]  access to some sensitive buyer data, first that application and then the worker itself needs to have the right
+[1911.00 --> 1918.10]  consent. So a worker that has not been granted the right consent by the merchant or the buyer will just not have
+[1918.10 --> 1924.68]  access to that data. So that is our solution for extensibility and allows us to partition the
+[1924.68 --> 1930.66]  problem of first-party and third-party content. It's based on remote DOM. And then we use the same
+[1930.66 --> 1937.04]  technology for our pixels or analytics as well, where we define an event bus, we emit all the events,
+[1937.38 --> 1941.80]  analytics providers are executed in the sandbox as well.
+[1942.42 --> 1947.30]  Is that a compromise in terms of functionality? Do you get 100% of what you could do before in terms of
+[1947.30 --> 1953.00]  what you all are providing or is it like a, are you constraining people through and losing features
+[1953.00 --> 1953.54]  along the way?
+[1954.04 --> 1958.98]  Yeah, you asked exactly the right question. So the answer is we've had to rebuild a lot of stuff
+[1958.98 --> 1964.46]  because a web worker, if you're familiar, is not the same thing as working in the top level page,
+[1964.52 --> 1967.94]  right? Like it doesn't, it doesn't give you access to the DOM. It doesn't expose all the same events.
+[1967.94 --> 1974.48]  So the reason it took us as long as it did to layer all of this infrastructure is because we had to work
+[1974.48 --> 1979.52]  with partners and replicate. So what do you actually need, right? Instead of raw access to the DOM tree,
+[1979.70 --> 1983.98]  like what are you looking for? For example, if you were building a heat map solution as an example,
+[1984.18 --> 1989.58]  right? Like some of our merchants are really keen on having very visual, clear understanding of how users
+[1989.58 --> 1995.62]  are behaving on their checkout page. Like you need a lot of different access to a lot of different events
+[1995.62 --> 2000.06]  elements. Okay. Well, let's, let's work through that and figure out what is the right subset that
+[2000.06 --> 2006.52]  we can expose via this bridge. So over time, we've built up a collection of these APIs and primitives,
+[2006.66 --> 2012.88]  some of which effectively replicate what is available on the parent page. One of the challenges
+[2012.88 --> 2017.74]  here, by the way, is like, if you ever work with web workers is they use asynchronous communication,
+[2018.00 --> 2022.70]  right? So you have to post message between a web worker and the top level page. Whereas a lot of the
+[2022.70 --> 2028.02]  DOM APIs are synchronous APIs. So if you're just naively writing code, expecting to be executed
+[2028.02 --> 2034.44]  on top level page, you would use synchronous APIs. So we've had to shim some of that. And like in places
+[2034.44 --> 2039.84]  we try to keep it as close as we can to what you would expect as a developer, because we don't want to
+[2039.84 --> 2044.96]  impose additional friction. But in certain places we had to provide replacement APIs where we said,
+[2044.96 --> 2049.88]  look, you're building for Shopify. It will operate across scale of millions of merchants.
+[2049.88 --> 2054.44]  If you're building an application, it is worth for you to do this extra step because then you have
+[2054.44 --> 2060.76]  all of these insurances in place. So a lot of handholding with partners and getting the developers
+[2060.76 --> 2067.62]  to adopt all of those APIs. But the benefit of all of that work today is I'm not going to say we're done
+[2067.62 --> 2072.18]  because there's still more things to build, but we're in a really good place because now all of our
+[2072.18 --> 2080.38]  merchants are running on the sandboxed primitive that I've described. And what we can provide is,
+[2080.62 --> 2084.30]  first of all, upgrade safety. We can safely roll forward our capabilities in checkout,
+[2084.84 --> 2092.62]  knowing that customizations that you've deployed will not break as you move forward, right? Because
+[2092.62 --> 2097.20]  we control the bridge, we control the API interface. So if we change the underlying API on our site,
+[2097.20 --> 2103.82]  we can still provide guarantees about that. We have reliability. We know that. So for example,
+[2103.90 --> 2109.34]  we saw examples where merchants would inject scripts where a partner would just time out.
+[2109.62 --> 2113.52]  So they would have some logic and for some reason their service goes down and then the checkout's
+[2113.52 --> 2118.30]  broken because, well, it's just waiting to render, right? Like Shopify, you've broken the checkout.
+[2118.46 --> 2122.98]  It's like, actually, it's your part. It's your script that you injected of a partner that failed
+[2122.98 --> 2129.78]  to scale to your flash sale, right? So now we have assurances about that. And then finally,
+[2129.88 --> 2136.24]  performance and security. Another benefit of putting work into the sandbox is it moves all
+[2136.24 --> 2143.16]  the work off the main thread. So you can't have code that monopolizes the main thread and renders the UI
+[2143.16 --> 2147.44]  unresponsive, which gets back to our web vitals conversation, right? Like we can make better
+[2147.44 --> 2151.92]  performance guarantees about how the page is loaded, how responsive it is, and the rest. And finally,
+[2151.92 --> 2157.20]  there's a security bit, which is, we know that you can't inject arbitrary content in top level page
+[2157.20 --> 2162.58]  and exfiltrate data. And then finally you have PCI compliance because now we have a clean partition
+[2162.58 --> 2169.98]  where we say, as Shopify as a platform, we will provide all of the inventory authorization and
+[2169.98 --> 2175.20]  integrity checks for the first party scripts that are executed in top level page. And, oh, by the way,
+[2175.32 --> 2179.96]  you can totally bring third party content, but we will execute it in this isolated context
+[2179.96 --> 2186.74]  that allows us to punt that problem and not have to worry about all of the integrity
+[2186.74 --> 2190.52]  problems that happen when you just include it in a top level page.
+[2190.52 --> 2196.18]  So did I hear you right that you said all your merchants are already using this? You're able
+[2196.18 --> 2199.84]  to deploy that without, or did you not say all?
+[2199.84 --> 2199.94]  Yeah.
+[2200.40 --> 2200.70]  Yes.
+[2200.78 --> 2200.92]  Yeah.
+[2201.24 --> 2207.24]  Yes. All. Yeah. So this has been a long journey to move all of our merchants onto this new platform.
+[2207.24 --> 2213.34]  But as of earlier this year, like 99.9% of our merchants are on this platform. There may be like
+[2213.34 --> 2218.66]  one or two exceptions, but effectively any Shopify checkout that you, Shopify power checkout that you
+[2218.66 --> 2221.16]  visit today as a consumer is running on this infrastructure.
+[2221.70 --> 2227.72]  Hmm. And that was something that they had to opt into or that you just did on it. Like,
+[2227.76 --> 2231.20]  how did you all roll that out? You said it took a long time, but what was, what did it look like?
+[2231.20 --> 2235.76]  Well, it took a long time because of the right question that you asked, which is,
+[2236.10 --> 2242.04]  Hey, did you, what did you have to take away? Right. And the answer is we had to rebuild a lot
+[2242.04 --> 2246.06]  of the capabilities because we've created this isolated environment. We've had to recreate a lot
+[2246.06 --> 2252.80]  of APIs. So a lot of our work was working with other developers, partners who provide capabilities
+[2252.80 --> 2259.72]  that merchants want in checkout to make sure that they can bring the same capabilities into this new
+[2259.72 --> 2265.60]  world of sandbox execution. That's, that was the long haul. And then for some merchants that
+[2265.60 --> 2272.62]  had ability to manipulate content in the top level page, you know, it was a combination of
+[2272.62 --> 2279.42]  documentation, handholding, consulting, and just getting them to move to the new world so they can
+[2279.42 --> 2285.76]  benefit from all of these capabilities. But we're, we're there and the time is right because now you
+[2285.76 --> 2288.36]  have PCI v4 compliance effectively taking care for you.
+[2289.72 --> 2298.76]  And do you think that PCI v4 compliance means you cannot be skimmed in the way that you could prior,
+[2298.76 --> 2302.12]  or do you think it could still happen in new and exciting ways?
+[2303.12 --> 2310.30]  Right. Right. So I think this actually is, is another layer that we should add here. What I've
+[2310.30 --> 2317.54]  described is runtime compliance or runtime guarantees, right? So the thing that we've built actually
+[2317.54 --> 2322.58]  allows us to provide assurance or like extend some guarantees over, we just know that it's not,
+[2322.58 --> 2328.06]  it's not possible to inject third party content. So if you have a supply chain attack on that,
+[2328.18 --> 2330.62]  like it's isolated into a thing that doesn't matter.
+[2330.62 --> 2339.06]  Right. In practice, I think what a lot of other players are and e-commerce providers will end up
+[2339.06 --> 2346.88]  using or how they'll provide compliance is retroactive monitoring. So PCI does not enforce a requirement
+[2346.88 --> 2352.46]  that you have to have runtime guarantees. What it says is, Hey, you should have a process
+[2352.46 --> 2359.10]  that provides an inventory, make sure that scripts are authorized and you have the integrity. It
+[2359.10 --> 2363.76]  doesn't specify that it like, it needs to be guaranteed. So practically how, how could you
+[2363.76 --> 2369.52]  implement this and how, how do most, like if you go and search for PCI compliance security products,
+[2369.52 --> 2374.60]  you will find plenty that will basically say, Hey, I know a great solution for your PCI problem.
+[2374.60 --> 2381.56]  You know what it is? Deploy my JavaScript into your page. Right. Because more JavaScript is always a
+[2381.56 --> 2387.58]  solution. Right. And I will instrument the page and listen for all the things that are happening.
+[2387.96 --> 2393.54]  I will observe all the other scripts. I will build an inventory. I will monitor if it changes.
+[2393.90 --> 2399.34]  I will try to provide hashes and effectively I'll like, you can delegate this problem to me.
+[2399.34 --> 2405.08]  Now you can see a flaw in that reasoning, right? It's like, how do you know that your script is not
+[2405.08 --> 2409.96]  going to get compromised either? Right. Who watches the watchers? Right. Well, there's that. And how do
+[2409.96 --> 2415.18]  you know that the malicious thing doesn't come up with a clever way to obfuscate itself from you?
+[2415.18 --> 2422.20]  Right. It's the antivirus problem, right? Like virus hiding, exactly. Virus hiding from the antivirus
+[2422.20 --> 2429.32]  problem. Um, but that is likely a solution that many will adopt as a retroactive.
+[2429.34 --> 2435.68]  So effectively you observe if anything has changed. It's like, oh, well, that's odd. I'm
+[2435.68 --> 2441.90]  seeing a set of reports for a script that I did not expect relative to my inventory as I defined.
+[2442.54 --> 2447.50]  Does that indicate that I have a problem on my site? Probably. Right. So there's some guardrails
+[2447.50 --> 2454.24]  that PCI sets for like how long that period can be and how you need to react to it. But it is strictly
+[2454.24 --> 2459.70]  lesser and less secure experience, which gets back to your question. Like if you have these
+[2459.70 --> 2466.16]  assurances, it doesn't mean that the class of attacks is eliminated. The answer is it depends
+[2466.16 --> 2471.54]  on how you implement it. Right. So in our case as Shopify, I would feel pretty strongly about
+[2471.54 --> 2478.70]  extending a promise of like, yeah, unless our content, first party content is compromised,
+[2478.70 --> 2486.34]  it would be very hard to, to compromise this page. Now we can control a Shopify is the buyer
+[2486.34 --> 2491.02]  has installed a browser extension that injects arbitrary scripts into the page, right? Like
+[2491.02 --> 2495.28]  that is outside of our control because that's a, that operates at a higher layer, or maybe
+[2495.28 --> 2500.92]  you even have malware on your computer that does things and inject content into the page
+[2500.92 --> 2506.84]  or otherwise intercepts like when you're typing. Like those things are still possible. It is not a,
+[2507.48 --> 2512.88]  we've completely eliminated this type of attack, but it certainly makes it a heck of a lot harder
+[2512.88 --> 2519.00]  because now it means that at least there at a minimum, there's a way the merchants are required
+[2519.00 --> 2525.74]  to detect these changes or the, these, um, attacks and remediate so they can't just go unnoticed.
+[2525.74 --> 2533.10]  Mm-hmm. So this all sounds like a lot of really good work you all have done at Shopify for Shopify
+[2533.10 --> 2543.12]  and Shopify's customers. Thinking bigger, it would be great if your hard work and years of rethinking
+[2543.12 --> 2551.20]  this runtime and sandboxing and actually providing the security that PCI v4 wants everybody to have,
+[2551.20 --> 2557.20]  whether or not they do or not to be compliant. Can't some of that get into the browser? Like,
+[2557.20 --> 2563.08]  couldn't we just build it? Like, could, could, could your work extend beyond Shopify's borders
+[2563.08 --> 2564.56]  and help other people too?
+[2565.20 --> 2568.96]  Yeah, absolutely. And this is, I think this is key. This is not just about Shopify. It's about
+[2568.96 --> 2575.22]  improving the buyer experience on the, on the web holistically. So two things to answer that.
+[2575.30 --> 2579.82]  First of all, the remote DOM library that I mentioned, it's an open source project that we've built
+[2579.82 --> 2585.74]  in open source. So if you go to github.com slash Shopify slash remote DOM, you'll find that there,
+[2585.74 --> 2590.02]  uh, take a look at it, use it. This is, that's the technology that powers Shopify checkout.
+[2590.66 --> 2594.68]  Other large companies have already adopted it. I believe Stripe is using it for their apps.
+[2595.04 --> 2600.62]  Actually fun story. Uh, when we built the project, I think Stripe beat us to using it in a production
+[2600.62 --> 2601.82]  product. Really?
+[2601.82 --> 2607.74]  Even though we were the ones developing it for checkout, but like it is, um, it is used
+[2607.74 --> 2615.18]  at Shopify and, uh, by other big, big players to provide this form of isolation. And the general
+[2615.18 --> 2619.74]  pattern is, Hey, I have a trusted first party surface into which I want to bring in third
+[2619.74 --> 2625.66]  party content. And I, I don't want to compromise integrity of my first party top level surface.
+[2625.66 --> 2630.94]  Right. Well, remote DOM is one of the like technical solutions for that. So please take a look at that.
+[2630.94 --> 2635.90]  That's answer. Number one, second though, and coming back to the browser conversation.
+[2636.86 --> 2642.06]  Absolutely. The primitives that we have in browsers today, content security policy and SRI,
+[2642.94 --> 2647.74]  we can make better. And, uh, we've actually done a bunch of work on exactly that at Shopify. Um,
+[2648.78 --> 2654.22]  we don't want to do work in JavaScript that we could push into the browser because the browser is
+[2654.22 --> 2658.14]  just much more efficient and it has capabilities that we otherwise would be very hard for us to
+[2658.14 --> 2664.06]  replicate. So first let's like enumerate some trivial examples of gaps, script integrity.
+[2664.06 --> 2669.66]  So, uh, sub resource integrity, uh, for those not familiar on your script tag, you can pass in
+[2669.66 --> 2674.86]  effectively a hash. So when you inject the tag into your HTML, you can pass in a hash that is a
+[2674.86 --> 2680.70]  fingerprint. And when the browser loads the script before it executes that it can compare the hash of
+[2680.70 --> 2685.34]  the thing that it fetched versus what you've defined and say, Hey, if those two things match,
+[2685.34 --> 2690.86]  great. I will execute the script. Otherwise I'm going to raise a violation and not report this.
+[2691.74 --> 2695.34]  That's a, that's a big capability, uh, in that existing browsers today.
+[2696.14 --> 2701.98]  Uh, it's not simple to deploy, but it is doable, right? Because you need to figure out how do I get
+[2701.98 --> 2706.38]  these hashes and how do I inject them at the right place? But then one of the gaps that existed for a long
+[2706.38 --> 2713.10]  time was, um, module imports. So SRI worked for top level scripts, but if you're building a JavaScript
+[2713.10 --> 2718.62]  application and you're using an import, uh, you just could not pass in an integrity hash. Why?
+[2719.34 --> 2725.26]  Well, because module imports came after sub resource integrity was designed. So it was just never,
+[2725.26 --> 2730.30]  it was never a thing. Um, that was a pain point for us because we use module imports at Shopify. So we
+[2730.30 --> 2737.34]  worked with Chrome and Safari to upstream, uh, some patches to get that supported for module imports.
+[2737.82 --> 2743.74]  So the good news is that's now baked in, I believe as of May of 2024, I think when Safari shipped it
+[2743.74 --> 2749.66]  in the, in their, in their release, uh, both Chrome and Safari support, uh, SRI for module imports. Like,
+[2749.66 --> 2755.82]  so that's, that's one. Um, another thing that, uh, came up in our, uh, thinking when we were exploring
+[2755.82 --> 2763.58]  CSP compliance and how do we make our law, our own life simpler is, uh, this idea of require SRI for.
+[2764.46 --> 2771.50]  So what if you could express a content security policy that says, Hey, all scripts must have an SRI
+[2771.50 --> 2779.34]  or some, or integrity hash. Gotcha. Right. And why is that useful? Well, then you can make, um,
+[2779.98 --> 2785.26]  a strong claim that if you have that policy being enforced by the browser, then if for some reason
+[2785.26 --> 2790.70]  you sneak through by accident or malicious act, a script that doesn't have it, they would just be
+[2790.70 --> 2796.94]  rejected, right? Which today it would just execute normally, um, without any questions.
+[2797.82 --> 2803.98]  And, uh, even though that might be hard to deploy in an enforcement mode, it could totally work and be
+[2803.98 --> 2809.42]  really useful in a report only mode. So for those not familiar with content security policy, you have
+[2809.42 --> 2814.38]  an enforcement mode and a report report only mode where you can get violations, which is incredibly
+[2814.38 --> 2818.94]  useful because you could say, Hey, this is a policy I would like to enforce. What are the violations?
+[2819.98 --> 2825.58]  So with the require SRI for you could deploy this in report only mode and say, great. Now I'm going to
+[2825.58 --> 2834.14]  get reliable reports from the browser, from the wild for any time a browser detects that a script
+[2834.14 --> 2840.86]  is missing an SRI capability, right? Um, this is great because sophisticated attackers
+[2840.86 --> 2847.26]  would not emit these scripts on every single page load. They might target specific users or
+[2847.26 --> 2851.90]  a class of user, or maybe they target specific browser, or maybe if it's an extension, it'll, it'll
+[2851.90 --> 2856.62]  apply some sort of other heuristic, right? It's very hard to, this kind of mirrors our conversation on
+[2856.62 --> 2863.58]  why ROM is important, real user measurement, gathering violation reports from real users gives you a much
+[2863.58 --> 2871.50]  much better and reliable signal for where the problems are. So, uh, require SRI for is another
+[2871.50 --> 2878.70]  capability that we've, uh, shipped into Chrome and that allows you to get violations on missing SRI
+[2878.70 --> 2883.42]  attributes, which allows you to build an inventory of like, this is the list for me to burn down and
+[2883.42 --> 2889.18]  figure out why. Right. And if anything changed, how do I, how should I react to it? Yeah.
+[2889.18 --> 2897.10]  Yeah. Another example is, okay, great. Now we have these reports coming in. Uh, wouldn't it be nice
+[2897.10 --> 2902.38]  if we could also get the hash of the content, right? It's today you would just get a report saying,
+[2902.38 --> 2912.30]  Hey, I detected script from example.com slash X, Y, Z dot JS. But what was the content of that? Uh,
+[2912.30 --> 2917.74]  you don't know, right? Wouldn't it be nice if you could also get a hash in the report such that you could,
+[2917.74 --> 2923.58]  um, audited, um, audited and say, Oh, well maybe that's totally okay because the partner revved their
+[2923.58 --> 2928.14]  version and it just happens to be the V2. I just put that, I put that into my approved list and
+[2928.14 --> 2935.66]  everything's fine versus I have no idea if that was a compromised version or a legitimate version of
+[2935.66 --> 2943.26]  the script. Interesting. So pardon my ignorance for a moment, but where does the reporting take place or
+[2943.26 --> 2948.86]  post to the browsers doing the reporting? Is it correct? Who gets the report and how is it?
+[2948.86 --> 2954.86]  The browser sends it off somewhere or? Yep. So on the wire, um, you would, when you emit a page,
+[2954.86 --> 2961.42]  uh, you can define a content security policy or CSP policy in a header and you would define, you know,
+[2961.42 --> 2966.70]  for script source list, for example, a list of origins from which you're allowed to fetch,
+[2966.70 --> 2974.94]  um, for images and all the rest. You also have a report to target and a separate report to header
+[2974.94 --> 2980.94]  that provides a specification for you specify the end point to where you want the report violation
+[2980.94 --> 2987.10]  report to be reported. And you know, as good hygiene, that reporting endpoint should ideally be like a
+[2987.10 --> 2994.70]  distinct origin, um, in all the rest, but you provide a destination. So you can find services that will do this
+[2994.70 --> 2999.90]  for you, right? They'll just say, point your report to, to us, and we will provide a dashboard,
+[2999.90 --> 3002.94]  which you can drill down reports. We will aggregate, we'll give you metrics,
+[3003.50 --> 3007.74]  and all the rest. That's something that we do in house at Shopify. And I think many other
+[3007.74 --> 3013.34]  large providers will do on their own. Uh, but you could outsource that problem, but just having the
+[3013.34 --> 3020.70]  ability to even get the report with, Hey, um, a report has been emitted because the script is missing
+[3020.70 --> 3027.98]  an integrity hash is by itself a really useful capability because otherwise you'd probably
+[3027.98 --> 3032.06]  have to set up some sort of a crawling infrastructure that periodically checks your page and says,
+[3032.78 --> 3036.78]  you know, I, I accessed this page from five different points on the globe.
+[3036.78 --> 3037.90]  Right.
+[3037.90 --> 3043.26]  Every 24 hours. And I observed that nothing has changed. Well, that's good, but we could do much
+[3043.26 --> 3048.14]  better by just actually observing what the real users are seeing and getting the actual reports of
+[3048.14 --> 3048.62]  violations.
+[3048.62 --> 3058.14]  Gotcha. So this new one require SRI for would work in like manner as the CSP violations in terms of
+[3058.14 --> 3058.62]  reporting.
+[3058.62 --> 3064.62]  Right. So you would, you would, the, the CSP policy is require SRI for scripts, right?
+[3064.62 --> 3070.70]  So you're saying all of my script resources must have a hash. And then you can configure that to
+[3070.70 --> 3077.58]  be a report only policy such that it would still execute if the script is missing the hash, but you
+[3077.58 --> 3083.26]  would get the violation fired in the background. And the browser has its own logic for prioritizing
+[3083.26 --> 3087.10]  batching delivery and doing all of that, uh, to, to get you the report.
+[3087.10 --> 3089.90]  Mm-hmm. Now do you deploy this one in Shopify?
+[3089.90 --> 3091.26]  Yep.
+[3091.26 --> 3091.82]  Yep.
+[3091.82 --> 3095.34]  And do you use it in report mode or do you let lockdown mode or how do you use it?
+[3096.62 --> 3105.26]  So for this one, it would be a report report mode. Uh, but it depends on how, like, it depends on the
+[3105.26 --> 3109.58]  shape of your checkout and how much control you have for your first party or third party content.
+[3110.22 --> 3115.98]  Just to double back on that for, for Shopify, we, uh, for our checkout, we enforce a CSP policy.
+[3115.98 --> 3123.90]  Actually, let, let me run through the whole list. Uh, for our first party content, we have a well-defined
+[3123.90 --> 3130.14]  process for vetting all the dependencies and a process for updates, auditing, to make sure that
+[3130.14 --> 3135.02]  we provide some guarantees over, you know, if the library that we depend on has been compromised,
+[3135.02 --> 3141.02]  how can we detect that? We have change management process for it. So this is the reviews, testing,
+[3141.02 --> 3146.54]  CI, all the things that you would expect that allows us to create the inventory. We know from where it's
+[3146.54 --> 3151.50]  served, which means that we can define a strict CSP policy that says you should only fetch from these
+[3151.50 --> 3152.78]  sub origins that we trust.
+[3152.78 --> 3159.42]  In our build step, we can inject the hashes, the SRI hashes for known content.
+[3160.54 --> 3167.98]  And we can also emit the require SRI for policy to ensure that if anything else, for some reason,
+[3167.98 --> 3173.42]  if we emit a missed some script that we would get a violation on that because we don't want to break
+[3173.42 --> 3179.10]  checkout, but we want to be notified immediately. If those things are detected, then we can react to it.
+[3179.10 --> 3185.18]  And we have our own reporting endpoint, which we aggregate, we look at the reports. This is a
+[3185.18 --> 3188.46]  thing that merchants don't have to worry about because we do this work on their behalf.
+[3189.10 --> 3195.90]  Right. And we can provide this guarantee over overall integrity. And then finally, we've protected
+[3195.90 --> 3202.94]  the parent page, but the payment credentials page or the payment form itself is also isolated into its
+[3202.94 --> 3208.22]  own iframe, just as it was before. So this is a defense in depth, right? We protected the parent,
+[3208.22 --> 3213.90]  but we also have our own implementation of the iframe and like the full PCI compliance
+[3213.90 --> 3215.18]  behind that particular form.
+[3215.18 --> 3221.26]  Well, that's a lot of stuff for PCI compliance, Ilya. What happens with V5? How many years are you
+[3221.26 --> 3222.14]  going to put into that one?
+[3223.98 --> 3229.74]  I don't know. That's a great question. I'm pretty sure that V4 will keep us busy for a long while.
+[3231.18 --> 3235.10]  Yeah. Because this is only section 6.4.3, right? That's all we're talking about right here.
+[3235.10 --> 3238.70]  That's right. There's all the others.
+[3240.78 --> 3245.34]  Okay. So interesting stuff. It sounds like you've solved some really
+[3246.94 --> 3253.18]  difficult technical challenges in order to do this in a way that's not just compliant, but actually
+[3253.18 --> 3257.42]  in the spirit of the compliance as well, like trying to actually make it more secure.
+[3257.42 --> 3264.14]  What are some takeaways for listeners out there? Maybe they're doing their own checkout. Maybe they
+[3265.02 --> 3270.78]  have compliance they need to do. Maybe they just want some more secure websites. What do you think
+[3270.78 --> 3277.02]  they could be thinking walking away from this if they're not in the actual situation that Shopify's in
+[3277.02 --> 3280.62]  and having to implement this stuff? What could they learn from this conversation?
+[3281.74 --> 3289.58]  Yeah. I think the meta pattern and message takeaway here is broadly the integrity and security of
+[3289.58 --> 3295.34]  first party versus third party content. We mix first party and third party in most contexts, but
+[3295.34 --> 3302.14]  even outside of checkout, there are many surfaces. Let's say you have an admin surface or
+[3304.62 --> 3311.42]  a privileged surface that you only want certain users to access and you want some sort of extensibility
+[3311.42 --> 3317.50]  in there. So you want to bring in third party content or customization in all the rest. The pattern that
+[3317.50 --> 3323.82]  we're describing with isolating third party content is a generic pattern that you can deploy there.
+[3323.82 --> 3329.34]  So we use the same sandbox and technology in checkout. We use the same technology in our admin.
+[3329.34 --> 3333.42]  So for merchants, we allow customizations and third party developers to bring in their custom
+[3334.06 --> 3339.98]  UI and other aspects. As you can imagine, that's a very sensitive surface. Order data is there,
+[3339.98 --> 3345.74]  customer data is there. You don't want to just open up a Pandora's box of injector arbitrary
+[3345.74 --> 3354.30]  JavaScript because who knows where the data might travel. So the isolation primitive, it may be a
+[3354.30 --> 3363.74]  remote DOM, it may be something else, but this way of thinking of isolating into either an iframe or a
+[3363.74 --> 3368.14]  worker, I think is a pattern that we should be adopting more widely. And it has these additional
+[3368.14 --> 3373.34]  benefits. You have better assurances about security, yes, performance as well, because you're isolating
+[3373.34 --> 3380.38]  content and moving it off the main thread. You get to define the API contract. So you have better
+[3381.82 --> 3386.62]  upgrade ability if you need to maintain that. And I think that's just something that we need to get
+[3386.62 --> 3392.14]  better at on the web. Now, the challenge, I think for all of us and kind of as industry practitioners
+[3392.14 --> 3397.18]  is to think through, boy, the worker is kind of this like naked environment.
+[3397.18 --> 3402.86]  Mm hmm. We can probably figure out, we should think about how do we figure out some better set
+[3402.86 --> 3407.34]  of APIs that where we don't have to reinvent the entire wheel, just as we did with, you know,
+[3407.34 --> 3414.54]  at Shopify for great. Now I want to build a heat map thing. What does that mean? How do I mirror the
+[3414.54 --> 3422.14]  entire stream of events from top level page into this isolated environment? I think we can do some
+[3422.14 --> 3427.82]  thinking and innovation there. Very cool. Anything else that's on your mind that we haven't discussed
+[3427.82 --> 3432.06]  in this context or honestly, in any developer context, I always love to hear your opinions
+[3432.06 --> 3437.50]  on stuff. Anything else on your mind? I think one really interesting topic coming, coming back to
+[3437.50 --> 3443.18]  the world of checkout and commerce is of course, agents and how agents will interact or how they might
+[3443.18 --> 3450.54]  affect any of these behaviors. Oh, yes. MCP. Are you done with MCP? That's the newest acronym,
+[3450.54 --> 3456.70]  model context protocol. It's burgeoning. Yep. Yep. MCP is definitely top of mind and we're
+[3456.70 --> 3461.90]  looking at it intently. We're using it for a number of tools and internal services at Shopify.
+[3461.90 --> 3467.34]  We're also considering if and how we should be exposing MCP as a, as a protocol and endpoint,
+[3467.98 --> 3473.98]  as a service on behalf of merchants. So, so imagine you could have a merchant storefront as a remote
+[3473.98 --> 3480.62]  MCP endpoint. But more broadly, like if you think of, let's imagine you interacting with an agent
+[3480.62 --> 3488.30]  asking it, Hey, I'd like to have a pair of white sneakers, size 10, 15 to or 50 to a hundred dollar
+[3488.30 --> 3493.74]  range. Please go find me a pair and check out under the hood. The agent might crawl the web,
+[3493.74 --> 3498.46]  find the storefront, add to cart, head to checkout. And what does it do then as it's looking at a
+[3498.46 --> 3503.50]  payment form? Is it a responsibility of the agent to hold onto your payment credentials? And what are
+[3503.50 --> 3509.74]  the implications of that for entering? How does it enter those credentials? Are there any security
+[3509.74 --> 3515.66]  and compliance problems or challenges in that? I think that's a wide open question that we as an
+[3515.66 --> 3523.74]  industry are yet to figure out an answer. Is the human required in that loop? What if it's an
+[3523.74 --> 3527.58]  accelerated checkout where maybe information is vaulted? Right. I think there's a range of
+[3528.38 --> 3532.14]  questions and answers that we need to figure out in the space.
+[3533.66 --> 3537.98]  What's your personal thought on is the human required in the loop? How do you feel
+[3537.98 --> 3540.94]  confidence wise on removing the human from that loop?
+[3540.94 --> 3546.30]  I think it's context dependent. I think there's definitely a class of commerce and types and
+[3546.30 --> 3552.38]  certain types of transactions where I know exactly what I want. There's very low risk and it's kind of
+[3552.38 --> 3559.50]  a predefined flow where I just say, look, I need another carton of milk. You know exactly what
+[3559.50 --> 3563.02]  I'm looking for. You know where to shop and please go finish it. And I just want it to my front door.
+[3563.98 --> 3569.90]  And then there is other types of experiences where maybe this is your first time engaging with a
+[3569.90 --> 3577.18]  merchant. Maybe a merchant has a set of rules where they actually require you or require the agent to
+[3577.18 --> 3584.54]  decelerate because, hey, for compliance reasons, I may need to verify your age or I need you to read
+[3584.54 --> 3590.54]  this disclaimer on this product before you purchase it. Right. You can't just have the agent blindly
+[3590.54 --> 3597.90]  ignore that context or click approve and then proceed with the transaction. So I think we're,
+[3597.90 --> 3604.22]  we'll need to define some protocol or a shared mechanism to signal to agents that like, hey,
+[3604.22 --> 3612.78]  in this particular case, I need you to pause and ask for human to either confirm or take over control
+[3612.78 --> 3617.26]  and complete the transaction. There's so many questions there. I just don't feel like
+[3618.14 --> 3623.98]  I even have the brain right now to analyze all the things that have to be considered. I'm glad that
+[3623.98 --> 3629.18]  you're, are you going to be working on this for Shopify? Are you going to stay in the NPC island? What,
+[3629.18 --> 3633.50]  what's next for you inside? Is this an active thing that you're thinking about for Shopify?
+[3633.50 --> 3639.50]  It is definitely, um, definitely an active area of, uh, of exploration for us. That is one of the
+[3639.50 --> 3644.06]  things, um, I'm looking with, um, our team and many of our partners who are building these agents.
+[3644.06 --> 3649.34]  We're trying to figure out what is, what is the future of checkout where agents drive some meaningful
+[3649.34 --> 3654.06]  portion of, of that experience? What does a good experience even look like, um, in that context?
+[3654.06 --> 3659.18]  So I think those are all, um, very interesting and pertinent question given where we are today.
+[3659.18 --> 3665.82]  Hmm. Well, I have to have you come back in a year or two and let us know what you end up building as
+[3665.82 --> 3670.86]  you've, you figured it all out. You seem to have figured it out. At least this, this hairy technical
+[3670.86 --> 3674.54]  problem has come to this new PCI stuff. So I'm sure you'll figure out something.
+[3674.54 --> 3679.66]  Yeah. We'd love to be back. And at the rate that we're moving in AI world in a year or two for now,
+[3679.66 --> 3682.06]  who knows what will be there. So.
+[3682.06 --> 3690.70]  Yes. I'm trying to think of the most recent person who said six to nine months and LLM will be writing
+[3691.66 --> 3695.50]  a hundred percent of code. So, I mean, who knows, man, maybe we'll be,
+[3695.50 --> 3697.98]  you and I will be out on the street corner talking about this stuff.
+[3697.98 --> 3701.18]  I doubt that is the case, but well,
+[3701.18 --> 3702.38]  who knows?
+[3702.38 --> 3708.86]  Yeah, me too. But you know, it's not a week goes by that somebody doesn't declare software
+[3708.86 --> 3713.02]  engineering is dead or dying. So I had to squeeze that one in there.
+[3713.02 --> 3716.70]  Yes. I think, I think what we're actually saying is the definition of what software engineering is,
+[3716.70 --> 3723.18]  is changing, right? I am constantly amazed by what, um, AI is capable of doing in terms of
+[3723.18 --> 3729.18]  software development, but I'm also constantly surprised by the silly and stupid mistakes that it makes.
+[3729.18 --> 3731.74]  And oftentimes those mistakes are actually due to
+[3731.74 --> 3739.02]  misunderstanding or lack of poor definition of the problem that is being solved. It's kind of
+[3739.02 --> 3742.70]  putting the mirror back to yourself, right? Because oftentimes I'll find that like, actually,
+[3742.70 --> 3746.70]  you know what, you did exactly the right thing, the way I expressed it, but that's not what I meant.
+[3746.70 --> 3750.14]  And I didn't even know what I meant when I typed it, because now that I've seen the mistake,
+[3750.14 --> 3755.10]  I understand what I was actually trying to get to. So it is this like art of defining the problem
+[3755.10 --> 3761.02]  and rubber duck programming. And I think we're heading more and more towards the world where
+[3761.02 --> 3766.38]  we're actively collaborating instead of hands on keyboard typing, uh, if statements.
+[3766.38 --> 3771.58]  Yeah. The best rubber duck programmers might be the best programmers of the future. It's the
+[3771.58 --> 3775.34]  ones who can just talk it out the best, you know, figure it out as you go. All right,
+[3775.34 --> 3780.06]  Elia, appreciate you coming on the show and chatting with us and looking forward to having you back soon.
+[3780.06 --> 3781.26]  Thank you, Jared.
+[3784.78 --> 3790.22]  Okay. So it turns out securing e-commerce checkouts has never been more complicated,
+[3790.22 --> 3796.14]  but thankfully brilliant engineers like Elia and his team at Shopify are putting in the work. And
+[3796.14 --> 3801.66]  some of that work is making its way back into the web platform. I love when that happens. And when you
+[3801.66 --> 3807.58]  think about it, the complicated nature of it all makes sense. The stakes have never been higher. I read the
+[3807.58 --> 3815.18]  other day that last year e-commerce sales soared to a record 1.2 trillion dollars. That's a lot of
+[3815.18 --> 3821.58]  moolah being transferred. And if you can hack it, well, you can jack it. So yeah, it's complicated for
+[3821.58 --> 3828.46]  a good reason. Let's give one more thanks to our sponsors of this episode, Retool, Augment Code,
+[3828.46 --> 3834.46]  and of course, Fly.io. Check out their wares to support their work, which supports our work,
+[3834.46 --> 3841.10]  which we appreciate. Thanks also to our beatmaster in residence, Breakmaster Cylinder. Did you know
+[3841.10 --> 3846.38]  our next full-length album is almost ready? And I'll tell you right now, it's called After Party.
+[3846.38 --> 3851.10]  And I'll also tell you right now that I've been bumping it all week. I dig it. Hopefully you will
+[3851.10 --> 3857.34]  too. Soon. So soon. All right, that's all from me, but we'll talk to you again on Changelog and Friends
+[3857.34 --> 3859.26]  on Friday. Bye, y'all.
+[3887.34 --> 3906.70]  Bye, y'all.

Spec-driven development with Kiro (Interview)_transcript.txt

@@ -0,0 +1,445 @@
+**Jerod Santo:** Today we are joined by Deepak Singh, one of the leaders of the developer agents team at AWS working on Kiro, a very exciting and interesting new take on a tool that a lot of people have takes on right now. Deepak, welcome to the show.
+
+**Deepak Singh:** Thanks for having me.
+
+**Jerod Santo:** Y'all announced Kiro, I think it was back in July, early July, to much fanfare and excitement. I even got excited, and that's hard for me these days, to get a little bit excited about a tool... So I was happy when y'all reached out and said "Let's talk about it on the pod." Welcome to have you, and why, when there are so many agentic coding things going on, is AWS throwing your hat into the ring?
+
+**Deepak Singh:** Yeah. I mean, we threw our hat into the ring a little bit before Kiro, and actually we learned a lot doing that. If you go back in AI world, like donkeys years, like less than two and a half, three years ago, it feels like another era/epoch in AI time. You had all these assistants that were basically fast typists, they helped you complete sentences... And then you went into these chat-based assistants that you could write functions with, like "Hey, give me a function that does blah..." And it was pretty good, but we looked around inside the company, amongst our customers, and one of the nice things about being at Amazon is we have a lot of software developers in the company, who are very opinionated about how they use software. And so talking to them and talking to our customers, it became pretty clear that this sort of chat-based auto-completion/'give me a function' was useful, but it wasn't going to change the world. And along the way, LLMs got better, a lot better. They added thinking and reasoning to it, and that's when we started building sort of what we call our next generation of agents. And we started launching them; actually, earlier this year, we launched something called the Q CLI. This is actually the most heavily used agent inside Amazon.
+
+\[04:33\] In the process of building things like the Q CLI, learning from our customers, a few things became very clear to us. The first one was we were moving towards a world where code was not going to be typed by humans. So the type faster, write blocks of code for me was not going to be the game changer. The game changer was going to be I have a pull request converted into something. So I've written an issue, converted it into something that's meaningful that I can make part of a bigger project, or write my entire software for me, and a human was not going to do the typing. We had already started seeing examples of that, and it was pretty clear that that's where the world was starting to go.
+
+The second thing we did was we found that senior engineers had a really, really hard time with the original sort of era of assistance, where invariably they would say "I understand the code base. I understand the problem. I can type faster. Yes, I might use it for some help", like I use the - I'm dating myself - Perl Cookbook to find a function. And we were like, "Okay, that's not useful. How do we get our best engineers wanting to use these tools and actually finding them useful in their day-to-day work?" As we started talking to them, the thing that became very clear was the way they wrote code or solved a problem was by basically breaking it down into smaller problems. Like how do you take a larger problem, what are its constituent elements, and how do they express them? Sometimes they write them on a whiteboard. Sometimes they write a design document. Sometimes they write pseudocode, and then they work with a set of junior engineers and other engineers to convert that into a final product. This was happening before AI. So the question we asked ourselves was "Can we convert that into an AI-based system? Can we take the way our best engineers think, make it easier for them to do that, and along the way also assume that they're actually not going to type the code, and the agent is going to do it for them?"
+
+Along the way, as we were doing this, this whole idea of vibe coding, as we call it, came up, which is this prompt-based loop where you keep prompting an agent until you get to an answer... And in that world also, we noticed there were people who were for the more complex problems going into something else, maybe Claude, or ChatGPT, and creating a set of tasks and breaking it up and then throwing it into the coding agent. The whole premise of Kiro was that's just the way it works. The code typing part is hidden. It's there, you can type code if you wanted to, but the assumption is the only time you'll type code is to tweak things that an agent may have gotten wrong... But even changes that you want to make will be done by talking to an agent.
+
+So the way Kiro works is like our senior engineers work - you express a problem, Kiro will convert that into a specification of specs, which is sort of the core piece of Kiro. It's a spec authoring tool at its heart. And a spec is a set of requirements in Markdown, design document, also in Markdown, and a set of tasks where it wants to do also written in Markdown. And that's what you work on with the agent, and the code part of it is just the output that the agent generates. There's more to it, but that was the starting point. That's how we came to the user experience that Kiro has today.
+
+**Adam Stacoviak:** It's interesting to go that deep. I mean, even Q CLI - if I got the pullback right, Q CLI enables completions for hundreds of popular CLIs... Is that what you mentioned initially?
+
+**Deepak Singh:** No, that's how it started.
+
+**Adam Stacoviak:** \[08:00\] Gotcha.
+
+**Deepak Singh:** Yeah, so there was an open source project called Fig that became part of AWS about two and a half years ago. I used that as part OhMyZsh on my shell even before they became part of AWS... And they do a lot of CLI auto-completion. In March of this year, the Q CLI team put an agent inside the terminal, inside the CLI.
+
+**Adam Stacoviak:** I see.
+
+**Deepak Singh:** And you can basically start vibe coding. That's what it is. You start prompting it... It's advanced quite a bit now. For example, you can start a conversation, and if you want to experiment with something, you can actually create a branch of your conversation. You can go into a tangent; quite literally, it's called tangent. Have a conversation, and if you don't like that part, you can come back to your checkpoint, or you can now go down the different branch, and make that main later, effectively.
+
+So that's been very successful, but it's all an interactive conversation, right? So the problem is what happens on day 20, when you've forgotten what you did, you have to go look at the conversation history... And what happens when you give that code to somebody else? They have no context of what was in your head, and no artifact that they can look at in a team. And we work in teams. A lot of engineers work in teams. So in many ways, what we were trying to do with the Kiro spec UX was "How do we capture that vibe coding magic, while making the code more robust over a period of time, where it's still useful a year later, and people have the ability to inspect the specification, modify it if they want to make any changes?" The two teams work pretty closely together, so we learn a lot from each other, so it's kind of fun.
+
+**Jerod Santo:** I think perhaps why this idea, at least, if not the implementation, resonated with me and I think with a lot of people - well, first of all, it's AWS, so you guys have a loud megaphone, so when you announce something, people pay attention. And so that got some of the attention. But also, it's a new take on something that we're all working on, which is like this agentic coding thing... And it's kind of jumping ahead, I think, in the process we're all kind of getting to maybe in our own ways, of like realizing that chat isn't it; and chat's cool and all, but you just can't build serious software in a chat dialogue. Like, you can do stuff, and you can make progress, and you can experiment, but there's a certain point where you're like "Yeah, I don't want to live my entire life this way."
+
+**Adam Stacoviak:** Way too cumbersome, yeah.
+
+**Jerod Santo:** Yeah, I'm just not going to be productive. And you guys formalizing --
+
+**Deepak Singh:** Fully agree.
+
+**Jerod Santo:** Yeah, formalizing the idea of - I think you're calling it spec-driven development, or you can give me other taglines, but like this concept that is...
+
+**Deepak Singh:** Yeah, that's the right tagline.
+
+**Jerod Santo:** Okay, perfect. Like, even just that to me, I was like, "Yeah, that makes some sense. That's something worth trying." And Adam can tell you about agent flow, I'm sure he will. He's kind of developing his own spec-driven development style inside of Claude Code, or with Claude Code, as he builds things...
+
+**Adam Stacoviak:** It's agent-agnostic.
+
+**Jerod Santo:** And Kiro seems like an attempt to -- well, you're using it with Claude Code, right?
+
+**Adam Stacoviak:** Well, AMP and Augment...
+
+**Jerod Santo:** Okay, so lot of things.
+
+**Adam Stacoviak:** Yeah.
+
+**Jerod Santo:** Okay, well said. Regardless of the agent, he's doing that... And I think that what Kiro attempts to do, I guess, is to formalize that into the entire experience. Is that fair?
+
+**Deepak Singh:** Yeah, that is fair. And it'll evolve. I think we are learning how to build software in a world where the engineer's role, the role of an engineer is going from being the -- I'll say typist; I don't mean it quite like that. But they are the ones who are the fingers on the keyboard, converting what they have in their head into a language, which is usually - pick your language of choice here, inside Amazon, AWS. If you're building a backend system, it's probably going to be Rust. And that's what they excel at.
+
+Now, you're acting a little bit differently. You're a driver, you're driving the behavior of an agent. It's about how you give the agent the right context, structured thinking; specs is one way.
+
+\[12:01\] There are other components to it, which others -- I'm suspecting your agent flow thing that also comes up, which is how do you provide access to the right tools? How do you provide access to the right steering files, that give the agent the right context to be more effective? And this combination of -- actually, inside a Kiro project, the core components are the spec, the tools, which is usually MCP servers, the steering files, which are essentially, again, markdown files, JSON files that tell an agent 'This is what I want you to do, how I want you to behave", give it a persona, as it were... And downstream of that, you have this idea of hooks, which are these automations that happen -- like the moment you check in code, it's watching for events, and it could be "The moment you save a file, please run an optimizer for me, because there may be a better way of optimizing the code, or send somebody an email saying "I'm done checking in code. You may want to look at your merge request" or whatever. That combination is the Kiro UX, but the spec workflow is the heart of that.
+
+**Jerod Santo:** So what does Kiro look like? How does it work? What is it today? Where is it heading, etc? But first, lay the field out, so we can all see what it looks like.
+
+**Deepak Singh:** For those of you who haven't seen Kiro, it lives on its own website called kiro.dev. The best place to find anything about Kiro is go there. You can download it, you can read the documentation, but my recommendation is get it -- right now we have a wait list. So one of the -- we talked about the Kiro launch... It got really popular really fast. We got over a hundred thousand developers signing up in the first few days, and we had to put a wait list on, which - we mostly cleared the original wait list, which got very, very large, but it's still there... But the assumption is if you sign up for the wait list now, you'll be out of the wait list very quickly. You don't have to wait for a month or so.
+
+But once you download it, you'll recognize a few things. One, I think the most obvious one is that it is a Code OSS-based editor. So if you have VS Code profiles and preferences already set up, it'll import them for you. So if you had any plugins, et cetera, running, most of them, it'll run pretty successfully. Some there are clashes and we'll warn you that "Please do not use this plugin with Kiro", for whatever reason. And if you have -- we use what is called VSX exchange, which is our open source plugin ecosystem for VS Code. So that's where you can go and get most of your plugins. That's the obvious thing.
+
+The second thing that'll jump out at you is when you open a project, instead of getting your code windows etc. the default is you'll get a question. "Do you want to just talk to an agent, or do you want to start building code using spec workflows?" And you can actually just start talking and Kiro will detect whether you need to do a more structured development, or just -- sometimes vibing is fun. Sometimes you just have questions to ask the agent. But you'll notice very quickly that the emphasis in Kiro is not in the code editor part, the code editing part, the syntax highlighting of code. That stuff is all hidden. The main window that's open, and the way you can set it up is all around the spec authoring experience, where you're creating a specification and you're working with an agent to create the spec.
+
+So that's the first thing you'll notice about Kiro. And you can use it on your favorite laptop, all the operating systems that you care about... And it's pretty easy to get going with, because once your brain adjusts to the fact that what you have in front of you is not code, but actually, as I like to joke, Kiro is more of a markdown editor than a code editor, because most of your interactions are going to be natural language with a markdown, and what you see is markdown.
+
+The first thing that will come out is a set of requirements that are written in EARS format. So those of you who have worked with user requirements etc. before. EARS is a format that I think came out of Atlassian, to represent user stories and requirements, all of them in markdown. You can throw text at it, you can throw sudocode at it... It's multi-modal. You can throw your classic "I drew my startup idea on a napkin" and throw it at it, and Kiro will convert that into a specification.
+
+\[16:24\] Once you like the requirements - you can take a look at them and say "Yeah, I like them." If you don't like them, you can type into the chat window that you want to change them, or you can directly edit the markdown, which we rarely see people do.
+
+The next thing it does is actually come up with a design. Like an engineer, it'll come up with "Here's my design proposal", and you get to decide if you like it. And if you're okay with it, then you say "Okay, I'm good." And it'll create a set of tasks; it could have one task, 20 tasks... Where a task could be "I am going to create a window and a button on the right corner, which is blue in color, and says Subscribe." That could be one of the tasks. It'll be written in. And within that, there'll be sub-tasks, which are the actual things it'll do.
+
+You can do task by task, you can -- because some of those tasks are optional. Or you can say "Implement it all." It'll go ahead and do it. If you go into the file browser, you'll see all the files getting created. You can inspect them. Sometimes it'll come back to you for help, even if you're not in pilot mode, where it needs a user permission because your file system requires it, for example.
+
+And that's how your projects are created. That workflow is around the specification. And if you want to go back and change the code, the most recommended way is actually to change the specification, and not go in and manually edit the code. It's like, "No, I actually don't want this to be blue. I want this to be green." Or "I didn't like what you did here. Change your implementation a little bit." You can also do it in real time. If you see it's going in the wrong direction, because you get the chain of thought in front of you, you can essentially hit Control+C and go "Nope, I want you to go in a different direction." So that's kind of how it works. I'm simplifying greatly, but that's kind of how it works.
+
+**Break:** \[18:10\]
+
+**Adam Stacoviak:** You know, it sounds super weird to say this, but I resonate with not editing the code yourself... And the reason why is because the LLM, the agent has this context that is beyond my context. And it kind of has this plan. And so when I start telling it what to do, I kind of derail its plan. And that to me is jarring at first, because I'm like "No, no, no. I still want to be in charge here, I want to have agency." Not to be too punny, but... I find that even a small change I really don't want to do, because I want to tell it what to do. Jerod mentioned agent flow - this is just a term I coined. I think it sounds cool, for one. I think it's the coolest term out there, so...
+
+**Deepak Singh:** Yeah, it is cool I think.
+
+**Adam Stacoviak:** So if everybody wants to use agent flow, back off. I'm just kidding. I like it. It's cool. But my flow is not as sophisticated as yours is in terms of like MCP servers and tools yet... It's really just how I think in my brain about the problem I'm trying to solve, develop some version of a spec... I've been calling them just documents. So I've been calling it document-driven development.
+
+**Deepak Singh:** Yeah. They are documents.
+
+**Adam Stacoviak:** They are documents. And it's nothing against the word spec, but generally, in our industry, specs have been a version of a pejorative that it's just kind of like "Ewgh... Do you want me to do that on specs?" I just personally didn't jive with that term. Cool with you all using it, it's just my personal preference.
+
+**Deepak Singh:** Yeah.
+
+**Adam Stacoviak:** But it was not as kind of sophisticated as yours. That being said, though, as I'm playing with Kiro... You know, you're generating this initial document, the spec, which I get, and I've written those specs before... I was in an agile team, I was a product manager, led teams, and I would speak in these terms, like "As a user, I want X" etc. That whole format. And I get that. And then the design document that comes from that is even cooler. That's where I think my brain gravitates most. So that's kind of what I've been generating, that version of a design document, not so much that initial spec that creates this design document... But not wanting to touch the code I thought was really strange at first, until I started to get into the agent's way with what I told it to do via this document-driven, spec-driven agent flow. I just became the architect of the idea, and the researcher, and the viber... I like to call it riffing, really, because I would ask questions about different things, poke holes in the CLI or the API, or "How does this truly work? Did you research that? Is that really based on facts and truth? Here's the truth here", or "Here's an example of how I've hit a pitfall there, so avoid that."
+
+And then we come up with this thing called a spec, or a document, or whatever, and I say, "Okay, great. This thing represents as best as I can think of the idea." And then usually, the last thing I do is I say "Where are the blind spots at with this?" Okay, we've defined this, I think I've got good eyeballs on the spec or the design... Where are the blind spots here? And we'll go back and forth on what those are, and we'll update that thing. And then when you get into that flow, and you say, "Okay, go." And I want to be autopilot, right? Your button says Autopilot, and I kind of like that, too. It's like auto-edit - sure, go ahead; just make the world. I don't want to really get in that thing's way, you know?
+
+**Deepak Singh:** I wanted to call it YOLO Mode, but \[unintelligible 00:22:56.11\]
+
+**Adam Stacoviak:** Yeah, YOLO mode actually -- I mean, I would almost advocate for that, because that kind of makes the tool fun. What we don't want to take away from this process of creating software is removing the fun of it, and making this tedious task of just like writing specs and writing documents. We can't lose that fun aspect of it. So I rambled a bit, but I totally vibe with the idea of not touching the code, not because I can't change one single value, but because I kind of get in its way, and it has this context window... And I really don't want to disturb it, because I've done my job, which is be the architect and the visionary of the idea, and then we've blessed this thing called a document or this spec... And let's just go. And if I get in its way -- it can get back in motion again, so it's not like a total derail, but sometimes you can really jack it up, and you're on a tangent, or I call it a side quest.
+
+**Deepak Singh:** \[23:56\] Yeah. No, so I think you're completely right. There's two reasons for this also. There's a technical reason, which is context is king. Your results are only going to be as good as your context. And especially if you've been in a long conversation or a long session, going in and manually changing code actually kills the context, and you get out of sync. And getting back into sync is just really hard. I mean, even if you're not using structured work, but just documents, giving your instructions, the moment you go start touching the code, you've lost that context completely. So that's one part of it.
+
+I also feel -- at Amazon, we have this leadership tenet for our principal engineers called "Illuminate and clarify." Our senior engineers are very good at illuminating and clarifying a problem. They usually do it with their team or group of junior engineers. I think of what you just described as illuminating and clarifying to an agent, to an AI agent. And I think the metaphor stands. And the better we are at doing that -- because I think it's more art than science right now. A lot of folks are learning how to work with these agents. You've developed your own workflow. I like the name, agent flow. I've seen other engineers have their own flow, so to speak, and I think the thing that we, as the Kiro team, will have to do over time is - today there's a particular flow that Kiro uses. You can do some stuff around it, but you can't go randomly and completely randomly build your own flow.
+
+I think we learn quickly from how people want to build these, because we want to help developers stay in the flow that they like. That's a big part of what we want to accomplish, is keep it fun.
+
+**Jerod Santo:** Right. So there's a couple decisions that you guys made - you probably had to make it pretty early in the process - which I think are interesting to me, the why's behind the decisions. And one of those decisions was it's going to be an editor. It's going to be a VS Code -- is it a fork, I assume? Or is it a...
+
+**Deepak Singh:** It's based on Code OSS, which is the open source framing project underneath VS Code.
+
+**Jerod Santo:** Gotcha. So it's not exactly a fork of VS Code, but it's based on the same foundation as VS Code. It's VS Code-esque. Not really interested in that decision... I think once you decide "We're going to build an editor", then that's a really good decision to just be like "Let's build off of that." But a lot of people are going to the terminal. And I have found, as an engineer, that I've enjoyed using terminal-based agentic tools more than I've enjoyed using the exact same tools in my editor, whether it is VS Code, whether it is Zed... Those are the two that I've been using things in.
+
+And so I'm curious, first of all, about the choice of "Where does the agent live?" And your guys' answer is "It lives inside your editor." Can you speak to that?
+
+**Deepak Singh:** Yeah. So actually, it's interesting that both the agent living inside Kiro and -- one of the nice things is we have the terminal-based agent as well with the Q CLI. When we say 80% of builders inside Amazon are developing code with AI tools on a very regular basis, a very large number of them are using Q CLI. There's a couple of -- because as I said, it predates Kiro. There's a couple of reasons we did that. There's a class of people who just love working with what you would call an IDE. And I think the IDE as we know it, which as a code editor is going to go away... Because as we just discussed, that's the less important part. It's more of a visualization system. So if you look at people doing terminal editors, they will use an IDE or some text editor at some point because they need to visualize, they need to do diffs etc.
+
+\[27:41\] There are some things that we've found even with Q CLI, which are very, very hard. So if you want to build a spec and have a spec workflow to it, that's a much richer interface. And it's very difficult to put that interface into a terminal. We think there will be a lot of folks, especially power users, who will use terminal-like systems, because they give you that freedom and flexibility. It's the classic Vi versus Emacs war at some level. It's just the flow that they like. But if you're interacting with an agent, it's not a bad system to use.
+
+We also felt that there was a very large number of people who wanted that richer user interface, and the richer user interface allowed us to try a few things. One was the whole spec workflow thing, that we can try building it inside a terminal; it's going to look clunky and ugly, and it won't work. There was one, and we had strong belief in that. But the whole helping people author an MCP server, helping them author a set of steering files... The whole library of hooks that Kiro has is much, much easier.
+
+The way you would do hooks in a terminal was you would probably have a /hooks button, and you would get a list of the hooks that you wanted. But if you had 500 hooks, you'd have to scroll through all of them to find them. In a visual interface, that's much easier. You can have a library of hooks that only your team cares about, and over time you may have a library of hooks that, in fact, your company may publish. It's like, "This is the code review agent that we're going to use. Here's an MCP server that goes with it. Here's a steering file that goes with it." All that context is shared.
+
+So there's things you can do from a user experience perspective that are just hard in a terminal. If all you're doing is just driving an agent with prompts, terminals are great. I don't think the IDE gives you any advantage. But within an editor, if you're working on this rich thing that you want to visualize diagrams, work with multimodal inputs, have these other affordances that you want to make part of your context, it's a more powerful, more visual interface to do it. You're not just prompting at that point of time. You're doing a lot more than that. But I agree - if you're just prompting an agent, CLIs are amazing.
+
+**Jerod Santo:** Okay, good thoughts. The second decision that you all made - at least I'm assuming you made this - was "We are going to be..." I guess not model agnostic, but "We're going to pick a model that you're going to use, and then you can change that model..." There's different angles of this. Like, obviously, some people have their own model they want you to use, like Anthropic ones, you know. The model behind the agent is like a huge part of the system, right? And so Kiro, as I downloaded it, it selected Cloud Sonnet 4... And there's a dropdown selector. For me, I can't pick any other option. I'm assuming I could configure that to pick different options.
+
+**Deepak Singh:** Not yet.
+
+**Jerod Santo:** But just your thoughts around model selection... I mean, it's a big part of it, because at the end of the day, you are wrapping a model that you have to decide what to do with that.
+
+**Deepak Singh:** This is a bit of a philosophical debate we have within ourselves as well. So the reason the dropdown is there - there was a time that Kiro allowed you to pick between Sonnet 3.7 and Sonnet 4, because sometimes you run into availability issues with Sonnet 4; there was just not enough capacity, so you could choose Sonnet 3.7 and do your work there. But that problem went away, and nobody cares about Sonnet 3.7 when Sonnet 4 is available. You could always pick Sonnet 4. So we took that out.
+
+**Jerod Santo:** Okay. So that it's not going to say GPT-5 next to this, ever.
+
+**Deepak Singh:** No, but if you download it now, you will get two options. The default model -- I won't say model. The default agent in Kiro is something that we call Auto, and I'll go into this in a second. I'll actually go back to Sonnet 4 as well. What you're using in Kiro is not Sonnet 4. You're using an agent that uses Sonnet 4 for the core thinking part of its work. There's a set of scaffolding that goes around it; everything from your system prompts, to the way you handle caching, to the way you are handling any guardrails that might be in the picture... Sonnet 4 -- let's take two editors; they both say Sonnet 4. They're not equivalent. It's how you drive context into the LLM, and the scaffolding you have around it that makes it useful, or less useful, or more.
+
+So in some ways, you could argue that saying Sonnet 4 is also almost a comfort level for people saying "I've got Sonnet 4 inside it", because it is an agent that's doing the work. It happens to be an agent that is built on Sonnet 4. There could be pieces of it, and actually that's true for many agents out there, where for saying hello I don't actually want to use Sonnet 4. That's a very expensive hello, as an example.
+
+**Jerod Santo:** \[32:23\] Yeah. Well, I was just thinking about that, because you -- I didn't want to call it a wrapper, because that seems like that's just belittling the amount of effort that you all have done. But basically, in wrapping a model like that, you're making a pretty large bet on the provider of that, and... Choosing them, versus being like "Well, you can use Gemini, you can use ChatGPT, you can use Anthropic."
+
+**Deepak Singh:** The good news is because we have a dropdown, if we want to go in that direction, that's fine. Now, where we have decided to go, at least for now, is an agent called Auto. Auto is using multiple agents underneath the hood. We can choose whichever largest frontier model we want to choose. Right now, we tell people - it's there in the documentation - that one of the models that we use is Sonnet 4, because that's a model that we know well, we understand well. All our agents use Sonnet 4, Q CLI also uses Sonnet 4... So we have all the scaffolding in place and we understand how it works. But we want more -- there's a number of ideas we have around how we handle the model based on the tasks we have, including maybe some specialized models for certain things.
+
+So Auto is an agent, it's not a model. It uses multiple models underneath the hood, one of which happens to be Sonnet 4. And over time, we may have different categories. Sonnet 4 sets the bar, in some ways, for how good the results need to be.
+
+The way we think about it is our quality of results with Auto have to be at least as good as the agent that uses Sonnet 4, the one that says Sonnet 4, but we want to do it at a fraction of the cost and hopefully better performance. That's a mental model. But again, we haven't taken \[unintelligible 00:34:06.10\] because I think this is a philosophical debate within the community, within our team, on what's the right thing to do. Because the funny part is, if you give people five agents, the chances are that most of them will just pick the one that they think is the best anyway. So that's where we are right now. Where the future goes, we'll see.
+
+**Adam Stacoviak:** Choosing models as a -- I think there's a cost perspective there, right? And if you remove the cost perspective to the individual user... So I guess the context here really is like we're drastically different types of developers, in terms of like, I'm a solo. I'm not doing it as a team, even though Jerod and I are both here. I'm not working on like Changelog web proper stuff. It's like scratching my own itch, CLI type things... Just really learning and exploring. And so a lot of my context and even my preferences are largely based around a team of one. Whereas Kiro is really a team of many, and so the choices you made were really around teams, and context sharing, and things like that. And in a world like that, you usually have some version of an enterprise cutting the check, and there's less concern about ultimately the costs of an agent and how it may impact you. Obviously, you've done your cost analysis, but that's largely something else, usually.
+
+**Deepak Singh:** Yes and no, because I think if you go directly with a cheaper model, the challenge is people will do that, but then will get very frustrated with the quality of the output. What I care about is the quality of the output. And that works for individual developers, that works for your self startup that's funded by Y Combinator, it works for the enterprise. I actually think, in some ways, enterprise developers work under more constraints on how creative they want to get than developers in other places.
+
+\[35:59\] So the way Kira thinks about is, the customer that we want to really make happy is an individual developer. That individual developer could be a solopreneur, somebody sitting at home and scratching an itch, somebody sitting as part of a small startup, maybe the technical person in a two-person startup, for example... They could be a developer at a startup that we all know, one of these series D unicorn startups, or it could be somebody at Delta, right? ...to pick a big Q developer customer. All of that is possible. So there's a few things that we did in Kiro that point to that. One is that the default login system in Kiro is social. You either log in with your Gmail, or you log in with your GitHub login. It's not you need an enterprise system. That's a big part. Again, going back to decisions that we made, a very conscious one - it's actually one of the few, if not only products that have been built by an AWS team that you can sign up with Gmail and have the ability to pay. You don't need an AWS account. In fact, you don't need anything. You need no AWS account, you don't need to worry about AWS permissions. Zero. There's no requirement at all. If you happen to be an AWS customer, there's ways you can get signed up.
+
+So that was a big part of it. What we did decide to do was assume that people want to write code that's going to -- a big part of the idea was "Yes, there's a lot of fun projects out there that people write." I'll call them toy projects. Again, I'm being sort of general. Projects that you do to learn, etc. and you can do them with Kiro. You can vibe-code with Kiro, just like with any other agent. But our assumption was people want to take code and actually do something meaningful with it, and grow it over time, and evolve it over time. So that was very important, and it is also true that a lot of people do that as part of a team. So we wanted to keep that in mind. Because the team could be two people, it could be a two pizza Amazon team. But the fundamental assumption was that the person signing up is an individual somewhere, whether they're at home or in a company. And I think finding that balance between the fun part - I forget which one of you mentioned, you have to... Like, development is fun. You can't take it away. Especially developing with AI agents. You don't want to make it like a boring, put in a bunch of checks and balances because you assume it's going to be in the enterprise. And I think that's kind of where we are right now. We'll see as it grows and where people push us which direction we end up in, but there's a lot of white space right now in this whole area.
+
+**Adam Stacoviak:** We're just all vibing our way there.
+
+**Jerod Santo:** We're getting there eventually. We'll get there eventually. How many -- are there AWS, are there two pizza teams shipping code with Kiro? I mean, I know you guys have a big wait list, but you have a great internal team that you can just test stuff out on and say "Hey, use this sucker."
+
+**Deepak Singh:** Well, the best example is the Kiro team. The Kiro team has just been using Kiro from the day Kiro started.
+
+**Jerod Santo:** Really?
+
+**Deepak Singh:** So they built Kiro in Kiro. I don't have any stats on it, but the one use case we mentioned is this one engineer who wanted to ship a feature, a notifications feature in Kiro, and he did it over a day because he wrote a spec for it and then the agent wrote the code, and maybe somebody else shipped it... But in general, part of one of the tenets we had in Kiro, and I have it for a lot of my teams is you want to build it almost for yourself. Developers work best when they're building things that make them happy developers... And your development team builds something that makes you happy. Then, of course, you give it to a bunch of customers and see whether their happiness and your happiness happens the same way; developers are picky people. And that's when you start making the choices.
+
+\[39:51\] The spec editor in Kiro actually evolved significantly, because the way the team was using it worked for them, but as we gave it to other people, it didn't work quite as well. And we almost rewrote it -- I mean, we did rewrite it thrice, I think, in the user experience, to get it to a point where the majority of our customers really liked the experience; most of the people in our beta cohort. And so that's one.
+
+I'll give a more general one, which is we are at a point inside the company which actually helps us a lot learn behavior. It's that most developers now, 80%, as I said earlier, are using AI systems to build code, where we have now got significant projects which are of significant complexity, which are 80%, 90% written using AI. So we are moving quickly in that direction.
+
+**Jerod Santo:** That's really cool. I would love to see some of your guys' specs; like, from the Kiro product team. Not open source the whole thing or whatever, but like "Here is a large production code base that was built with spec-driven development. And here is a bunch of specs that you can look at to see what we've ended up with as we've built out this code base."
+
+**Deepak Singh:** That's a really good idea. We've also -- I mean, some of the folks on the team are toying around with the idea of can you have a spec sharing... You know, just a place for people to share specifications, which include not just the spec files, but also steering files etc. Because they all go together. In the end -- I think this is what Adam was saying... Context is not as one-dimensional as people like to say it is. It includes a spec, it includes a bunch of other things. Your prompts, your steering files... Is there a package way of sharing all of that? You have agents.md now, right? It's a very powerful sharing mechanism, because it's a standard format. So I think there's going to be interesting evolution even in that space. And as you said, I think the community is sort of converging into the idea of specs. It's whether they are doing it deliberately, or sort of everyone's heading there... So we are very excited to be part of that, because we've had this religion for a while.
+
+**Jerod Santo:** So inside of QKiro, for our listener, the left-hand side, when you have the Kiro panel open, there's one, two, three... There's four subsections: specs, agent hooks, agent steering, and MCP servers. So Deepak, how many of those produce artifacts that end up checked into your source control? How many of those are trackable things?
+
+**Deepak Singh:** Yeah. I mean, they're text files, you can check all of them in.
+
+**Jerod Santo:** All of them are.
+
+**Deepak Singh:** They're sitting on your file system. Yeah. Everything in Kiro is a text file on your -- I laugh because for me they're on my right side. The way my brain works, I've moved all of them --
+
+**Jerod Santo:** Oh, sorry.
+
+**Deepak Singh:** But it's different. By default, I think it's on the left.
+
+**Jerod Santo:** This is the default. I just popped it up open and --
+
+**Deepak Singh:** Yeah, I moved everything to the right. I like having them there. There's a reason workspaces are configurable, because we're all picky.
+
+**Jerod Santo:** Yeah, totally. And our own way is always the best way, isn't it? Like, my way is better...
+
+**Deepak Singh:** Always.
+
+**Jerod Santo:** Yeah, you can have yours, but mine's the best.
+
+**Deepak Singh:** Yeah. But they're all markdown files in the end, or JSON files. They can all be checked in. And that's the beauty, that in the end Git is your source of truth.
+
+**Adam Stacoviak:** It's so wild how we've come back to markdown files and JSON files... I mean, honestly, you can have data serialized into JSON... You can read it as a human; it's not that fun, you can still do it... But you can still build an interface out of that through a script that says "Turn this data into human readable", or "Markdown it", essentially. And those two different document types to me are key.
+
+One thing I'm not seeing in this, and tell me if this is something that's on your roadmap, is I feel like you've got the initial spec, you've got the design, and you've got some version of a task list that takes you to an implementation. Feature is done. Or feature is in quotes done; it hasn't been QA'd, it hasn't been deployed... It may spike an incident, there may be a bug...
+
+\[44:06\] And something which I really feel like is what I'm calling the agent flow part of it is it's not just this initial spec and the feature that gets written, it's the bugs that may come after it... I started to do things called builder logs; they're blogs, but I've enjoyed reading what the agent wrote about the feature we delivered... Because it helps me learn. So I'm using it as a means to be a better illuminator and clarifier, essentially. You know, how can I better understand the system I'm defining? And the builder logs are part of that, incidents are part of that, bugs are part of that, knowledge-based articles... You know, what do we learn as a result of this exploration? Or even in this little riff, or this vibe; not even writing software, but what do we learn from poking holes in the API or the CLI, or what really is or is not there? What really is or is not possible? When we have to go from API to SSH, because the API lacks an interface to certain parts of the system, then we actually have to go in via SSH and sniff around, or whatever. So all these little weird things that happen as you build software is largely part of like bugs, incidents, knowledge-based articles, builder logs...
+
+**Deepak Singh:** Yup.
+
+**Adam Stacoviak:** Do you have plans for the full ecosystem beyond this?
+
+**Deepak Singh:** I'm glad you asked this question. I will say the way we are thinking about it right now is where hooks come in. Because I don't think we know exactly what's the right way to do it. Like, Kiro will summarize what it's done for you and say "Here's all I did", right? You can take a look at all the actions it took. That you can go find. Our current mechanism for doing that is the hooks part. Hooks is a great way for people to hook into - no pun intended - into Kiro. The hooks can come from you, they can also come from a third party, by the way. It's a great integration point, because it's easy for somebody else to write a hook...
+
+So for example, you want to take a code and document it properly, create a -- let's say you're writing Java code and you wanted Java docs created, or something like that. I'm not a Java guy, so I have no idea how that works. But you wanted to convert that into a really nice documentation that you can upload onto a website, and maybe even publish it. You could create a downstream hook that is essentially a documentation agent or another agent that takes a look at it and converts it into documentation. Maybe somebody else has written it, there's another startup or some open source project that does it... You can plug it into Kiro as a hook, and that will do it automatically. So you still don't have to drive it. The good news is that the hook is using the same context as the rest of the project. So the context gets preserved. So that's one example.
+
+On the bug side, one of the things I would love to do actually as learnings is "Can you convert your learnings into a steering file that can then be used as a feedback loop?" The good news with agents is they're really good at reflection. You give them error logs and feedback, they'll get better and better over time. How can we do more there? It's not something that's built into Kiro, but you can imagine that that's something that ends up happening.
+
+And I think what we learn over time is what are the kind of hooks that everybody does? And that makes -- to me as a product person, that's a capability that we should probably just take care of within Kiro. But over time, we want people to be able to extend it in the ways that they like and learn from that.
+
+If you look at all the successful things out there that people enjoy using... I'm a heavy Obsidian user, for example. The whole plugin Obsidian extension communities is a big, rich community that's out there. But there's a set of core plugins. And what's the core versus what comes from the community is something that we learn over time with hooks. It's early right now.
+
+**Adam Stacoviak:** It is early. So the hooks essentially is your way to allow yourselves to operate within a certain known user experience, a certain clarified way that you expect Kiro to be used, but it allows folks to extend it beyond and add to it that you would not maybe otherwise want to prescribe, because it's white space, as you mentioned before. There's a lot of vibing happening. I think we're all learning too, as we do this vibe code type stuff, and allowing the agent to write all of it, versus stepping in, even if we have the technical abilities to do so... How mature is that? Is there anything out there in the hooks part of it?
+
+**Deepak Singh:** \[48:30\] The mechanism is fairly mature. There's a bunch of hooks that have been written by our beta customers and early users for various things. One of my favorite hooks that's out there -- the classic hook that's actually part of the sample hooks in Kiro is a hook that translates dictionary. If you change your English language dictionary, it'll go and update every other dictionary for you. So for localization -- it's a localization hook. Localize my code, basically. It's the kind of thing that people hate doing, but you can do it automatically, and if you make it a hook, every time you check in code, it'll automatically go and update all your dictionaries for you.
+
+I know when somebody else is writing a game in a JavaScript framework that they were not very familiar with -- I think they were writing it in Vue. They didn't know Vue very well. So their Vue code -- and they didn't know if the agent was perfect. They had an optimizer that ran every time they checked in code. There's a lot of things that you can do. I've seen people hook it up into code review tools... So I think that'll happen, and I think part of what we need to figure out is what's the right mechanism to make it scalable, to use that term.
+
+The bug thing is kind of interesting. I'm going to go outside Kiro... In CloudWatch we have -- and by the way, you can hook that into Kiro quite easily. We have an agent which you call incident response, that every time an alarm gets triggered, it starts creating hypotheses on where in your system you have an issue that the alarm may have been triggered. You can probably convert your root cause into a spec. I'm speculating. It's probably not that hard to take "Here's my root cause, here's where the code -- this is why the issue happened." You can create an issue -- that's what people will do. They'll file a ticket to do it. But you can even create a spec that says 'This is the problem that we need to go solve." Feed it back into something like Kiro. And with hooks, you should be able to automate something like that, is my guess. I haven't tried. I don't know if anybody else has. But these kinds of workflows will become more common.
+
+**Adam Stacoviak:** What's interesting about this, hooks, if I understand it correctly, is there's things that as you're doing something with the agent, these hooks are sort of like side things you want done, like a formatter, or as you mentioned, a localization thing... Maybe you're even running a build. And in the case of, say, Claude Code, I'd have to open a new tab and instantiate a new version of Cloud with brand new context to not be blocked. So I may do something that a hook may do, but I'm doing it, because there's no concept of hooks in the Claude Code world... That I'm not using at least. Maybe an MCP server can do that, but still, you're gonna have that context window or that active window you're working in consumed by this kind of like a side task... And you're kind of blocked. You're blocked until you can do something else. So that's kind of cool, how you've solved that problem there.
+
+**Deepak Singh:** Yeah. It's something that the team did pretty early... I mean, I think one of the nice advantages of building a tool and using it to build your own - it's somewhat meta, but that's kind of how it works... And hooks came out of some of those ideas, because "Hang on, we need to do that." So it was a very early feature in Kiro.
+
+**Jerod Santo:** Yeah, it seems like it has tons of potential. What are the timings of the hooks? When can you actually fire off a thing, or...? I'm sure there's a life cycle of some kind.
+
+**Deepak Singh:** Yeah, at least the way it is today - it's like a watch API on a file system or a repo. It's looking for some change.
+
+**Jerod Santo:** Okay, so when files change, it does things.
+
+**Deepak Singh:** Yeah. It's an if-this-then-that kind of thing. If this happens, do it. But the key part is what Adam mentioned, is it retains the context of the project, which is very, very important. It's not gonna go randomly start from scratch and start it; now you have to feed it all the context back in. It already has all the context, so it's gonna work within that context.
+
+**Adam Stacoviak:** \[52:29\] Yeah, I feel like I'm just a context mover and saver, is all I am. I wanna keep the context, I wanna move the context, I wanna save the context... And then in the case of Claude Code, when that infamous auto-compact comes around, I'm like "No!! My gosh, please! Stop destroying my world", you know?
+
+**Deepak Singh:** Auto-compaction is such an interesting thing, whichis necessary, because you want to conserve tokens, but it kills your caches, it kills your context... And all kinds of challenges.
+
+**Adam Stacoviak:** Well, then you -- you know, as a user, I just feel like it's a ticking time bomb constantly. I just feel like such anxiety... I'm like right in the middle of the best part, we're finally to the nugget... You know, we've riffed back and forth enough to finally get the clarity, and it's like, "Okay, draft or update this spec." Or in my case, it's a PEP. I've actually borrowed the Python way, and even the acronym too, PEP... But I call mine Project Enhancement Proposals, versus Python Enhancement Proposals. But by and large the same concept, where you've got this initiation of whatever it might be, it's got statuses and all those things... And that's what seemed to work for me, updating that. And I can't do it, because I've got to wait for the auto-compact to happen... Or then it comes back and it's like "I think I was doing this based on what this history says", but then you kind of reset the zero. I understand what happens, but it's always like this anxious moment as a user to map around.
+
+**Deepak Singh:** Yeah. As you were talking about it, I felt like driving an electric vehicle and having range anxiety... \[laughter\]
+
+**Jerod Santo:** Yeah, similar.
+
+**Deepak Singh:** It feels like that.
+
+**Adam Stacoviak:** It's like, how many miles till I can get there to charge up? Yeah, exactly. I think eventually that will get solved, and I understand technically why it's there... But as a user, all it does is basically give me anxiety. And then I've got to work around it, and moving context, and saving context, and... I mean, I've even started to save chat sessions, because you can do /export, copy to your clipboard, or move it somewhere and kind of save at least the chat history so that you can read back from it. You may not have the full actual context that was saved, but you at least have some version of a map of how you got there.
+
+**Deepak Singh:** I'd be very interested when some of the new Kiro builds come out. We've got a couple of enhancements coming out this week in the clients... I'd be very interested in you running long-running sessions and seeing what happens... Because we've tried to do a lot of work in managing that context, in managing that and making sure you get the right behavior out of it. It's not perfect, but I'd be very curious to see your experience, whether the range anxiety gets mitigated or just gets a little bit better.
+
+**Adam Stacoviak:** Let's talk about that then, because I want to talk about stack... Because my stack for what I'm doing is pretty much any agent CLI. And Jerod, you'll be happy to know that Zed is my daily driver now. I've thrown every other possible editor to the wind, because Zed obviously is fast and beautiful... And the only thing it lacks is some of the things that Sublime Text still uses, but... I digress from that point there. But I'm using some version of a CLI in the terminal, and then a separate editor. And I would not call Zed an IDE, although it has IDE tendencies... That's my stack. And the reason why it's my stack is one, it's accessible to me... And I've just learned about Kiro, and just got access to it... But even with Augment Code, I used it inside of VS Code, and me as a user, as a visual person, as a designer, VS Code is great, but the text is small, the windows are spaced out...
+
+\[56:16\] What I like about the Claude Code, Augment Code, AMP Code CLI flows is that the user experience of using them - it's a larger text. I can see it better. I feel like I've got more clarity with... With the vibing, for a lack of better terms. And then I still have this really awesome, Zed is the best editor, to still go and review and look at the actual code, go around the entire documentary that is the project...
+
+To me, while I would love to use Kiro, the thing that would hold me back is that you're in this VS Code-like world, and it's just less than desirable, I would say. It's not bad, it's just not my preference.
+
+**Deepak Singh:** Yeah. I think we talked a lot about back in the day building something completely new, like Zed did. I like the fact that they call it Zed and not Z... Or working off Code OSS... And the reason we ended up -- it was very clear to us, because we used to have a plugin, a classic VS Code plugin. We still do, which is on the Q developer side. But it became pretty clear to us from a UX affordance side that was going to be a very, very difficult place to be. We wanted to do things in the user experience that as a plugin were just hard. So having code OSS as a starting point was great.
+
+I think the interesting point is will we stay with code OSS, and at the end you fork it so much it becomes its own thing... I think that's to be determined. There's a lot of that. One of the other advantages of using code OSS is there's so many people using VS Code, and they all have their plugins and extensions, and just pulling them into Kiro to do things makes their life easier, as opposed to starting from scratch. That was actually probably one of the biggest reasons that we stuck with the Code OSS part. But the goal was always making a UX that was optimal for the flows that we want to get to.
+
+And I don't disagree with you, I think UX is also a very personal thing, and how it works... I'll always go back to "I just like, personally..." I never liked using code editors. I liked using text editors. But I'm not a really good programmer either. But I think with Kiro -- because now you're not in code editing land, you are 'vibing with an agent' land to create specs. Our user experience is going to keep growing, and how can we make that a true, rich multimodal experience and keep evolving it? Where it ends up again, I don't know, but that's the goal.
+
+But actually, what you said, the stack part is very important. I do feel like most people will end up using a) what I'd call a more rich experience, which you can call an editor, an IDE, whatever... I almost feel like calling an IDE is doing the thing a disservice right now. So we call it Kiro IDE, but in my head it's like a desktop product, which happens to have a code editor inside it. And a terminal-based product, like a CLI. And I think those two - I mean, because it's VS Code, I actually run Q CLI in my Kiro terminal as well. I'll often also have a separate shell open all the time.
+
+And one of the ideas I think that's going to be interesting in how it evolves - you see that a little bit with some of these CLI things that now have editor plugins, where you're using the editor just to visualize what's happening... I think there's going to be some very interesting, again, UX evolution of how terminal-based agents and richer UX agents evolve over the next several years.
+
+**Jerod Santo:** \[01:00:07.16\] So progress with Kiro means the VS Code side, the code side, gets kind of tucked further and further away, and the spec editor and this immersive, interactive building of a specification, and then some sort of end product kind of continues to level up and become more and more primary. So perhaps Kiro down the road - you know, maybe you don't have to tell people that yes, you can look at the code underneath if you need to.
+
+**Deepak Singh:** I remember having a debate with the team and "Do we even need a code editor in here?"
+
+**Jerod Santo:** Yeah...
+
+**Deepak Singh:** But they won that battle. Currently so.
+
+**Jerod Santo:** Yeah, I think right now you do. My question for you is, with the current state-of-the-art models, do you think that world that we just described, of Kiro as this interactive app building thing with a code editor hidden in it - do you think that's possible today with just better agents, Auto getting better and Kiro getting better, and leaving all else the same? Do you think we can get there? Or do you think the underlying models also need to improve to bring us to where you wanna go?
+
+**Deepak Singh:** Oh, both. Both.
+
+**Jerod Santo:** It wasn't a both question, Deepak. You can't say both. That's cheating. \[laughter
+
+**Deepak Singh:** I'll be blunt.
+
+**Jerod Santo:** It's either or. \[laughs\]
+
+**Deepak Singh:** Fine, touche. I think we need at least one more generation of models. There are probably models out there that can get you there, but they're really expensive. And their context windows aren't right... There's a lot of limitations right now. I mean, I say that given the fact that over the last year the models have become so much better.
+
+**Jerod Santo:** They've gotten a lot better.
+
+**Deepak Singh:** Some of the things that we're doing today, you could not do a year ago.
+
+**Jerod Santo:** Yeah, for sure.
+
+**Deepak Singh:** You just couldn't. I do think there's one more generation of models before we get -- I still feel like to write a full application... Not a simple CRUD app, but a proper distributed backend... And we have examples of people who've done that, but it requires a lot of skill on the part of the developer to be sure that you're in the right place. I'm not saying that's a bad thing. I still feel like the people who are going to be most effective using these tools are skilled people. There's a lot of folks who believe and hope that you can give somebody like me and somehow make me a better programmer, developer. That's not happening. Can I write applications more quickly and easily? Absolutely. I can prototype better than I ever could. I've written more in the last year than I had in the five years before that. And mostly it's me trying to make a molecular viewer for myself, because that's what I like doing.
+
+**Jerod Santo:** \[laughs\] Have you gotten one done?
+
+**Deepak Singh:** I've gotten like 15 versions of it... That's my kind of test pet project.
+
+**Jerod Santo:** Do they work though? Can you ship them?
+
+**Adam Stacoviak:** A molecular viewer, is that right? That's cool.
+
+**Jerod Santo:** A molecular viewer.
+
+**Deepak Singh:** I used to be a bioinformatician, so I like my molecular viewers. That's my standard test project.
+
+**Jerod Santo:** Would you ship any of those -- you've got 15 of them. Would you ship any of those as products?
+
+**Deepak Singh:** I have no idea. I think it works, but I'm not sure if it works for everybody.
+
+**Jerod Santo:** Right... Okay, keep going.
+
+**Deepak Singh:** So there's a level of skill and understanding that you need to get the maximum out of these agents that we have today. But as these reasoning models get better, as we start adding new techniques to it... For example, there's a fun term these days called Neuro-Symbolic AI, which is basically bringing mathematical techniques into generative AI systems.
+
+\[01:03:45.13\] So we have a very strong Neuro-Symbolic AI team over here... By the time this podcast publishes, this is probably going to be well-known news, but Joe Hellerstein, who's well-known in the database community, he has a project called Hydro. Hydro is an open source project to basically verify the correctness of a distributed system written in Rust. We want to bring those kinds of techniques and build them into the spec engine of Kiro. So as you're building software, you don't need a human at the other end to verify correctness; that you can verify correctness using a mathematical model, in this case.
+
+**Jerod Santo:** That'd be a game changer.
+
+**Deepak Singh:** I think those are the kinds of techniques that take us to the next level. We're not there yet, but I think it's going to happen within - I don't know if it's six months, one year, two years, three years... But it's gonna happen.
+
+**Jerod Santo:** Yeah, that'd be a game changer, for sure. It seems like everything's 80% solutions right now; or maybe even better than that, maybe like 95%. And then like that last mile is where you need to be a senior engineer to actually recognize and finalize the last mile of delivery on all of this stuff. And it's limiting --
+
+**Deepak Singh:** There's an intuition that they have that you just don't get otherwise. And as I said, to the question you specifically asked, I would say "No, not yet, but we're not that far away."
+
+**Jerod Santo:** Let's talk about pricing. This thing's a product you all are trying to sell... How do we buy it?
+
+**Deepak Singh:** So we changed our pricing twice already.
+
+**Jerod Santo:** Okay.
+
+**Deepak Singh:** So this is a pricing that we announced - and I'm going to get it wrong - last week, and it'll probably go fully live, we'll start charging people in sometime in the next few weeks. It's a classic per user pricing. You get a free tier... You actually get a free trial, because the free tier is mostly just for vibe coding. The free trial gives you the full experience. I think we put you into the pro tier for about a month, and you can play around with it, and then if you like it, you can actually start paying, or you can flip back to the free tier... But there's a \$20 tier, a \$40 tier, and a \$200 tier, that give you 1,000, 2,000, and 10,000 credits. There's a credit system.
+
+The way we talk about it is if you're using the Auto model, you will get those -- credit consumption for Sonnet 4 is 1.3 times the credit consumption of Auto. You just get a set of credits, and as you can see, if we gave you a selection of some kind, we would have different credit consumption rates based on the cost of running those models... So you get two agents - an auto agent and a Sonnet 4-based agent. The Sonnet 4-based agent uses credits 30% more faster typically, because it depends on the prompt and the -- LLMs aren't deterministic. But for the most part, a 30% more than faster rate of consumption than the auto model. And we also have overages.
+
+Then another thing that we did was we added the idea of fractional credits. So when you go, you can go from 1 to 1.1. You don't go from one to two in these step changes. We've found people were going through their credits a lot, because they used a little more than the boundary, and then they would get sent to the next one. And that meant that you just consume credits really quickly.
+
+The other thing we have is overages - so if you're in the \$20 plan, you don't need to go to the \$40 plan if you just want to spend \$21 for the project you have in hand. So there's overages there, and then the 201 assumption is most people will rarely get there, but we know there's some people who will, because they have.
+
+But we spent a lot of time in trying to optimize that credit consumption, and at least in our early testing it looks like you're not running out of credit so quickly. We add in the middle, separated out, the amount of usage, the way we were charging for specs versus vibing, and people found that too difficult to wrap their heads around, so we've combined it into a single credit bucket.
+
+**Jerod Santo:** Gotcha.
+
+**Adam Stacoviak:** I'm still confused.
+
+**Jerod Santo:** \[01:08:10.21\] You're still confused...
+
+**Adam Stacoviak:** I'm still confused.
+
+**Jerod Santo:** Yeah, it's probably simpler on the webpage.
+
+**Adam Stacoviak:** Well, I don't even mean your description of it. I mean just generally, because I think AMP Code and Augment Code, they use messages as their terminology, whereas you're using credits...
+
+**Deepak Singh:** That's what we used before. We used credits.
+
+**Adam Stacoviak:** And how that maps to usage is sort of like a black box to me, because -- I guess in those cases... What I learned about AMP Code when I talked to Beyang Liu, CTO of Sourcegraph - he helped me understand how because I had a larger context window, it was actually consuming more credits... Which I thought was like "Well, that's convenient for you, because obviously I want context", and so I'm burning faster. It's super unclear to the user... And I'm like "Gosh, you're just super expensive." He's like, "Actually, you're just holding it wrong." I'm like "Well, I'm not actually holding it wrong, I'm just holding it, period. I don't know how I'm holding it."
+
+**Jerod Santo:** Right. "How do I hold it right?"
+
+**Adam Stacoviak:** Right. So this mapping of usage to messages, to credits, to me -- and it's not your fault. It's just... It's hard to understand.
+
+**Deepak Singh:** It's tricky. So one of the things you've gone back and done is - because it became pretty clear that's what folks wanted - was actually show you real-time usage. I don't know if we'll have it right off the bat, but I know we are working on it... We now show you, as you're using it, how your credits are getting consumed. Where we want to get to, and we'll get there pretty quickly, if not in the next couple of weeks, is actually showing you how a particular thing that you did, what kind of credits it consumed. Because to your point, consumption of credit, how much you're using is not just about your prompt. It's not even about the context window. It's about tools. If the LLM decides to use 200 tools instead of five, it's going to use much -- it's going to use more credits.
+
+**Adam Stacoviak:** That's the danger of MCP servers. They're amazing, but it also enlarges your context window, too. So the more tools you use from those servers...
+
+**Deepak Singh:** Yeah, if you preload your project with 5,000 MCP servers, you'll run out of credits in like five minutes, because that context will get so large. I also think as a community, we need to start becoming much better at educating people on best practices.
+
+So I think part of where this will end up doing is you'll end up having really, really nice -- I won't say really nice, but mechanisms to understand what's the optimal way to set up a project, how do you do things that you always want, versus bring them in as you need it, showing you the right visual cues... And the other problem is - and I'll go back to tokens here, because in some sense that is the currency. A token in one model is not equal to the token in another model. Tokens are also different. People put token weights etc. on there, but that's a heuristic.
+
+So I think the thing that people are struggling with is how LLMs work is not deterministic by the way people's brains work. How things get consumed and cost accrues... And quite honestly, I think you're seeing the entire industry deal with it over the last six months, in different ways. This is our way. We think it's going to work for what we want, or what our customers are going to do, but a lot of the responsibility is on us to be very transparent about when you submit a prompt and something happens, you at least get a sense of "This is what this takes." And if you have different agents, the fact that they behave differently, you can make decisions based on quality and speed, on whether it's worth -- you may want to go with the more expensive agent for whatever reason, but you have to know why. And I think that's going to be -- and that's hard, but that's what we're trying to do.
+
+**Jerod Santo:** \[01:11:53.25\] Well, it's certainly still the early days, I think. I mean, I think this is evidence of the trailblazing that is going on, and just the figuring out of how it all has to work, and how it should work, and then educating ourselves on how to hold things optimally, not wrong... And picking the correct model, or having a model picker who's orchestrating behind the scenes, like where you send any particular prompt to to save money and speed... I mean, it's all very much the land of trailblazing, isn't it? I mean, this does not sound figured out, by any means.
+
+**Deepak Singh:** Yeah. I think the other thing is how people use these agents and how they get success out of it is sometimes learned only after doing it. You don't before know "This is the best way to use this agent to be successful." The whole idea of what became specs and these memory banks and prompt libraries and steering files - I think this is the equivalent of trying to learn how to walk falling down, trying to learn... Again, it's almost like your first baby steps; that's how you learn. And I think we've gotten to the point where we are, not because all of us were smart and knew what we were doing. I think as an industry, we are constantly learning from people's behaviors, how they use these things...
+
+Inside Amazon we have this power user Slack... I get more understanding of what's the best way to run an agent because of people who are using these agents in anger every day, and learning better and better ways to use them. And they come up with things like agent flow, and that gives product people product ideas.
+
+**Jerod Santo:** Right.
+
+**Deepak Singh:** But that's just us. I think the whole industry is doing that. And so it is early days, but it's also a lot of fun...
+
+**Jerod Santo:** Yeah.
+
+**Adam Stacoviak:** That is the hand of innovation, is using things in anger.
+
+**Jerod Santo:** That's right.
+
+**Adam Stacoviak:** It creates itches that you scratch, and next thing you know you're doing something completely --
+
+**Jerod Santo:** You're innovating.
+
+**Adam Stacoviak:** Yeah, you're on the edge there. I'm gonna mention something that's slightly dystopian, even at this close to end of the sho --
+
+**Jerod Santo:** We almost made it through without any dystopian -- yeah.
+
+**Adam Stacoviak:** Well, are you guys both familiar with the idea of a toll booth?
+
+**Jerod Santo:** Of course.
+
+**Adam Stacoviak:** The way to get rich is to create a toll booth. Right?
+
+**Jerod Santo:** One way.
+
+**Adam Stacoviak:** Well, yeah, sure. One way. Well, by and large, if you look at most of the way people develop wealth, it's some version of a toll booth, in metaphorical terms. It could be real estate - that's a toll booth, because you need somewhere to stay. I've got the toll booth, you want to stay? Boom, I make my money. Or in the case of something like this, I'm thinking of it like - the future of software development. I see it all going this way. We all see it going this direction. And for a while there, the cost to be a developer was your ability to hold something in anger, maybe have a license for a code editor, or in the case of Vim, for free; it's open source. The machine you had to buy - there was no toll booth in front of your ability to create and innovate. The toll booth was maybe the books, or the education process, or the schooling required, things like that. But you could bypass those things if you were smart enough to learn on your own and just avoid the toll booth altogether. And I guess what I'm struggling with as we're in this conversation is maybe just the fact that if this is the way of the future, what a massive toll booth there is in front of the innovation that drives the world. And I'm just not sure how I feel about that right now.
+
+**Deepak Singh:** Yeah, I actually think it's the other way around... I'll give you some examples. Maybe there's a toll booth, I don't know. But the way I think about it is there are things happening where the cost of doing that thing is very, very high. Maybe I am a semi-technical founder, and I'm busy spending time looking for a deep engineering team, because I can't even build a prototype, because I'm not technical enough to go do that. I have an idea in my head, I can write some JavaScript maybe, but my idea requires some heavy-duty backend stuff, and I have no idea how to do it.
+
+\[01:16:12.15\] In theory - and I've seen enough examples that this is possible - you can do that on your own, or maybe with one person to do much more work. Once you have that, you can go further. That's one example.
+
+I've also seen people who've had a project that's been sitting on the back burner for a decade, because it is something that they've built -- it's an open source project, they've got a day job, and they don't have the time to go and evolve that open source project, because it takes too much time. They don't wanna spend their weekends doing maintenance of an old thing, maybe. Or you have a backend that's not scaling, and you need to change it, but the cost -- you know you need to change it, but the cost of doing it is like 20 people and a year and a half of development, if you did it the old way. I've seen enough evidence now in all three cases of people building using these agentic development systems, going and tackling those problems.
+
+The first one's an obvious one, people talk about it all the time; that's a founder thing. The second and third one, I think is where the magic lies, which is you are doing things now that you would not have done before, which results in better code, better user experience etc. because your barrier to entry and ability to do it is low enough that what you thought would take you two weeks is now gonna take you a day and a half, maybe even less. What it would take you 20 engineers and a year and a half, you can now do in much shorter time etc. My favorite example is actually of a team that does sprint planning every four days instead of two weeks, because they were finishing their backlog that quickly...
+
+So to me, that's -- I don't know if you'd call that a tollbooth, or shifting of the tollbooth, or whatever, but I do think there's this innate barrier reduction. There's an activation energy to everything - I'll speak in my physics terms - and you need a catalyst to cross that activation energy. And that catalyst could be funding, because you get more people, it could be hiring really talented people... I actually think it's not just the catalyst part. Your activation energy has gone down significantly. So the catalysts you need are -- you don't need catalysts at all in some cases, or you don't need the best ones out there. And I think that will have a meaningful impact on software development and how teams are structured and how we think about it, in a very positive way, because you're getting things done that you're either not doing or wouldn't have done, or would have done much more slowly. At least that's how I think about it. Maybe I'm the glass half full guy, and \[unintelligible 01:18:47.01\]
+
+**Adam Stacoviak:** No, you're spot on. And I concur and agree with everything you just said. But what we cannot deny is that there seems to be a subscription of some sort. And this is where I'm struggling, really... Even though I'm emphatically happy about the circumstance, I think it's cool, I think it's the best thing almost ever... But now I'm struggling to think that there's now a subscription attached to being a developer, or to doing developer things that build software. And the reason why I say that is you go back to a couple of years ago when they said "You will own nothing and you'll be happy." Well, now we don't really even -- like, if this is the way, right? ...and if -- I was just telling Jerod this other day... If Adam and Jerod from 20 years ago, which is today - let's say there's versions of me and him today that are entering the software world... If this is the way, maybe they don't even own their intellect and their innovation ability anymore. Like, now we're renting that, too. And I'm kind of just struggling with that in this moment.
+
+**Deepak Singh:** \[01:19:53.29\] Yeah... The ideas are still yours. I remember the days when I was at AWS a long time ago, when -- I used to have a stack of Linux servers at home that I finally... I don't even know why I have them. It was before I joined AWS also. And I got rid of them and started using EC2 for everything. I don't need -- like, I have this little corner of my house picking up heat... Not required. Like, that was an early one. So I still remember when you switch -- you kno, the time you'd switch from assembly language to more of the modern languages, or even from some language to a Rails type structure... There's always an abstraction or something that comes up that allows you to shift your attention somewhere else. I think the big one is where does that attention shift and how valuable is that? And I think those have historically defined whether something is going to be successful or not. I do think - and maybe all of us have drunk the kool-aid on this one. In this particular case, I don't think we're talking about a 5% -- like, my ability to innovate or my ability to act and my ability to think, it's not becoming 5% better. I actually think I have a new tool that allows me to operate fundamentally differently... And I don't think we've quite grasped what operating fundamentally differently as individuals and teams actually means. I think people are starting to find out, but I think it's going to make an impact in a very positive way. Maybe it's a shifting of the toll booth, and where the toll -- because there's always a toll booth, as you said, somewhere.
+
+**Adam Stacoviak:** Yeah... But don't take it, even though I described it as dystopian, and I don't want to be necessarily negative, although it is a negative thought when you sort of go down the rabbit hole... I 100% agree. I'm personally doing things, just simply exploring and scratching my own itches, that I just would never ever do, because I just didn't have the time to even go to the depth necessary to do it in a way that was even worthwhile doing. And you take that same idea and you apply it to an enterprise team, that has different context windows and challenges and issues... Gosh, yes. For sure.
+
+I was just telling Jerod too the other day that for a while there I was bummed that I thought that "Well, there's going to be a market shift and a market correction in terms of the way enterprises or even teams are built these days." And you may see layoffs, but I think by and large we're going to see a massive, massive influx into folks becoming what are called software developers, because there's tooling that allows them to operate like a senior engineer, or with senior engineer abilities. And that's amazing. I'm all for it. In the moment, I was just like "Gosh, do we now have a toll booth in front of our ability to show up as developers?" And that was just kind of a bummer.
+
+**Jerod Santo:** I think subscription fatigue gets to us all.
+
+**Adam Stacoviak:** It does.
+
+**Deepak Singh:** Yes.
+
+**Jerod Santo:** You know, when you're paying per token, you're basically paying that subscription fatigue. Well, Deepak, I know you are hitting up against a hard stop. We appreciate you spending a few extra minutes with us... Kiro looks really cool, I love all the innovations going on... I think that spec-driven development is a very awesome stake in the ground, to claim like "Here's a way of doing agentic things that is better." And I hope you guys all the best in building it out and realizing that vision of what you're trying to build. It's cool stuff, man.
+
+**Deepak Singh:** Thanks for having me. It was fun talking to you. And as I said before, I think we started -- I've probably heard... I started listening to the Changelog at episode one, so it's great to actually talk to you folks.
+
+**Adam Stacoviak:** That's so cool.
+
+**Deepak Singh:** I appreciate it.
+
+**Jerod Santo:** So cool.
+
+**Adam Stacoviak:** That's beyond cool. Thanks so much for being a listener all these years. Appreciate that.

The 1000x faster financial database (Interview)_transcript.txt

@@ -0,0 +1,489 @@
+**Jerod Santo:** I'm joined today by Joran, creator of TigerBeetle. Welcome to the Changelog.
+
+**Joran Dirk Greef:** Hey, Jerod. So great to be here. Thanks for having me.
+
+**Jerod Santo:** Excited to talk to you. I've been excited ever since you were mentioned by Glauber Costa on our Turso episode. He spoke very highly of you, very highly of TigerBeetle, and I thought, "We've got to get this guy on the show."
+
+**Joran Dirk Greef:** Yeah, so great to hear that. Glauber is sort of part of the mix in some of our design decisions for TigerBeetle... So I've always held him in high regard. And yeah, so I'd love to dive into that later, but yeah, thanks again.
+
+**Jerod Santo:** Absolutely. Well, we'd love to hear what you've been up to. I read on your TigerBeetle blog that you said in July of 2020 you stumbled into a fundamental limitation in the general purpose database design for transaction processing. Can you tell us that story?
+
+**Joran Dirk Greef:** Yeah, so it was the strangest performance optimization challenge that I've had to do... So I was doing performance engineering at the time, consulting on a central bank switch with the Gates Foundation. They had created an open source nonprofit central bank exchange, and I was brought in to see "How can we make it go faster?" So how can we do lots of transactions between banks? How can Alice pay Bob - and those are little payments, so this is the classical database problem, a transaction between one or more people. And if you look at the Postgres docs for Postgres transactions, the canonical example there is Alice pays Bob, databases need transactions, because they need to record people transactions.
+
+**Jerod Santo:** Sure.
+
+**Joran Dirk Greef:** And so the central bank exchange was a good example of why you need a database... Because it's literally money moving from one person to another. And so I came into this project and I had done a lot of performance optimization work, and specifically in Node.js. So I'd been involved in Node.js since Ryan Dahl announced it. I'd kind of already been writing JavaScript on the server before Node.js, using Rhino on the JVM, and I was convinced that JavaScript was going to be server side. So I was already writing it, and then Ryan Dahl came. And then I was in Node.js for 10 years and doing a lot of performance.
+
+**Jerod Santo:** Interesting.
+
+**Joran Dirk Greef:** Yeah. So this switch was written in actually Node.js, to make it a really accessible, open source project. And so it was like "How do you optimize Node.js?", and I had experience with this. But the surprise of this whole thing is that, you know, how do you make these transactions go faster through the switch? Inside, it's essentially just a general purpose database, 23 years old. MySQL was the one, and it's very basic. They're just doing -- if you look at those Postgres docs on transactions, they're doing classic debit/credit in a SQL transaction, nothing more. And that's it. And then I thought, "But how do you make it go faster?" So my first experiment was, "Well, let's give MySQL NVMe." And it didn't go faster. The system could only do 76 debit/credit financial transactions a second. And it didn't go faster. I thought, "Okay, that's odd..." Because normally, it makes the database go faster, Jerod.
+
+**Jerod Santo:** \[08:23\] Yeah. Throw a little hardware at it, it'll go faster.
+
+**Joran Dirk Greef:** Yeah. So I thought, "Well, maybe I did it wrong. Maybe it's not NVMe. Maybe the database is memory bound. Let's give it more cache."
+
+**Jerod Santo:** Okay.
+
+**Joran Dirk Greef:** So we gave it a lot of RAM, and nothing changed. It stayed 76. And then we thought, "Okay, there must be like some CPU-intensive algorithms." So we profiled Flame Graphs, everything. And there was nothing. The CPU was like 1%, you know... Even the database, there was -- everywhere you looked in the system, there was just no bottleneck, and yet the system could only do 76 TPS. Network also. You know, what we call the four primary colors: network, storage, memory, compute; we gave them all hardware, and there was no change. And this was so puzzling for me, because always with performance work there's some bottleneck somewhere, and you fix that, and then you fix the others. And here, it was the strangest problem, because you had to optimize a system where the hardware was doing nothing. It was idle. And people who had worked on other payment systems, they sort of took me aside and helped me along. They said, "You know, it's the row locks." And what they meant was when you do Alice pays Bob through a SQL transaction to represent a real world transaction, you go to the database from the app and you say, "Select me Alice and Bob's accounts. I want to get their balances." And then in the app, you check, "Does Alice have enough money to pay Bob? Yes." Then you do another query back and you say, "Okay, increment the debits or the credit columns, and also insert a record of the transfer." So you've only done two SQL queries within one SQL transaction.
+
+**Jerod Santo:** Right.
+
+**Joran Dirk Greef:** Yeah. And this is what I saw. In reality, the switch was doing 20 SQL queries within a SQL transaction, but they were all necessary and essential. And I've since then chatted to lots of other FinTech people, and usually it's like a -- there's about 10 SQL queries per debit/credit. If you want to do it really well, you could do it just like two, as I described.
+
+And so actually, this was the problem. The SQL transaction was holding row locks on the account rows across the network latency. And even if that's, say, one millisecond round trip and you're doing two round trips, then you've got locks for a millisecond... And that means if your transactions are all sequential, your database tops out at a thousand TPS, a thousand debit/credits a second. And obviously, the next question we're going to say is "Well, let's go horizontal." But it doesn't change anything, because you have these row locks.
+
+And the other problem that we saw is that there's a power law of contention, like the Pareto principle. 80% of real world transactions are through 20% of hot accounts. So for example, on the stock exchange, anyone who's bought shares has probably bought NVIDIA or sold NVIDIA, you know?
+
+**Jerod Santo:** Sure.
+
+**Joran Dirk Greef:** And so 90% of trades are NVIDIA or the super-stocks - Microsoft, and some of the others, you know.
+
+**Jerod Santo:** Right. And a bunch of other ones just sit there, untraded.
+
+**Joran Dirk Greef:** \[12:03\] Exactly. Then there's the long tail. But if you're the brokerage behind the stock markets, if you're powering like 30% of Nasdaq's volume, then most of your SQL transactions are hitting like only 10 rows, your 10 super-stocks. I mean, how many are there? And so they're permanently row-locked.
+
+**Jerod Santo:** The Mag 7. There's seven of them.
+
+**Joran Dirk Greef:** Oh, there's seven. That's it, yeah. And so that's like the brokerage. But the central bank was much the same. They only have like typically eight or so big participant banks around the table. And so all the money movements are literally going through eight SQL rows... But even if you're -- say you're building like Earth's biggest bookstorel; you've got the same problem. You've got the bestseller list of books. And so all your Black Friday transactions are hitting that.
+
+So then we started to see - this is many businesses, many different kinds of industries. But let's go back to the beginning, back to the question. So it was row locks, and we realized it's a fundamental limit, and it's got nothing to do with the SQL database. Whether you pick Postgres, or MySQL, or something horizontal, or cloud... It actually doesn't matter. The problem is the interface of row locks across the network. And that was when we realized "Actually, we can't go faster. We need a transactions database, not a general purpose database." So that was the impetus, yeah.
+
+**Jerod Santo:** Okay. So you're in 2020, you are trying to basically squeeze more transactions per second out of MySQL for kind of a specific type of transaction, right? Like, you are literally meaning financial transactions, not SQL transactions, which could be a bunch of different things that kind of roll out into a single unit of work. We talk about a SQL transaction - that could be lots of different things, and then it gets rolled back, or it gets committed. But you're actually saying "We're tracking and doing debits and credits, financial transactions, and we can't do these particularly simple type of SQL transaction any faster. We can scale horizontally, but we have the network row locks..." I'm sure there's a way in MySQL to say, "Well, let's just fly closer to the sun and turn off row locks", or something. But maybe you're also a central bank...
+
+**Joran Dirk Greef:** That's --
+
+**Jerod Santo:** Yeah... Not a great idea?
+
+**Joran Dirk Greef:** In Postgres, it's almost the default. \[unintelligible 00:14:38.24\] Read committed isolation.
+
+**Jerod Santo:** Oh, okay... \[laughs\]
+
+**Joran Dirk Greef:** You fly close to the sun just by default, you know?
+
+**Jerod Santo:** Right.
+
+**Joran Dirk Greef:** You actually have to -- there's a lot of like tweaking you need to do just to get it safer, but...
+
+**Jerod Santo:** Yeah, no, I mean, at the end of the day you decided "This requires a different kind of database. We should have a transactional database." Is that how you think of it? ...where it's like the core unit of work is a credit and a debit combined together. Is that right?
+
+**Joran Dirk Greef:** That's right, yeah. And I had a little bit of luck in seeing this, because I've been coding for 30 years, since I was 11. Now I've given away my age...
+
+**Jerod Santo:** You've dated yourself.
+
+**Joran Dirk Greef:** Yeah... And been coding for 30 years, and it's been my passion, and self-taught. So I didn't actually study computer science, because I was already reading papers, and I thought, "Well, I'm more like an autodidact. That's going along already." Actually, I've always loved business too, and people told me, they said, "You know, a great subject to study if you wanna understand the schema of business across all sectors, and anywhere in the world, it's double entry accounting." That's sort of -- and there's a quote from Warren Buffett, "Accounting is the language of business." So that's why many -- well, yeah, that's why I went and studied financial accounting. I majored in that at university, because I had a love for business and wanted to understand the schema, general entries, debit/credits. Because with debit/credit you are doing transactions. You're doing multi-row transactions between multiple parties.
+
+\[16:25\] A debit/credit transaction could actually be you have many debits on one side and many credits. You're expressing something essential in life across -- and it's not only businesses. You could use this not only for money, but for inventory, stock counts, or just counting things, counting API usage, or counting kilowatt hours... It's basically "a counting", you're just counting. So any domain where you need to transact with counts, and it's quantities... It could be like valuable or not valuable, but it's just things are moving, moving around, and that's transactions.
+
+And yeah, that's sort of what TigerBeetle was meant for. Sort of the canonical, general purpose example would be debit/credits. And so often we see this example given... But we thought, "Wouldn't it be great if you just got a database that did this out of the box?" That lets you in one query to the database execute 8,000 debit/credit transactions, instead of having to do 80,000 SQL queries to do the same. So that's one query and you do 8,000, instead of 80,000 queries. And so that's just TigerBeetle, you get these debit/credits. So we thought, "Let's build it in." It's kind of like financial asset; not only row-column consistency, but multi-row column consistency, so like transactional consistency. Double entry consistency, yeah.
+
+**Jerod Santo:** Interesting. Now, I have not been in financial technology as much as you have. I've definitely dabbled, and I was a contract software developer for many years, and so I worked on contract, and I got into lots of different industries... And one thing I've found is lots of different industries have their little niche databases that people on the inside know, people on the outside don't... Or sometimes they're homegrown. I imagine inside these large financial institutions there's probably many - maybe homegrown, maybe there's a proprietary thing that's like "This is the database we use" because of the exact specific use case that you are speaking to. Now, am I right about that? Are there TigerBeetles living in large financial institutions and you decided "We want one for the world"? Or why did you decide to build this thing besides maybe go find one, or...?
+
+**Joran Dirk Greef:** Yes, I think we're mostly right.
+
+**Jerod Santo:** Okay.
+
+**Joran Dirk Greef:** And the reason is because yes, everybody has to build this database, but what you usually see - and I also had no experience in payments specifically as I worked on the central bank switch. I came in from the technical and performance angle. I had the accounting experience. I could see that, yes, we're doing debit/credits, and... It was the payments people that also took me aside and they said, "You know, every single FinTech or -- this is only FinTech, but it applies to any business that needs to record transactions." Any money, you know.
+
+**Jerod Santo:** Sure.
+
+**Joran Dirk Greef:** But they said, "You know, they're all reinventing these debit/credit databases." And typically, what they do is they do it over a general purpose database. So I think this is also interesting, is that it's a kind of, again, lucky on timing, because SQL is like 50 years old, and it's been able to power OLTP transaction processing in the business transactions for 50 years. And so this has always been a latent problem, but we were starting to hit it now on the system.
+
+**Jerod Santo:** \[20:19\] Gotcha.
+
+**Joran Dirk Greef:** And that's because things I think are becoming more -- you know, the volumes are increasing, but... Yeah, back to your question, so I think, yeah, all these FinTechs were reinventing these databases, but they were building a database and didn't know it. So they were building a database in the application around... So that's why there was this central bank switch, and you can see the whole thing as a debit/credit database. It's like tens of thousands of lines of code around MySQL. But if they just had an OLTP database with a debit/credit interface, the switch would be so, so simple. And that was -- so that's how this work started. There was no desire to build a company, build a product, sell something... I was also passionate about the mission, because the switch was open source non-profit. I was told - and this was the mission, and it is - the cheaper, the more performant you make the system, you will reduce fees. The developing countries that run this will be able to lift a few million people above the critical poverty line. And it is literally actually, I don't know, on the order of about a billion people around the world that don't really have access to banking, like we take for granted in the West... Because they just don't have -- in those areas they have to walk miles just to move money. But everybody has a cell phone. So these systems were being built that you could power people to move money instantly on cell phones. But the performance is so important because as you make things faster, you make them cheaper. And now, for someone who's got -- they're sending like $10 back home... If you're taking 5% fees or 1% fees, that's life-changing, because that 4% means so much to them. They're going to maybe have an extra meal, and it's going to compound. So this difference in fees - that is why they are doing this.
+
+So for me, it was really like -- we do want to make it faster. I'm not just consulting and I'm not -- there are like a thousand of MySQL knobs I could have tuned. We could have -- we did double performance. We could have made it a little bit more... But it was kind of like "We actually really want to make it really fast out of the box."
+
+**Break**: \[22:47\]
+
+**Jerod Santo:** Well, to just spoil the end of the story - it's not the end, but it's further down - you did end up designing something that's a thousand times faster... So I think you said three orders of magnitude. That's a big win. But let's talk about getting there. You decided "Okay, what we want is a database that's designed from the ground up for this kind of transaction." Then what do you do? You didn't wanna make a business out of it here. You have a business now, but did you decide "I'm gonna code up in my free time an open source database for the world", or "I need to go raise money"? Where'd you go from there?
+
+**Joran Dirk Greef:** Yeah, so I didn't have really any of those thoughts. I think the first thought I had was I figured we could fix the interface of the database instead of having row locks and multiple queries. We pack a lot of debit credits, first class, one database query back again, and you're done. You actually do get that 1000x performance, and it isn't -- you know, you have to only improve on 76 logical finished product TPS.
+
+I think people sometimes get confused. They think we're talking row inserts per second. We're actually talking logical transactions, debit/credit with the contention row lock problem. If a database is only able to do 100 to 1,000 TPS, depending on the latency of the network and the contention, those are the two variables. You can plug them into Amdahl's law. Typically, if there's a round trip time 10 milliseconds, looking at 100 TPS, 50% contention, looking at 200 TPS; that's your max general purpose limit. And so if you wanna go 1,000 times faster, but you are now packing 10,000 transactions in one database round trip, it's only one meg of information. You're amortizing so many costs. It is not actually hard to do 100,000 a second, 200 -- we actually did a million a second with primary indexes only. TigerBeetle ships with 20 secondary indexes turned on. So it does between 100,000 to 300,000 a second. So it actually is -- I hate benchmark wars. That's why we never really do it. We just try to explain to people first principles, why it makes sense.
+
+But that was sort of the first step. Like, that was my first feeling. And we only thought that, the 1,000x, because that was what --
+
+**Jerod Santo:** Who's the we? You keep saying we, but you haven't described your partners yet.
+
+**Joran Dirk Greef:** Okay. Yeah, yeah. So actually, it was me, and then Don Changfoot was working with me, guiding me around the system. I was doing the work, and then Don was helping me, and he was like my bridge and manager. There were also other people at Coil, my managers there, and surrounding team... Really bright people. My manager was the co-chair of the W3C Web Payments Group. He was the co-chair, but there was Visa and MasterCard and all of them also on that. So it was people who really understood payments. And yeah, so I say we, because I always like to include the team.
+
+**Jerod Santo:** Sure.
+
+**Joran Dirk Greef:** So today there's a team... But I did just create it, and it came out of this experience. So the first feeling was just "We need to go 1,000 times faster, because that is gonna drive down costs, and because the scale is actually 1,000x more." Countries that deploy these kinds of systems, you tend to get very small value transactions and massive volume. In India, for example, they've done four orders of magnitude scale in 10 years only. So in other words, 10 years ago, if you picked MySQL today, you need to -- you're telling MySQL "Go 10,000 times faster, please." There's no cloud database designed, you know... Most cloud databases were designed 2012, or 2015... None of them were designed for like four order magnitude increases. So there's no OLTP database on Earth that can power India or like Brazil's \[unintelligible 00:29:32.04\] These systems are already past three orders.
+
+So we were like "Okay, we need at least 1,000x", because it's not -- this is something I think people also...
+
+**Jerod Santo:** \[laughs\] "It's not as impressive as I'm impressed by it", is what you're saying. Like, there's people who've done 10,000, you're doing 1,000... I'm impressed, but India is not impressed, for instance.
+
+**Joran Dirk Greef:** Yes. Well, you could actually take TigerBeetle today, run it on your laptop and your laptop could power India's transactions volume going through \[unintelligible 00:30:00.04\]
+
+**Jerod Santo:** Wow.
+
+**Joran Dirk Greef:** \[30:03\] You really could, and it would be pretty easy. So it resets the experience... Which I think makes sense, because we've got 30 years of hardware and software research advance; how you could build a database today, a lot has changed. So if you just work with the grain... You know, in those days it was spinning disk; that was the bottleneck. Today it's memory bandwidth. So all the parameters of the design have been inverted.
+
+**Jerod Santo:** Right.
+
+**Joran Dirk Greef:** But again, it's actually not that hard, because there's just so much seismic differential waiting to be just -- I was just excited that we could now do something.
+
+**Jerod Santo:** Yeah. Because all the major ones, the general-purpose ones, MySQL, Postgres, SQLite etc. they're 30 years old. I mean, they're '90s, sometimes in certain cases '80s... And so they were built with a different world of constraints. And they've shown amazing malleability and resilience, and the fact that they're still good enough for lots of things today is an amazing feat of engineering by all those teams. But to start fresh, and to start with a - is it a limited domain? I mean, you're not also bolting on general purpose on top of TigerBeetle. You're saying "Use TigerBeetle for these transactions, and then also use something else for the rest of your workloads"?
+
+**Joran Dirk Greef:** And that's quite right, yes. And I agree with you, Jerod. It's so nice that you said, because I was wanting to say it, too...
+
+**Jerod Santo:** Oh, yeah.
+
+**Joran Dirk Greef:** It's a testament to the designers that these -- you know, Postgres powers the world of general purpose transactions, or if you need to build an app, whatever... But I think the interesting thing here, to your point about fixed domain - for many years what we call OLGP, General Purpose Database, that was used for OLAP, up until '93-'96. And then OLAP actually became a term. Online Analytics Processing. And then, because there was this whole paradigm shift, that "Look, Postgres is great. It's row major." Inside the elephant anatomy, that's how it stores the information. It's sort of designed for a 50/50 read/write split. And so it has like a B-tree that's maybe a little bit better for OLAP... But the OLAP people came along and said "No, no, we need to go column major." It's like a paradigm shift.
+
+**Jerod Santo:** Right.
+
+**Joran Dirk Greef:** So the anatomy of DuckDB is totally different to the elephant. And today we wouldn't say that Postgres is an OLAP database. It's a general purpose database. It's not Snowflake, or BigQuery.
+
+**Jerod Santo:** Sure.
+
+**Joran Dirk Greef:** And that's fine. And vice versa. The OLAP is not always OLGP. But I think today, for the same reason, because of increasing scale, and because you can specialize, there's like a paradigm shift... OLTP is not just rows and columns, row major or column major; it's multi-row major. You're doing things across rows with contention. So the concurrency control of the database inside looks totally different. The anatomy of a TigerBeetle, I would say, if I may...
+
+**Jerod Santo:** Sure.
+
+**Joran Dirk Greef:** ...it's different. It's a lot similar to Postgres, but it's also -- you can think of like the group commit that MySQL or Postgres has, and just dial it up 10X, make a much bigger group commit, much bigger batches. Also, again, like you said, fixed schema. So you don't need all the overhead of serialization, there's no inter-process communication... Just so many things.
+
+It was just incredibly fun to design TigerBeetle, because it was such a simple problem. Just debit credit at scale. So in other words, I think this is it - TigerBeetle is not an OLAP database. It's also not a general purpose database. It's just an OLTP database, and that's all, you know?
+
+**Jerod Santo:** Right.
+
+**Joran Dirk Greef:** \[34:13\] And then Postgres is fantastic. So you never replace Postgres, just like if you're using OLAP you don't. But what you would do is OLGP is like your control plane in your stack. So you put all your entity information. People call it master data, or reference data. So it's the information -- you know, your users table, your usernames and addresses. If you're building Earth's biggest bookstore - usernames and addresses, names of your catalog, your book titles... Those are not really OLTP problems, because that's just... You know, you update your top 10 every now and then, or... You know what I mean. OLTP is like when people move a book into the shopping cart, because that's adjusting inventory that's held potentially for a shopping cart. After a certain amount of time, that debit/credit times out and rolls back. Then if people do check out their shopping cart, those goods and services are moving through logistics, supply, all of that, to warehouses, and delivery... That's all debit credit. Quantity is moving, and it's moving from one warehouse to another, to a driver, to the home, back again, okay, back again... All debit/credit. And then you've also got the checkout transaction with the money, that's also debit/credit. And so like that would be OLTP, and that's sort of the Black Friday problem. The Black Friday problem isn't "How do we store the book catalog, or update that?", because it doesn't change often. So that's a great general purpose problem. Just like users - they don't change the names often. So your database - that's great for variable length information... That information is actually very different to transactional information. Transactional information is very boring, essentially; just multi-row debit credit.
+
+**Jerod Santo:** Right. And so this multi-row major that you describe...
+
+**Joran Dirk Greef:** I made it up, yeah. I would always say multi-row, but let's call it multi-row major. \[laughs\]
+
+**Jerod Santo:** Yeah, multi-row major, versus row major or column major.
+
+**Joran Dirk Greef:** Yeah.
+
+**Jerod Santo:** That makes sense to me, because every single transaction has -- you're gonna assume that over here there's an addition, over there there's a subtraction. There is this double entry thing where there's gonna be more than one row in pretty much anything that matters in TigerBeetle, right? And so that's an interesting way to think about it, and fundamentally a different way, like you said, versus thinking about it column or based on rows. How does that fundamental primitive manifest itself in your decision-making? I assume there's storage concerns, maybe like memory allocation, maybe there's protocol... I don't know where all that works its way out as you design this system. How does that affect everything that TigerBeetle is?
+
+**Joran Dirk Greef:** Great. I'm so excited to dive in. \[laughs\]
+
+**Jerod Santo:** Please do. Let's do it.
+
+**Joran Dirk Greef:** Yeah, yeah. So let's apply this. Let's take the concurrency control, for example. So let's say we've got 8,000 debit/credits. So one debit credit would be like take something from Alice and give it to Bob. Then take something from Charlie and give it to Alice. Take something from Bob and give it to Charlie. And you've got 8,000 of these. Some of them might be contingent on the other, and you can actually express a few things around this... But let's just leave it like that. So you've got 8,000 debit/credits in one query. So the first thing - that comes off the wire into the database. The database is gonna write that to the log, and it's gonna then call fsync. What's great there is you've called fsync once for 8,000... So it's fsync once for one query, but that is amortized across 8,000 logical transactions. And fsync usually has a fixed cost. So it has a variable cost, but there's also always a fixed cost, and it's like half a millisecond, or a millisecond, or... It depends all on the hardware, but there's always a fixed cost. But now you've amortized that massively, over 8,000... Typically, the group commit for MySQL might be around 15 things. So it's much smaller. It will amortize fsync, but not by so many orders. That's the first thing.
+
+\[38:32\] Now we've got the D in ACID. You know, atomicity, consistency, isolation, durability. Before the database processes it, it makes the request durable on disk, and calls fsync. The next thing it does is it'll take these 8,000 and apply it to the tables. It'll actually update the state on disk. And if it was something like a general purpose database, what it'll do is it'll take the first debit/credit, it'll read Alice's row and Bob's row for their accounts, then it will update the rows, and then it will write them back. Then it'll go on to the next one, read the two rows, update them, write them back. So on and so on. So you're looking at about 16,000 accounts that you read in and write out, and that typically takes what is called latches; little internal row locks also, inside the database.
+
+**Jerod Santo:** Okay...
+
+**Joran Dirk Greef:** So 16,000 little micro cuts; you're also contending there. And then -- but you see, here's the catch, is that the domain is usually... Again, everybody's buying NVIDIA. So if like 80% of your 8,000 are all NVIDIA, you're reading NVIDIA, writing NVIDIA; reading NVIDIA, writing -- latching NVIDIA, latching it. And you're doing it 16,000 times. And so what TigerBeetle does instead - and this is, again, with anatomy changes. Now, we first look through all 8,000, and we prefetch all the data dependencies in that request. So all the accounts, for example. And so we load NVIDIA once. Then we load the other six hot super-stocks, and then there's a long tail that we load... But actually, there's not many accounts, and they're usually hot in L1 cache, so they don't even go to disk, because we've got a specialized cache just for these accounts. Everything in TigerBeetle is CPU cacheline-aligned. And we think of that these days as like 128 bytes. Think cachelines are getting bigger. M1 started there. And everything is cacheline-aligned. We don't straddle cachelines for false sharing. Everything is zero copy deserialization, fixed size, very strict alignment. Powers of two, and we don't split machine words, or just... It doesn't always make a difference on the hardware, but it can. So these are all the little edges. The cache is optimized...
+
+But okay, let's go back. So now we load just NVIDIA, the super-stocks and a long tail, but actually they're all in L1 or in L2... And then we've got the data dependencies cached. Then we push all 8,000 through, and then we write them back. And so it's just -- I think you've got it now. It's just drastically simpler. It's kind of like -- you just do less, and that's how you go faster, because you're doing so much less. You don't have SQL string passing.
+
+**Jerod Santo:** Right. It almost feels like you're cheating, but you're not. You're just doing exactly what needs to be done and nothing more, because you're not general purpose, right?
+
+**Joran Dirk Greef:** \[41:45\] Yeah, exactly. I often like to say we didn't do anything special. It's kind of embarrassing. It's so simple. The 1000x trick, it's -- yes, we use \[unintelligible 00:41:54.03\] we do direct IO, DMA to disk, zero copy... All the stuff. Actually, it doesn't make a performance difference. It makes -- it does, that's why we do it. It makes a 1%, 5%... And it all adds up. But that wouldn't get you 1000x. Just like stored procedures in a general purpose also wouldn't get you 1000x. Stored procedures will get you -- you get those 10 SQL queries down to one. But now you're still doing one for one, and you've still got -- so you only went 10x faster or 10x cheaper. If you really want to go 1000x, you actually have to just have first-class debit/credit in the network interface, change the concurrency control... We even -- TigerBeetle has its own LSM storage engine, LSM tree. We designed it from first principles, again, just for OLTP. So it's actually -- it's an LSM forest. We have an LSM tree for every object that the database stores. So transfers and accounts. Accounts and transfers between them; debit/credit transfers between accounts. Those are two trees. And then all the secondary indexes around each object is like 10 trees, and 10 trees... So for every size key and every size value, it's in a separate tree.
+
+And again, there's just things you can do now. RocksDB or LevelDB, which is what you find in a lot of general purpose databases - they use length prefixes for the key. So it's a four-byte length prefix... Or it's variable length, but now you've got the CPU branching costs etc. And then there's, again, a length prefix for the value. But if you know that your secondary indexes are only 8 to 16 bytes, and you then add the cost of length prefixes, four plus four is eight... You know, eight bytes of overhead just to store 16 bytes of actual user data - you're burning a lot of memory bandwidth, a lot of disk bandwidth, and you're going slower than a database that doesn't have length prefixes at all. And so TigerBeetle -- each LSM tree stores pure user data. Literally, we put the key and the value; there is no length prefix, because each tree knows at comp time what the size of the key is. So yeah, we can go on and on, but... And then the consensus protocol - we did also similar optimizations. Yeah.
+
+**Jerod Santo:** So hyper-tuned for this specific type of workload... Which also happens to be one of the most important workloads as well. So let's imagine that I'm an e-commerce shop, and I'm not gonna roll out Shopify. I'm gonna do it myself, because it makes sense. And so far -- I'm a Postgres guy, and personally I respect MySQL, but I just use Postgres as my example, because that's what I've been using for 15 years. So let's say that I've got my Postgres database, it's been running everything just fine, it's got my users in there, it also has all my transactions... And I'm hitting up against scale issues. This is a good problem for me, because it means I'm doing more sales. And so I'm selling a lot of books, and I'm hitting scale issues, and someone says "You should really look into TigerBeetle for your transactions specifically." And I think to myself "Postgres hasn't failed me for 30 years." TigerBeetle hasn't existed until 2020 at the earliest. You can tell me when it's 1.0, or when you guys actually shipped a product... Because conception to now, we're five years in.
+
+**Joran Dirk Greef:** Yes.
+
+**Jerod Santo:** I'm gonna trust my most precious -- my sales. I'm gonna trust that, to something that's this new. I'm sure you face this a lot, as you go and try to get people to try out TigerBeetle. What's your response to that concern? Because it's a valid concern. Not super -- what's it called? ...battle-hardened yet. Or maybe it is. Tell me about that.
+
+**Joran Dirk Greef:** \[46:02\] Yeah, great. I love the question. That was my second thought, as TigerBeetle was created. First thought is --
+
+**Jerod Santo:** Yeah, "No one's gonna trust us. Why are they gonna trust us?"
+
+**Joran Dirk Greef:** Yeah. And the second thought was this question, "How can you possibly be as safe as 30-year-old software that was created Windows 95? How could you possibly be as safe?" And I think the answer to that is actually a question. In the same way that "How could you be 1,000 times faster than something that is 30 years old?", the question is "What has changed in the world in the last 30 years?" So much has changed from a performance perspective... And then when we look to safety, and especially mission-critical safety, so much has changed.
+
+So 30 years ago, consensus didn't really exist in industry. Brian Oki had pioneered consensus a year before Paxos View-stamped Replication in '88. That was his thesis at MIT, with Barbara Liskov. The consensus did exist... You know, how can you replicate data across data centers for durability, to actually survive loss of a machine or disk? But that wasn't really an industry 30 years ago, so you don't really have first-class replication. Yes, you do have it these days, but it wasn't designed in from the start. And I think there's more examples around testing. So deterministic simulation testing, what FoundationDB did - and we actually got it from the people at Dropbox, not Foundation. James Cowling and them at Dropbox were also doing DST.
+
+The idea is like you design your whole software system that it executes deterministically. So you don't use -- if you use a hash table, the hash table, given the same inputs, always gets the same layout. You don't use randomness in the logical parts that users would see. you can think of it basically like fuzzing your software or property-based testing. Given the same inputs, you get the same outputs, and you design all your software like this. Then you can test it. But if the test fails, when you replay it, you'll get the same result. And so distributed systems were so hard to build 30 years ago, so not many people did. And then when they started really building them, they actually started with eventual consistency, because people were still figuring out how to do strict serializability. So there was a lot of fashion around eventual consistency, which has gone away, I think... But back then, to build a distributed system you just kind of assumed it'll just be eventual.
+
+So how to build distributed systems was quite hard. And it was hard because it was hard to test them, because the failure of one system over there causes the failure of another system here. And you can never -- when you find a rare bug, you can never replay it, and you need so many machines... And these bugs -- you know, before TigerBeetle, I worked on systems that were distributed, like full duplex file sync, hence my interest in Dropbox. But those systems were incredibly hard to test. There would be bugs, you know they're there, they're taking a year to find and fix. And then you realize it's a buffer overflow in the libuv. That was some of my first C code that I ever wrote, years before TigerBeetle. It was fixing a buffer overflow in libuv. It was the Windows event notification, some interaction with multi-byte UTF-8, and different normalizations of NFC, NFD. That was this distributed system's bug. And it also needed long file paths. So it took a year -- like, we knew it was there...
+
+**Jerod Santo:** Was that your Node.js days, or that was prior to Node.js days?
+
+**Joran Dirk Greef:** Oh, that was Node.js days, yeah. So around 2016. And we were using Jepsen style techniques, where you've got fault injection... Like, chaos engineering. That's what we were doing. So you could get the bug, but to even -- you could get this amazing bug, you knew it was there... It was like the tease. And you could never find it, reproduce it, or fix it.
+
+**Jerod Santo:** \[50:24\] Ugh. The worst.
+
+**Joran Dirk Greef:** And it literally took a year. So coming to TigerBeetle, I knew from Dropbox there were newer ways to build distributed systems. Just like you don't need to use eventual consistency anymore, there's proper consensus, no one should give that up lightly... You don't need to, because you can get great performance first, and so much performance fundamentally, you fix the design, and then you can add consensus. It's not expensive when the design is right. I mean, consensus is literally just, you append to the log in Fsync, and in parallel, you send it across the network to a backup machine in Fsync. That's replication, 99% of the time. And then consensus does the automated failover of the backups, the primary.
+
+So consensus doesn't really have a cost. That's a common myth. All these things have changed... But coming back to it, testing has changed, and how you test distributed systems. Because you can actually model a whole cluster of machines in a logical simulation that's deterministic. Now, you test it just like Jepsen, but the whole universe is under deterministic control. You can even speed up time. So you can like fast-forward the movie to see if there's a happy ending. If it does, great. Watch another movie. Each movie is like a seed; you know, how you can generate a Worms level. You know the Worms game, or scorched earth games would be like randomly generated... All from a seed.
+
+**Jerod Santo:** Oh, yeah.
+
+**Joran Dirk Greef:** And so this is just classic fuzzing, property-based testing with seeds, but you're taking these distributed systems and they're all -- the database was born to run in a flight simulator. It's deterministic. And if you can do that, you can build systems that -- you're kind of doing Jepsen, but you're also speeding up time. You can reproduce.
+
+**Jerod Santo:** Right.
+
+**Joran Dirk Greef:** And then once you've had that experience, you have to ask now "Do the 30-year-old systems have this level of testing?" Yes, they have 30 years of testing, but with DST, you can speed up time. So we've actually worked it out in TigerBeetle, one tick of our event loop is typically 10 milliseconds. We collapse that into a while true loop. We get a factor of 700X speed up when you take into account the CPU overheads of just executing our protocols. So we flatten everything and we execute it, and you get a 700X time acceleration. So we have a fleet of a thousand CPU cores dedicated to TigerBeetle. We got them from Hetzner. Thank you very much, Hetzner.
+
+**Jerod Santo:** Nice.
+
+**Joran Dirk Greef:** They're in Finland, so nice and cool... They're burning clean energy.
+
+**Jerod Santo:** There you go.
+
+**Joran Dirk Greef:** And a thousand CPU core fleet, they run DST 24/7... And that adds up. I mean, it is roughly 700X. It's a thousand cores, because we pay for it, dedicated. And they run -- we do a lot of work to like optimize how much we're using those cores... But it does add up to -- on the order of like 2000 years a day of test time. And but you see, now again we're simulating like disk corruption, and we're simulating things like we write to the disk, and the disk says "Yes, I have fsynced, but the firmware didn't." Or we write to the disk and the disk writes it to the wrong place. So TigerBeetle has like an explicit storage fault model. So we do assume that disks will break, but they don't only fail stop. They actually are like kind of -- we call them near Byzantine. So disks do - very rarely, but 1% in a year or two, a disk will have some kind of corruption or latent sector error. Then you get a latent sector error... A little bit less common is like silent bit rot, a little bit less is misdirected IO, where it actually just writes to the wrong place... And this can be the hardware, or the disk firmware... Even the file system.
+
+\[54:30\] So two years ago XFS actually had a misdirected write bug. And if your database was running on that particular version of XFS and you triggered that, your database would write to the wrong location on disk. And now the question is, "Well, who tests for this?" And you almost can't, unless you're using these new techniques.
+
+**Jerod Santo:** Yeah.
+
+**Joran Dirk Greef:** So I don't know. I mean -- yeah...
+
+**Jerod Santo:** It's cool.
+
+**Joran Dirk Greef:** I guess the question was it's not enough actually to be as safe as what was safe 30 years ago. We've got new techniques, and there's a few more of them in TigerBeetle.
+
+**Jerod Santo:** Yeah. No, I think that's super-cool. It reminds me of light bulbs, for some reason. LED light bulbs - they say they're supposed to last 25 years, or something. And I'm always like "You don't know that, because you haven't been using them for 25 years." In fact, the house that I currently live in, we're going on 10 years now, and they sold us on all LED light bulbs... And I remember as the installer was putting them in, he's like "You're never going to have to replace any of these." And you know what? I've replaced a whole bunch of them. So whoever did their testing can't do what you guys can do, which is - they can't just fast-forward time and prove that this thing's going to last for 25 years, because they haven't been building these for 25 years. But what's super-cool with this -- what's it called, deterministic simulation testing?
+
+**Joran Dirk Greef:** Simulation testing, yes.
+
+**Jerod Santo:** Yeah. What's cool about it is you guys can actually, just through CPU power and design, you can actually simulate all this time. And so you can claim, even though you've only been around for five years - I'm giving you all five, even though I'm sure it's technically less than that - that you've actually tested for hundreds of years, right? Like, you could just say that because you've done that work through this three-dimensional simulation that I'm imagining that you put the system through at all times. I think that's pretty cool.
+
+**Joran Dirk Greef:** Yes. At least we're trying to get there, you know? So we would also say it's only as good as our coverage, and our combinations of coverage, and our state space exploration... But we invest so much into that as well, you know? And so it does -- because we also know how many bugs we find, and how rare they are. If you think of it like a hacker, they require like eight exploits to be chained, and each exploits is tough. And you just know that traditional software - there must be so many millions of bugs. But yet we don't find them in the real world, because they're so, so rare... But with the DST you do actually, and we find them pretty quickly. So yeah, so we wouldn't claim, but it's very strong.
+
+**Jerod Santo:** It gives you some confidence that you wouldn't have otherwise. And I think that's very valuable. But like you said, or maybe you didn't say it, but I was thinking it while you said it - there's no accounting for the real world. So like Mike Tyson's famous statement, "Everybody has a plan till they get punched in the mouth", or something like that.
+
+**Joran Dirk Greef:** \[laughs\] Which is how it works, right?
+
+**Jerod Santo:** You can simulate all you want, and it's amazing what y'all are doing, but then there's the actual reality, and there's always gonna be something. And so - is TigerBeetle out there? Is it in production, in reality yet? Are people using it? What's the state of that side of it?
+
+**Joran Dirk Greef:** \[57:45\] It is, very much so. So we had a customer reach out to us, they needed to -- I mean, we've got a few customers, but just an example of one... End of last year they reached out, they said "Look, some regulations are changing. They need massive throughput, because it can put their business ahead." We had to get them -- for like sheer business strategy, they needed to migrate within 45 days a workload of like 100 million transactions a month. Logical. 100 million logical transactions a month. They needed to migrate within 45 days from the old system. We migrated them, and they saturated their central bank limit, and they were happy, you know?
+
+**Jerod Santo:** Nice. Yeah.
+
+**Joran Dirk Greef:** And we pulled it off, and the system just works.
+
+**Jerod Santo:** That's gotta feel good.
+
+**Joran Dirk Greef:** It did, yeah. That was great, you know? And there's national projects, three different countries, whether it's for the whole national transportation system, or the central bank digital currency, or another central bank exchange... TigerBeetle is going into the current production version of that gate switch now, as we speak. But I think, just to go back, Jerod, to the DST - it's a lot like formal methods. The difference is that formal methods check the design of a protocol, so that you know that the protocol could possibly work. It's formally proven that it could work. But for me, always the challenge was, "Well, how do I know that I've just --" Because I'm not a -- I always feel like I'm getting slower every day. How do I know that I coded this correctly? I know the protocol's correct.
+
+**Jerod Santo:** Right. Implementation.
+
+**Joran Dirk Greef:** I know \[unintelligible 00:59:25.07\] replication, Paxos, Raft - they're formally proven, but the implementation is like thousands of lines of code. And so how do you check that? And so DST is actually checking the actual code, and the simulator is checking for split brain, it's checking linearizability, strictizability, of which linearizability is a part... And it's even checking things like fsync gate. So cache coherency - TigerBeetle's user space page cache, is it coherent with the disk at all times? Even if there's Fsync IO faults. So Postgres - they are fixing this, they're adding direct IO. If people want to find out, it's called Fsync gate, 2018, but most databases still can't survive fsync gate. They were actually patches where databases patched like MySQL etc. to panic. The problem is when they start up, they still read the log from the kernel page cache, which is no longer coherent. So actually they have to use direct IO. So Postgres has been on a long journey, laudably, to add async IO, direct IO... It's in -- I'm not sure yet, it might already be in as the default... But those are kinds of the things you need to survive. And that isn't even an explicit storage fault model, but TigerBeetle simulator is actually checking. So your simulator can reach in and check so many invariants.
+
+But then also, I think back to what we were saying about claims and coverage, you wanna have a very buggy view of the world. So you take your four primary colors, network, storage, memory, compute, you assign explicit fault models. So compute, we say "Look, that would be in the domain of Byzantine fault tolerance, so we don't solve that." And that's explicit. Memory would be ECC; that's our fault model. And then what TigerBeetle focuses on is the network fault model, so packets are lost, reordered, replayed, misdirected. Classic Jepsen; partitions, all of that. That's what makes consensus so hard, is just solving the network fault model is almost impossible. Then TigerBeetle adds the storage fault model explicit. So you write to the disk, wrong place, forgotten, you read from the disk -- you're actually reading from the wrong place. So you'll get a sector as a valid checksum for the database, but it's the wrong -- now you need to daisychain your checksums.
+
+And then what we do is with these -- this is sort of the buggiest view of these four. Okay, the first two, we've been explicit that we don't solve those, because they have different levels of probability. The probability of a CPU fault is astronomically more rare than a memory fault, which in turn is astronomically more rare than network fault. Maybe it's the other way around. Sorry --
+
+**Jerod Santo:** \[01:02:17.11\] Sure.
+
+**Joran Dirk Greef:** The rarest thing is the CPU, then it's the memory, then it's the storage, and then it's the network. So most people are just solving network, TigerBeetle solves storage. But storage is when -- at scale, it happens more and more. So 1% of disks around a two-year window.
+
+**Jerod Santo:** At enough scale, you're gonna hit that enough times that it seriously matters, versus at small scale where you're like "Ah, we don't ever have to worry about that." Yeah, that probability thing is one of the reasons why I think logically as humans we fail often to reasonably consider scenarios... Because we think about worst case scenarios, but we don't pair that with probability of scenarios. And so it's like "What's the worst that could happen?" And for some reason - this is like off topic, but in our hearts we give that like 100% chance, right? Like, if that happens - well, let's assume it does happen. So 100% chance on that. Then it's terrible. But we don't also think like "Pair that with probability." So that's kind of what you're saying with the CPU stuff. It's like, the odds of a CPU failure - okay, catastrophic, of course, but exceedingly rare, versus the likelihood of a drive failure, for instance.
+
+**Joran Dirk Greef:** Exactly. Drive, or network... Yeah.
+
+**Jerod Santo:** Right. Network is probably the most unreliable at this point, I would assume, since our drives don't spin anymore...
+
+**Joran Dirk Greef:** That's it, yeah.
+
+**Jerod Santo:** Yeah, the network is incredibly unreliable, unfortunately.
+
+**Break**: \[01:03:46.17\]
+
+**Jerod Santo:** Well, while we're talking about the simulation stuff, I do want to take a quick aside and talk about something I've found that was really cool on your website. SimTigerBeetle. sim.tigerbeetle.com. This is a simulator for distributed database scenarios. It's like a game... Well, for the video folks, we'll put it up for you guys to look at. This is really cool. You can play these different games, like the Mexican standoff, the prime time, and the radioactive hard drive.
+
+What's going on with this? This is like out of left field when I saw it. This is not -- I wouldn't expect this from you, Joran. What is this thing?
+
+**Joran Dirk Greef:** I thought you didn't expect it, Jerod. Yeah.
+
+**Jerod Santo:** I didn't expect it. I saw it, I'm like "Okay..." I started playing it, and it's like, it's polished, the graphics are awesome, there's sound, there's music... And it's like a game, all about distributed system failures. Tell me more.
+
+**Joran Dirk Greef:** Maybe our team, we just enjoy TigerBeetle too much. \[laughter\] You know, it kept me company in lockdown, in COVID. That was July 2020. March 2020 I started on the Switch. It was my birthday, you know? And a week later, the world went into lockdown. And I was locked down solving this problem. July, 2020, I sketched TigerBeetle, created the prototype, many versions, and never stopped.
+
+I don't know, maybe it's because it's out of that experience that we, as a team, we have so much fun or we really enjoy it... But some TigerBeetle is kind of an expression of that, because the feeling of our DST simulator -- the name of it is the VOPR, Viewstamped Operation Replicator. It's a homage or ode to war games. It's got the WOPR, that classic simulator in war games. And so TigerBeetle's simulator is called the VOPR.
+
+**Jerod Santo:** Nice.
+
+**Joran Dirk Greef:** But the VOPR for us was such an experience, you know? Switching it on for the first time... I was listening to Kings of Leon, Crawl as I switched it on.
+
+**Jerod Santo:** \[laughs\]
+
+**Joran Dirk Greef:** And we had been building TigerBeetle for a year, with the whole design, all careful default models... We had the interfaces designed... I knew we would do DST, and I planned it like that... A year in, it took about a week to build the first version of the simulator and then we switched it on. And just the bugs came falling out of the trees...
+
+**Jerod Santo:** I bet.
+
+**Joran Dirk Greef:** ...and you'd fix like five a day... And each of these was like one of those one-year bugs from before. Now you're fixing like five a day. And that experience - oh, it was just special as a programmer. I wish it for everyone. And I think people are getting into this style. But this runs in our terminal, so I did a demo to a friend for the very first time, Dominik Tornow - fantastic - of Resonate. I really look up to him in distributed systems. He's become a mentor and friend. And if I have a hard problem in distributed systems, I'll ask Dominic. But I had just met him, and I showed him our simulator running in the terminal, and I'd never shown anyone before. And he was like "Wow..." He'd done formal methods. He's like "This is formal methods like on the code." And I showed him the probabilities of the fault models for each simulation run, and he was blown away. He said, "No, you've got to tell people about this." And then I thought, "Well, how do you show this to people?" And I'm a dad, and I thought "How do I encourage my daughter?" Not encourage her, but just -- you know, how do you encourage the next generation to get excited about computer science?
+
+**Jerod Santo:** Yeah, intrigued.
+
+**Joran Dirk Greef:** Yeah. And to me, this was the most magical part of programming in my own journey. So I thought, "Well, let's make a game. Let's take our simulator and let's put a graphical frontend on top to hook into all these events. And then we can create different levels for people, showing them how consensus works." If there are no network failures, so everything's perfect, you'll simulate the network latencies and disks... Everything's simulated, the clocks, all of that, but there are no faults. And so it's perfect. And now you can actually just teach just normal replication through the consensus protocol. And then the next level is like, okay, now we're gonna introduce probabilities of partitions, and network faults, but the disk is perfect. And then the next level is, "Okay, disk is radioactive."
+
+**Jerod Santo:** Yeah. Well, I played it for 5 or 10 minutes and I had a blast. It's like "Here's a hammer. You can start hammering stuff and see how it reroutes", and... Really, really cool.
+
+**Joran Dirk Greef:** Yeah. And you know, each of those beetles that you hit, they're running -- each of them is running real TigerBeetle code, compiled to WASM.
+
+**Jerod Santo:** Really?
+
+**Joran Dirk Greef:** \[01:12:08.07\] Yes. It's the real code in your browser, for the cluster. When you take a hammer and hit the beetle, you're actually physically crashing the machine, and it's restarting. And you're actually getting to touch a simulated world, but against the real code, as a human.
+
+**Jerod Santo:** That's amazing.
+
+**Joran Dirk Greef:** Yeah.
+
+**Jerod Santo:** I feel like, as a TigerBeetle engineer, you could just be playing that game, and Joran asks you what you're up to and you're just like "I'm working", you know? "I'm simulating some crashes here... Come on."
+
+**Joran Dirk Greef:** Yeah. Well, I have done that... I do do that, Jerod.
+
+**Jerod Santo:** \[laughs\]
+
+**Joran Dirk Greef:** Sometimes, instead of doom scrolling, you just sim scroll.
+
+**Jerod Santo:** Yeah, you might as well.
+
+**Joran Dirk Greef:** My daughter says to me "Papa, can we play the TigerBeetle game?" And then -- but I'm glad you call it a game, because we meant it like in the walking sim genre. So it's a game that you can't win, because no matter what you do, things recover.
+
+**Jerod Santo:** Right. You can't knock the whole system out. It's going to go back to good. It's going to find its way back.
+
+**Joran Dirk Greef:** Yes. You actually can, if you're really lucky. When I do it in live presentations, it's never failed me. But theoretically, you could crash the cluster, because you see, those human tools allow you to inject more faults than the \[unintelligible 01:13:30.08\] tolerance. And then the system is designed to shut down safely. So you might run into that, but you would have to be very quick.
+
+**Jerod Santo:** Yeah. Well, I was too slow in my five minutes of playing around. I wasn't quite fast enough.
+
+**Joran Dirk Greef:** I'm also. Always.
+
+**Jerod Santo:** But I didn't know it was possible. So now that I know it's possible, maybe I'll commit myself to shutting that system down. Pretty cool.
+
+**Joran Dirk Greef:** Oh, but there is a game, Jerod, at the end. I don't know if you've played the credits in Radioactive.
+
+**Jerod Santo:** I have not. No. Is this an Easter egg?
+
+**Joran Dirk Greef:** Yeah. After Radioactive, at the end, the credits go, and that becomes like a little platform jumper game...
+
+**Jerod Santo:** Oh, man...
+
+**Joran Dirk Greef:** And you can jump and spin a beetle, and you do get a score. That one is quite hard.
+
+**Jerod Santo:** That sounds amazing. I'm going to go do that after we hang up here. Two more things --
+
+**Joran Dirk Greef:** Maybe I should just -- if I can just quickly add...
+
+**Jerod Santo:** Sure.
+
+**Joran Dirk Greef:** That game, it was created part-time. I met Fabio Arnold in Milan, the very first Zig meetup in Europe. He was there, and he created it part-time; just a few hours every week, with Joey Mach's Illustrator from Rio de Janeiro, Brazil; just the two of them. And they created it like very, very low budget. But they had such passion, and we then carried on and we put more into it, so it did become more polished. But that's just the skill tribute to them.
+
+**Jerod Santo:** Cool. Yeah, I think for those who go out and give it a play, you will notice immediately - and I did - just how much love is put into something like that. You mean it when you say "We just love TigerBeetle", and this whole system because, there's custom sounds, and there's music... This is a labor of love, for sure, and a really cool way to show off what you all have built, in a way that is difficult just with our brains, as you talk about it, for me to map it onto my brain and make sense of it. But when you put it out there in that visual way, it's very compelling. So shout-out to them for their labor, and for you to keep polishing that thing up.
+
+I did want to touch on the open source side, and kind of the business end. You mentioned some customers, you raised some money now, so this is like a serious business, but it also is an open source database... Can you talk about the give and take there, the push and pull, the decision-making process? Because we talked to a lot of people in your circumstance, and some of them have made other decisions, some of them have made the exact same decision you've made... And we're all trying to figure it out, how can we make this thing work. So tell us your side of the open source/VC story.
+
+**Joran Dirk Greef:** \[01:16:09.04\] Yeah, thanks so much for asking that, Jerod. That was my third feeling.
+
+**Jerod Santo:** \[laughs\] I'm hitting all your feelings here.
+
+**Joran Dirk Greef:** My first feeling was -- yeah, yeah, yeah, but really... Like, July 2020, as the project was started, I remember clearly there were three moments. The first moment was like "Wow, this prototype is fast. It's like, the design works... Wow, this maybe could change things for the open source switch." That was it. And maybe it could change things beyond, but I didn't know. The second moment I remember was the DST, switching it on Kings of Leon, Crawl. The third moment was this wondering of -- what actually happened was we designed it to be so much safer that yes, it gained adoption within this Gates Foundation project. We won trust, and they are integrating it today. It'll power countries at the national level, who knows in how many years, but as it gets deployed. So we did solve the trust problem of being so much safer, because we designed it like that as well.
+
+But then the third moment was people then within that project saying "This is all well, but open source is too cheap. Where's the company to support it? Where are the people that are gonna be available to service and really work?" You need open source... So this system is Apache 2 open source, and all the software it uses -- it cannot use software that isn't open source, because otherwise it just would be a non-starter. So TigerBeetle was also created -- like, contracting then, it had to be Apache 2.0. That was obvious to me open source, because otherwise you don't fulfill the mission... Which is what inspired the performance and the safety, is actually to make this really safe, because it is people's money.
+
+So then the question was people were saying, "Where's the company to support this open source? It's too cheap." But I still -- I didn't have a clear vision of like business model at that stage. And then it became clear, as startups said to me "Open source is too expensive." \[laughs\]
+
+**Jerod Santo:** Hmm... Too cheap...?
+
+**Joran Dirk Greef:** So on the one hand, you have countries saying it's too cheap, and then you have startups saying it's too expensive. And I'm like "This is Goldilocks. We're just trying to make some great open source porridge, and it's either too cheap or too expensive." But nobody's saying it's free as in puppies. It's too cheap... And then I realized, okay, that's it. Business model is orthogonal to open source. Business is about trust. People trust you, at the national enterprise, and it's always about trust. That is what you sell as a business, is trust; your brand, your reputation. It's actually -- I use the word "It's brand", and I think startups talk a lot about go to market... I think it's more interesting to talk about brand. Do we all appreciate the value of brand? Because brand is trust. And I must thank my auditing professors. They used to ask us, "What do accountants sell? Trust." That's the only thing we sell, is trust that the numbers are correct.
+
+**Jerod Santo:** Right.
+
+**Joran Dirk Greef:** \[01:19:44.08\] Yeah, so business is about trust, and it's also about value. And so startups - they want a push button experience. "Who will run TigerBeetle for me?" Because that can make it cheaper for me than if I had an engineer do three months of work around the SRE of a database.
+
+**Jerod Santo:** Right.
+
+**Joran Dirk Greef:** So you can actually have a business and sell something that's gonna make something cheaper for startups. And similarly for enterprise, you can have a business and sell something that is going to provide the value they need, which is - now, they might have SRE teams, but they need the infrastructure to support massive scale, like Petabytes. How do you connect TigerBeetle to object storage S3. Like, OLTP data lake; not only OLAP data lake, but let's just connect the OLTP direct to S3.
+
+And so this comes to your question about the tug of war, and licensing, and all of this. And I think the big mistake that we can make - and I used to make this until it became clear for me that third moment - was that an open source license affects the implementation, not the interface. But when it comes to competition in the business world, that doesn't happen at the implementation. Typically, it happens at the interface. So if you think of like some of the fiercest competition, when things were really on a knife's edge for the web, it was the interface, not the browser implementations. It was the interface that the war was fought. Mozilla fought that war, and we needed other browsers to fight it, because the interface was being embraced, extended and extinguished. Triple E. And then you think of like Android, and Java, and again, it wasn't about the Android implementation, it was the interface. And that was that Oracle/Google lawsuit, you know? And then again, you think of like "Well, Confluent --" Kafka's Apache 2.0 open source.
+
+Then Red Panda came along - and I'm a huge fan of Red Panda, because very similar design decisions around being direct with the hardware, static communication... All of this very, very similar. We came from a similar time period, but in that time period the things were changing, how you built things. But Red Panda came along and they saw the open source implementation of Kafka and said, "Well, thank you, but we don't want it. But that interface, that is great. That's where our business will be also. That interface, thank you very much." And so they built a massive high-performance implementation.
+
+And then WarpStream came along and they said, "Well, Red Panda, you are business source license, not open source source available. Confluent or Kafka, you are Apache 2.0 open source." But that's all implementation. I'm gonna do my own implementation, thank you very much, of object storage. But the interface - great. Okay, now we're all competing. And so I think the myth is that source available was kind of the thing that always -- I always feel that something inside of me dies when I see a great company relicense, or when I see a young startup follow that lead. Because to me, source available says that "We think it's gonna stop competition." It doesn't. You may as well be on the beach, building little moats and sandcastles, but innovation technology is like a wave. It'll just find a way around you. It'll WarpStream around you. You can't legislate competition away. And we shouldn't be trying to build companies where we think the success of the company is us creating a monopoly. The world's too big. There's too much for you to be -- you don't need monopolies to do really great. And that doesn't build trust, to say to your customers "You can only buy from me."
+
+So I think people think it stops competition, and they think it helps them sell, and it actually defeats both of those, because you get complacent, and you actually fail to build trust. So you burn trust when you relicense, and if you start source available, you're gonna be doing diligence with enterprising and say "Sorry, you're not open source." It's license confusion, you know. And maybe some people get it, but there's a little 1% headwind. And it's actually - it's a category error, because you're spending so much effort chatting to people about implementation licensing... And the rest of the world is competing on interface. And TigerBeetle's interface is quite simple. Very simple.
+
+\[01:24:30.17\] So we could -- I don't know, whatever license we apply, it doesn't matter. Debit/credit is where we compete. There are companies that offer debit/credit as an API and it's very similar to TigerBeetle. But we compete on trust. We didn't just take a general purpose database and slap on debit/credit. We went deep; we really cared and we built the whole thing.
+
+And people pay us. So before we were a company, we had an energy company reach out and we landed quite a good contract, very quickly. And I think it came down to trust. So open source builds trust. So open source is great for business. It's also orthogonal.
+
+I think the other thing is if you -- so there's a lot of things in TigerBeetle that are like the... I had done many experiments of my own, my passion projects. They're all in TigerBeetle. Many parts of the design of TigerBeetle come from these various experiments that I did. And so I was never going to put that all into a project if it's not open source, because it's just too valuable. I want to always be able to play with it no matter what happens to it. And I think we all feel like that, like our critical infrastructure just has to be open source. So yeah, I think that's kind of how I think of it.
+
+**Jerod Santo:** No, I think that's a great perspective, and a specific explanation that I haven't heard previously, so I definitely appreciate it. I'm mulling on it. I think I agree with most of what you said. And the implementation versus interface dichotomy is one that I hadn't considered that explicitly, and I need to think through it more. So I appreciate your thoughts on the matter.
+
+Question is, what happens? You know, Apache 2.0, your heart and soul is into this thing... What do you do? How do you respond if and when AWS comes by and says "TigerBeetle by AWS, for sale now"? Does that scare you? Does that threaten you? What do you think about that scenario? Because that's what a lot of people are concerned about most specifically.
+
+**Joran Dirk Greef:** Yeah, so I think one can try and stop it. You can build the castles in the sand... Or you can just say "Look, the wave is inevitable. It's coming. Let's prepare for it." So what we do is we get the surfboard ready...
+
+**Jerod Santo:** \[laughs\] And you ride the wave.
+
+**Joran Dirk Greef:** ...and we're on the beach, we're waiting for it to come.
+
+**Jerod Santo:** Okay.
+
+**Joran Dirk Greef:** And actually, we're already paddling up, and there's the swell. The wave will take its time to catch up, but we're already surfing. And so there are gonna be -- we just get into the water and... It's like we could have the cavalry in the castle, or we could get them out into the field and have great cavalry, and great user experience. Like, let's compete on -- let's actually add value and serve the community honorably at a profit. Let AWS catch us in that. And great. If they decide that they couldn't build a debit credit database as well as TigerBeetle so that they offer it as their OLTP flagship database, TigerBeetle - well, that builds trust. A rising tide lifts all boats, and now we're surfing the wave with AWS.
+
+**Jerod Santo:** You're swimming. You're out there, surfing. I love it. I love it. I think it's a great way to think about it. Because like you said, the wave is inevitable, so you might as well prepare for it. You might as well ride it; ride that wave.
+
+**Joran Dirk Greef:** Yeah. And I think also we've seen this play out a few times, that source available doesn't stop at AWS. So if it's valuable enough, they'll write the implementation. If you relicense, they will immediately fork your community. Now you've got two problems. They're still competing, and now they lead the interface. And that's when it's fatal for a company.
+
+\[01:28:22.29\] When I see them relicense and you see the classic AWS, I think "Oh, gee, they did it". Yeah, that's actually the thing you don't want, is when your open source clients are being bought up and being led now. Yeah, but I love AWS. I love their work. I've learned so much from their distributed systems. I've got friends that work there... So really to me that isn't the threat. The threat is -- you know, what's the problem with the world? I am. We are as a team.
+
+**Jerod Santo:** Sure.
+
+**Joran Dirk Greef:** So the threat is really that we stop investing in performance and safety. We stop being trustworthy, building trust.
+
+Maybe we should say Jerod too, is that your product is more than the open source. There's a principle here, too. I think it works both ways. If we connect TigerBeetle to a proprietary API, proprietary interface, our principle as a company is that's viral. So if someone wants to license their interface as proprietary, our connector will be proprietary. For example S3, if we connect to S3 for massive scale, then we charge for that, fairly. We make our honest profits; so serve the community honorably, at a profit. There must be a profit, because there's entropy in the world.
+
+**Jerod Santo:** Sure.
+
+**Joran Dirk Greef:** But if something's open source, we'll then reconnect it. And there's lots of -- just like people would pay Amazon for Aurora, because there's a great value in all the management around Postgres... Again, porridge is too hot or cold. So we'd be curious to get your thoughts once you've thought about it, and I can sharpen the argument... Because I really think we need more founders need to stand up... And we've all been given a lot from the previous generation of open source, and I think it would be great if we all say "Okay, we're gonna pay it forward as well. Make a technical contribution." There's no reason not to. I think it's better for business, actually.
+
+**Jerod Santo:** No, I definitely appreciate everything you said right there. And I will be listening to it back as we produce this and consuming it more. I love the surf the wave analogy. That's where you really sold me. So so far I'm amenable to your argument, but I'm very easily convinced on the air here. Is there anything else TigerBeetle or otherwise that you wanted to touch on, that I haven't brought up yet? It's been a great conversation.
+
+**Joran Dirk Greef:** Yeah, no, I've loved it too. I mean, I guess we should say we wrote it in Zig, this new systems language.
+
+**Jerod Santo:** Oh yes, I didn't bring up the language wars yet. We have to get our clip, because the flame wars must rage on. You wrote it in Zig, I assume from the day one; it was a day one decision. And you are probably happy with it, since you just brought it up now. So Zig for the win, it sounds like. Is that your overall message? You love it?
+
+**Joran Dirk Greef:** Yeah, it was also just a big wave. You could see the swell, and you're like "I'm going for that wave..."
+
+**Jerod Santo:** You're hopping on that wave. Yeah, it's been a good choice so far.
+
+**Joran Dirk Greef:** Yeah. When Node.js came up, I jumped on, because it made sense. And then I was really happy that I did. And when I saw Zig come, I thought "Well, it's not often that you have these moments." Rust was another one. But with TigerBeetle the timing meant that -- you know, we could have caught the rust wave, but there were so many thousands on already. We would have been a drop in the ocean.
+
+\[01:32:06.23\] Also, we wanted to do static memory allocation, and Zig is actually really ideal for that. It's really perfect for the TigerBeetle design. It would have been much harder to do our intrusive memory patterns, and direct hardware access \[unintelligible 01:32:20.25\] zero copy. Two of our team -- one of our team is actually the co-founder of the Rust language, Graydon. He was the project lead, Brian Anderson (brson) at Mozilla. His desk was next to Brendan Eich. And he's writing our Rust client in Zig. Well, in Rust, sorry; he's writing it in Rust.
+
+**Jerod Santo:** Okay. I was like "Wait a second..."
+
+**Joran Dirk Greef:** But he writes Zig normally. But he's writing our Rust client. And then Matklad is the creator of Rust Analyzer and IntelliJ Rust. He joined the company -- he's basically like a co-founder. So there were a few of us... And my senior partner from Coil - not many people know, but he came with, and Matklad, Federico, Rafael... But they sort of are the core team. And Matklad joined -- he wrote a blog post called "Hard Mode Rust", trying to do static allocation, very similar patterns. And then Jamie Brandon introduced us. And Matklad was like "I've been trying to do this in Rust", and they're like "But Tiger-style - our way of coding in TigerBeetle, Tiger-style... We've got our engineering methodology written up."
+
+**Jerod Santo:** Oh, really? Can you link me up to that? We probably don't have time to go into detail, but I'd love to read it. Tiger-style you call this?
+
+**Joran Dirk Greef:** Tiger-style, yeah. And that's sort of all the safety techniques. So assuming that we do still have bugs, we also have like 6,000 assertions that check the operation of TigerBeetle in production. If any one of them trips - it's like 6,000 tripwires - then there's probably maybe a CPU fault, or a memory fault, or a bug. And then we immediately shut down the whole database. You wanna operate safely, or not at all. And that way you get total quality control, and the system becomes safer and safer. You don't have latent bugs. Yeah, so Zig really suited not only the performance, but also some of the safety decisions.
+
+**Jerod Santo:** Right. It wasn't merely the trend. There's also technical reasons that you chose it.
+
+**Joran Dirk Greef:** Yeah, I think we picked Zig before Bun. The only other major Zig project at the time was River, by Isaac Freund, a Wayland compositor. Amazing, amazing programmer, Isaac. He contracted on TigerBeetle for quite a while. And then it was TigerBeetle, and then Bun... Also Mach engine was around the same time as TigerBeetle... But we really just picked it because I was doing a lot of C code, and Zig was just a perfect replacement for C, and for these new memory patterns. That was it, yeah.
+
+**Jerod Santo:** \[01:35:13.22\] Very cool. Ahead of the wave on that one. You were an early adopter on Zig, and probably one of Zig's - I don't know if largest codebases, but maybe most production grade and out there, successful projects to date. Would you say that's fair?
+
+**Joran Dirk Greef:** I would say yeah. Bun is also pretty massive...
+
+**Jerod Santo:** For sure.
+
+**Joran Dirk Greef:** There's also Ghostty, by Mitchell Hashimoto, which is like incredible code. His performance work there is very similar, trying to get as close to pure memory bandwidth as you can. So he's making a terminal... How can you make that as close to memory bandwidth performance?
+
+**Jerod Santo:** Right.
+
+**Joran Dirk Greef:** Which is also what we think with TigerBeetle, and same as Jared with Bun.
+
+**Jerod Santo:** Yeah.
+
+**Joran Dirk Greef:** And I think that goes back to what we were saying, the love for that \[unintelligible 01:36:04.19\] that you see. We're actually trying to show what it feels like if you're coding in open source, because we've really crafted everything. You may as well enjoy... And this sort of came from Antirez; his craft of Redis impacted me a lot.
+
+**Jerod Santo:** Well, speaking of all these things, we just put a clip out today as a record of our conversation with Antirez, which was just a couple of weeks back. And did you know he's hard at work trying to get Redis to be open source again, inside of Redis Inc.? He's advocating.
+
+**Joran Dirk Greef:** Oh, wow.
+
+**Jerod Santo:** He's returned to Redis, and he thinks he can get the company culture moved in a place where they'll get switched off of that proprietary new license, and... Probably AGPL, but I think you'll find that good news, considering your stance on open source. I'm excited about it. Hopefully that happens.
+
+**Joran Dirk Greef:** That's great. And again, I should be clear - I love open source also because I love how it enables business, businesses. So I actually think it's great for business. I don't just like open source because I like it, you know. But I actually think it is better for trust, for sales... It makes everything easier. But yeah, that's fantastic news. I can't wait to listen.
+
+**Jerod Santo:** Yeah. Cool, cool, cool. Well, Joran, I appreciate you coming on the show. I'm fascinated... I don't have any use cases for TigerBeetle in my life, but I respect it, and I'm sure our audience will enjoy this conversation as well. So appreciate your time, appreciate your work, and your perspective on open source and business, which is refreshing, especially in a sea of people who are kind of moving away from your perspective. But maybe back again. We'll see. I mean, Elasticsearch is back, maybe Redis is coming back... We'll see what happens, but I appreciate you coming on the show.
+
+**Joran Dirk Greef:** Yeah, there's always a new wave to come. Yeah, and I appreciate you too, Jerod. Thanks so much for this. It's been really, really special.

The CEO of htmx likes codin' dirty (Interview)_transcript.txt

@@ -0,0 +1,439 @@
+**Jerod Santo:** Today I'm joined by the CEO of htmx himself. It's Carson Gross. Welcome to the Changelog.
+
+**Carson Gross:** Hey, thanks for having me on. Happy to be here and happy to talk about software, and maybe some htmx and so forth.
+
+**Jerod Santo:** I'm happy to have you as well. I feel like our paths have crossed a few times. I think you've been on a show I produced, maybe a show that I co-hosted, but never got to sit down one-on-one and just talk software, so this is exciting to me. You have lots of opinions about software, don't you?
+
+**Carson Gross:** I do. I have a fair number of opinions. What are they? You're supposed to have like "something, opinions, strongly held", or weak opinions, strong -- whatever that is.
+
+**Jerod Santo:** Yes. Strong opinions, weakly held, I believe. It could be the other way around, who knows...?
+
+**Carson Gross:** Is that the right thing? Well, Jerod, I have the opposite of what you're supposed to have, so...
+
+**Jerod Santo:** \[laughs\]
+
+**Carson Gross:** But I do have opinions... But I don't think everyone has to agree with me, for sure.
+
+**Jerod Santo:** What's interesting about you - you've had these opinions kind of embedded in software over the years. Intercooler is where it started, htmx it became... You've spent a long time trying to get people to pay attention to htmx. I remember, because we had you on Go Time, we had you on JS Party, and you were kind of starting to get some traction... And then you just decided - I'm just guessing, decided - like "I guess I need to just become a Twitter troll, and that's how I'll actually get attention." Because eventually, you actually did rise to the point where htmx is out there. Like, you are a known quantity at this point. Was there an epiphany, where like "I just had to become the world's best troll"? Or how did you do it?
+
+**Carson Gross:** Troll is such a -- you know, trolling, for lack of a better term, is good. Now, I'd always had fun on social media, not taking it very seriously... And I think what really happened -- and there's a video you can look up online; the speech that I gave at the Big Sky DevCon, which is the big conference out here in Bozeman, Montana, where I live. And I talked about sort of my experience marketing htmx, and how I had had Intercooler, the predecessor to htmx, technologically very similar. Conceptually very similar. And it didn't have nearly the success that htmx had. And I don't think that there's anything that I can point to and say "That's what worked." I think there's a lot of survival bias problems, and just random things that happened.
+
+And so I did, when htmx was sort of going to go to 1.0, and also in our book, Hypermedia Systems, which your listeners can check out at hypermedia.systems, online for free, and you can buy it in hard copy if you want... Just sort of like a bunch of things came together. And I had been getting -- so when I went to htmx 1.0, I was like "Okay, I'm going to double down on being crazy on Twitter. I'm not going to troll people. I'm not going to attack people." That's one thing that a lot of people try to do, is you punch up to try and get a bigger account to say something to you... And I definitely engaged with bigger accounts, but I tried to be funny instead of sort of mean, and whatever.
+
+So in that talk - I think it's called "A theory of open source marketing", or something like that... I talk about sort of like the whole process, and what I've found has worked online. So I wouldn't say I set out explicitly to be a troll. I think people are used to very serious technical content, and I do try to have -- I published an ACM peer-reviewed paper on Hypermedia, and we're hoping to get another one into the ACM Hypertext Conference this year... So we're doing intellectual work as well, and more serious work, but at the same time, I think the medium is the message at some level. And on Twitter, if you're taking it too seriously - or any of these Twitter-like situations - you're probably not utilizing it the way it was meant to be. It's like, it's a hundred and whatever characters... Saying something funny and coming up with a good meme is probably going to be much more effective than trying to be all serious.
+
+So I think a lot of things kind of came together... Once ThePrimagen and Fireship\_Dev happened to cover htmx, like right when we were releasing the book... And so there was definitely this sort of moment on... And then the X algorithm changed briefly... It's changed back, so I don't get nearly the engagement I used to, but for just a brief period of time, the way I was posting was just perfect for --
+
+**Jerod Santo:** Boosting you.
+
+**Carson Gross:** ...yeah, for the algorithm. And everyone was like "Why is htmx on my homepage every day?" I'm like "I don't know, I just --"
+
+**Jerod Santo:** "I don't know, but I like it."
+
+**Carson Gross:** I just got lucky.
+
+**Jerod Santo:** \[08:01\] Yeah. That's kind of the way it is. Yeah, you've got to play to the crowd, and know your audience, and know the medium, like you said. For a very long time on Twitter my entire approach was just memes. And I eventually -- and I did pretty well. I had some viral ones, and we grew our account from like 2000 up to like 30,000 people... And I just kind of grew sick of the entire rat race of what it was. This is back in the pre-Musk days.
+
+**Carson Gross:** Right.
+
+**Jerod Santo:** But it was fun for a while, to try to play the game and see if you can get good at it. And of course, you've been using that to promote your open source, and your thoughts, which really - it seems like you're trying to get your way of creating apps out there, through htmx, but also through your essays... Which are technical and thoughtful, and still interesting reads. They're not like ACM-style essays. Or maybe they are. I don't read the ACM. It just seems like it's too dry for me.
+
+**Carson Gross:** Yeah, it's pretty technical and academic.
+
+**Jerod Santo:** But just getting your thoughts out and sharing them... Hypermedia, a big thing that you're into; I think taking HTML seriously is something that I've heard you say... Where does this viewpoint of the world come, of web development particularly?
+
+**Carson Gross:** Well, I think one thing that I've realized is I built Intercooler, and then htmx is -- I didn't appreciate hypermedia really when I built the things. There's this idea that like existence precedes essence. That's a philosophical idea where something has to exist, and then you can look at that and say "Okay, what's the essence of this thing?" Like, strip away all the particularities of this thing. Like, what's the essence of it? And that was very much my experience with htmx, where I had to build it and then understand "Okay, wait a minute, this is really generalizing hypermedia controls." When I created Intercooler and htmx, I didn't think "Oh, I know what I'm doing. I'm generalizing hypermedia controls." But when I looked at what I had done and I went back and read some of the historical writings on hypermedia, then I kind of developed that idea, like "Okay, this is why this is working as well as it is." Because it always felt like I wasn't writing that much code, and I was getting a lot of bang for my buck out of it. And I think the reason for that is that it was taking this sort of like higher level idea of hypermedia and then generalizing it, is the way that I would say it.
+
+So I just think this is one of those things that you have to do it before you can really talk about it, and so I did it with IntercoolerJS and then htmx, and that put me in a position to write these more sort of like deep essays on hypermedia. And the thing that I came to understand about hypermedia as I got into it is it's one of a few different ways to build distributed systems. If your listeners are familiar with Roy Fielding, he wrote a really famous dissertation sort of early on for the web... And it's pretty academic and hard to read, so I don't -- there's a reasonable summary of it in the hypermedia systems book that you can read if you want... But what he talks about in there and what I've come to appreciate about hypermedia is it's just one of these sort of fundamental ways you can build a distributed system. And it was a really effective way, as evidenced by the fact that the World Wide Web sort of took over everything, in the late '90s and early 2000s. So developing an appreciation for what's unique about it, what are its strengths, what are its weaknesses... That's just something I've been able to do over the, last, say, 5 or 10 years, as I've been working on on these things.
+
+And then I'm trying to communicate that to the broader development world, because the broader development world has sort of left hypermedia behind because of the weaknesses in particular of sort of the base HTML web model for hypermedia. htmx tries to address those weaknesses... And then it's a matter of convincing these people like "Hey, I know the old way was sort of clunky, and the big kerchunks, and all this flash of unstyled content and all that stuff, but the web's gotten a lot better."
+
+\[12:03\] And by the way, htmx can make it even better, and make it more effective. And then the advantages of that old approach, which were simplicity - there's a simplicity to it, and so forth - it can be available to you, even if you're building some subset of modern applications. It's not appropriate for everything, of course, and so I try to talk about the tradeoffs that are inherent in hypermedia-based applications... But it can be used for a lot of stuff. And even if you don't use it, I think it's interesting from a technical perspective just to understand the ideas and what the tradeoffs are.
+
+**Jerod Santo:** Sure. If you were to distill down what hypermedia is into a few bullet points or principles, so they're all on the same page...
+
+**Carson Gross:** Yeah.
+
+**Jerod Santo:** What would you say?
+
+**Carson Gross:** I'd say a hypermedia is some sort of media. So text is the most common media we work with, HTML, with what are called hypermedia controls in it. And that's nerdy, but bear with me. What a hypermedia control is is something like a form or a link, where you've embedded in the document what Ted Nelson would call a nonlinear control flow mechanism, where you can jump to some other place, or you can submit some information. And that control - so let's use the term control, and UI controls, like buttons and so forth... That control is embedded in the document in a really kind of unique and interesting way. So you have you have content, and then you've got this control sitting there, right in the middle of the content. And so those are hypermedia controls.
+
+And what HTML does is make it possible to sort of generalize that. So a lot more things can be hypermedia controls, and they can do a lot more, and they can replace only chunks of the screen, and so forth. And a lot of these ideas are very old. The idea of replacing only a chunk of HTML and a screen has been called transclusion for a very long time, and a lot of hypermedia academics have talked about it... But it just never really caught on in the Web, because the only way to do it on the Web previously was with iFrames. So htmx sort of like breaks out of that and makes it more general, and a lot smoother. You don't have to deal with iFrames and the limitations that that has.
+
+But that's what I would say hypermedia is. It's some sort of media - we're most familiar with text, so hypertext, that has these weird things in them called hypermedia controls, that can interact with the network, and do crazy things, like navigational, and an update style when you talk about forms, which are added in HTML. You sort of update style operations to interact with a remote system.
+
+**Jerod Santo:** So the H in HTML hypertext, right? HT, hypertext. HT hypertext markup language.
+
+**Carson Gross:** HT - yeah, hypertext markup language.
+
+**Jerod Santo:** So that would be a form of hypermedia. But htmx exists, and I guess a lot of these other styles of building applications exist because, like you said, of the lack of tooling, or the lack of power that's in the HTML spec, or in the elements provided. Was this something -- like, when the Fieldings and the Tim Berners-Lee's and, I don't know, early Microsoft IE browsers, Netscape people... Like, were they trying to build a hypermedia thing, and they came short of it, because HTML wasn't powerful enough? Or how did it get to where HTML is a hypertext, but it doesn't have all the controls that we need to actually build what we want to build?
+
+**Carson Gross:** I think hypermedia is just sort of a weird thing, and people aren't used to it. We developers, when we think about distributed systems, we fall back to this RPC, remote procedure call paradigm, just very quickly. And maybe that's because that's what we work in for the most part, is RPC... I see with a lot of SBA style applications as -- it's much more of an RPC style application... Which is fine, there's nothing wrong with that. It's just different than hypermedia.
+
+\[16:13\] So why did they stop with forms in HTML 2? ...which -- and that was a really important advance. To the credit of the web and what they did with the web, all of Web 1.0, which includes Amazon, and Google, and all that stuff - that was all done with Web 1.0 tools. That was a massive advance as far as distributed computing goes.
+
+**Jerod Santo:** True.
+
+**Carson Gross:** So the anchor tag and the form tag are both very powerful. Just those two little controls are very powerful, in some ways... But they just weren't as useful as it could have been if they had kept advancing the hypermedia aspects of the web. Instead, what they really sort of turned to is more -- they introduced a lot of APIs that were more client-side-oriented, and scripting. And so that makes some sense to me... You know, that's just the way we developers tend to think. And I think it gets big enough and the hypermedia people start getting ground out, and maybe they just weren't thinking too much about the idea of generalizing this notion of hypermedia controls...
+
+I think it was it was so powerful at the time, it's hard to look back and say "Well, guys, this could have been more powerful." It was doing really well. And then JavaScript was there to address any issues that came up when it wasn't powerful enough. So we just sort of slid into this world where JavaScript took over more and more, so therefore there was a necessity to keep going with the hypermedia ideas that were sort of inherent in HTML.
+
+I think it's reasonable -- and you mentioned before we got on here, the term REST has sort of been completely flipped in its meaning, from the start till now, because of that.
+
+**Jerod Santo:** Totally.
+
+**Carson Gross:** So there's a historical process there. You know, Google Maps, I think, shares a lot of the blame, because Google Maps came out and it was just this amazing app compared to like all the -- if you used Map Quest beforehand, you had to click a button, and kerchunk, and like go to the next thing... It was horrible. And Google Maps comes out, and that's not building in a hypermedia way at all. It's all RPC-style stuff, and it's amazing. And everyone freaked out and was like "Well, we've got to build apps this way now." And that's totally understandable, but it's just the historical process... So hopefully we can bring hypermedia back into the consciousness of web developers as a potential and good alternative for some class of web applications.
+
+**Jerod Santo:** Right. So HTML has the anchor, it has the form, and we do have Ajax for like -- you can build in partial page updates... We can build around all this kind of stuff. But what are some examples of the things that htmx adds that really gives it what it needs to be a full hypermedia API, or whatever you want to call it?
+
+**Carson Gross:** Yeah, the way I would say it is "How does htmx generalize this idea of hypermedia controls?" And so what I've eventually hit on was the htmx takes -- so anchors and forms. What do they have in common with one another? Some user action, a click or a submit, depending on which one, triggers an HTTP request. A request to some remote server. And then that server responds with HTML, and then that HTML is placed somewhere in the document. So just anchor tags and forms - that can be controlled with what's called the target attribute, and you can actually target an iFrame by name if you want. That would be transclusion, where you're only updating part of the page.
+
+\[19:53\] So what htmx does is it says "Okay, let's generalize those things." First of all, any element can be a hypermedia control, can participate in this sort of like remote request world as a hypermedia control. Any event can trigger it, so not just clicks and submits, but also, for example, when an element scrolls into view, that's a common time when you want to maybe make a remote request, or something like that. And then it can issue an HTTP request... htmx also supports things that aren't supported by HTML right now. So in plain HTML right now, all you can issue are get and post requests.
+
+That's, I think, a shortcoming of the HTML specification, that we're actually -- we have another project called Triptych, that Alex Petros, who's a core member on the htmx team, is really driving. And we're trying to get HTML to adopt some of the ideas of htmx. So we're trying to get those other requests - put, patch and delete are the ones most people are going to be familiar with - so that can actually use them from HTML as well. But htmx, you can use all those.
+
+And then htmx - the final sort of generalization of it is you can, say, when the request comes back -- and that request, when it comes back, it isn't going to be JSON. So you're making a request URL. You're getting back HTML, and you can say in htmx, "I want you to target this element." And the element that you target doesn't have to be an iFrame. It can be any element on the screen. You use CSS selectors to pick it out, and then it'll replace that element. Or there's a few different options. You can say either replace the whole thing, or put it inside the thing, or maybe put it at the end of the thing, or whatever. And so that's really what htmx does, is it takes the sort of specialized ideas that you find in anchor tags and forms, and then generalizes them. So any element can act like a hypermedia control now and can replace any part on the page.
+
+**Jerod Santo:** And if you were to provide the virtues of that, like "Here's why hypermedia is better" or, tradeoffs etc, as opposed to the way we're building web apps nowadays, which is very much more of an RPC style thing... What would you say?
+
+**Carson Gross:** I would say the biggest thing is you're reducing the amount of state that you've got in the total system. Because when you have an RPC system, you have server state and you've got a client-side state store of some sort, and then you've got the DOM, which also at some level is a representation of the state... And all three of those things need to remain in sync via callbacks, or whatever, and blah, blah, blah. And so that's complicated. That's just another layer. Whereas if you go back to the hypermedia model, there's really just the server and the DOM, and that's it. So you get rid, you collapse one layer of state and state management that you have to worry about... So you're not really talking about client-side routing or any of that stuff. That all goes away. And it's just interacting with the server in a stateless manner, relatively stateless manner, sort of like the original web... So virtues of that are really simplicity. It can, in some cases, dramatically cut down on the amount of code you have to write.
+
+The trade-off of that is that the interactivity that's available isn't always as high as it is if you have a much more rich client-side's model, and the sort of reactive feedback that you can get with client-side libraries. So it's gated on this network request model, and if your system can't do that well, then it's not going to be a good approach.
+
+**Jerod Santo:** There's some echoes of an old-style jQuery building of applications where you have some CSS selectors, maybe you hijack a form submit, and then you say "When the submit happens, go grab this HTML, target this thing, put it in there." And we built web applications that way for many years, with more or less success. You can build yourself a sprawling mess like that. It can happen.
+
+**Carson Gross:** Yeah. Absolutely.
+
+**Jerod Santo:** \[23:51\] So one of the things that -- maybe the greatest virtue of the React movement was this one-way data flow, re-render, don't worry about the target spot for this thing, and is it going to intercept with that, and now I've got to update, I've got listeners, now I've got a PubSub going on, because I have different things I need to update based on what's coming back from the server... Does this have something built into it that helps avoid the jQuery potential spaghetti, or is it just a different way of going about similar apps, but because it's in the HTML, maybe it's simpler or cleaner that way? What are your thoughts on that?
+
+**Carson Gross:** Yeah, I think there's definitely a line at which if you try to do too much with htmx, you're going to get into that sort of like "Oh man, this is just too much." And it's an interactivity line, in my opinion, rather than like an application size line. You can build very big applications with htmx, but if you try and get too fiddly with the UI, and you're putting a network request on every interaction, then that's probably not going to be good.
+
+So you can use scripting with htmx... That's sort of the style of scripting that you would use. I would call it very DOM-oriented. You wanna try and have like a model. Whenever you have an additional source of truth in your system, and htmx is in there, you're going to start having to deal with synchronization issues. That's just the nature of it.
+
+I think where htmx is superior to what you were talking about, the old jQuery mechanism of building these styles of applications... Older listeners might remember jQuery.load, which was the method actually that pretty much inspired Intercooler. Like, I looked at that and was like "Oh, that's cool. I didn't know you could do that", use that to solve this problem that I was having in a web app. And then I saw Angular 1 and the way they used attributes in the HTML, and I didn't like the way they were programming, but I liked using attributes. And so where I think htmx is superior and where you don't really get the sort of spaghetti code that you sometimes got with jQuery is because with htmx, you put the attributes that specify what to do on the elements themselves. So it's right there.
+
+Older developers who work with jQuery a lot have no doubt run into this problem where there's a button, it does something - what does it do? Well, you've got to find the ID, and a selector somewhere, and maybe someone did something crazy, like string concatenation... So it's really hard to figure out what the heck is going on with that. And htmx eliminates that by basically putting the logic -- just like the href or the action on a form tag, it puts it right there on the element. And I think that dramatically simplifies things from a cognitive standpoint. You can obviously still build a snake's nest with it, and again, if you try to get too fiddly with interactions, I think that can lead to that situation... But for a large class of applications, the interactions aren't that fiddly. What you don't want is that really old reset of scroll state, and like the flash of unstyled content... Which has gone away in modern browsers, for the most part. But just that feel of like resetting everything on every action that you've got with the old web. htmx avoids that, and I think that's why you can make pretty compelling applications in many cases just using that model. Now, they don't have to spiral out of control, sort of like the old jQuery apps did...
+
+**Jerod Santo:** No, I think you're right. The -- I don't know what you call it, but like the co-location of the code that operates on the elements, with the elements, is a huge improvement over... I think we went kind of crazy on separation of concerns back in the jQuery day, where it's like all your JavaScript is over here in this other file, that like randomly attached -- not randomly, but arbitrarily attaches to different elements... And then you're wondering -- in your HTML it's super-clean, but you're like, "This HTML has no idea what's going to happen with it." And so I think having it there does make a lot more sense.
+
+**Carson Gross:** \[27:50\] Yeah, I wrote an essay on that idea called "Locality of behavior." You can find it on the htmx.org/essays page, where I talk about exactly what you're saying, which is this idea of co-locating the logic with the element itself, and sort of stepping away from separation of concerns... Which I think has not been nearly as beneficial as many of the advocates for it said it would be. And we've had that experience now with these old jQuery apps. We just know it's better. Like, "Man, I just wish I could see what this thing does."
+
+**Jerod Santo:** Totally.
+
+**Carson Gross:** So as long as that logic isn't too complicated and doesn't overwhelm the markup - and in the case of htmx, it typically does, and it's like two or three attributes at the most - then I think it's fine. Tailwind's CSS sometimes gets mentioned as well, and Tailwind's -- I've not taken the Tailwind's build yet, because when I see a ton of Tailwind classes on an element, I think to myself "Man, that seems like that should be a CSS class to me."
+
+**Jerod Santo:** Right.
+
+**Carson Gross:** But it's just probably a skill issue on my part. So I think there's a balance there...
+
+**Jerod Santo:** There is a balance, and I'm definitely on that fence, where I'm not sure exactly -- I kind of agree with you, but there's a line that you cross at a certain point where it's like "I feel like I'm just writing my CSS in here." Anyways, we can get on a tangent on that. I think there's certainly, again, virtues to having it localized there... But sure, it'd be nice to have just like a couple of classes, and then have that stuff tucked away.
+
+**Carson Gross:** Right.
+
+**Break**: \[29:30\]
+
+**Jerod Santo:** One thing that's interesting -- I mean, what's normal you have on your website is porting stories, like "Here's a port from React to htmx", and there's a few of those. What's abnormal, and I've found refreshing and cool, is you actually have a counter port. It's not really a port, it's a lack of a port. It's Gumroad. They were going to choose htmx for a new system, and they actually decided not to... And so you host that there, on the htmx.org essays, written by Gumroad founder Sahil. I'm not sure how his last name is pronounced... Lavingia? Yeah, we'll just leave it right there. That's cool. How did that come together? And then maybe just give a summary of like what did they find? Why did they decide not to for their particular case?
+
+**Carson Gross:** Yeah, I think that's a good essay, it's worth reading... And I just saw them on Twitter talking about htmx and being pretty negative about it. And one thing that I've learned about online behavior is that often your first reaction is not the good one. And if you do the opposite -- you know George Costanza, when he did like opposite? He was just like "I'm going to do the opposite of everything I would do."
+
+**Jerod Santo:** Yeah, totally.
+
+**Carson Gross:** So I was like -- I've been trying to do that, and it has been surprisingly effective in my life. And I was like "You know what? Hey, it didn't work out." I don't want people to use htmx if it's not going to be good for them. No one's winning in that scenario. Just someone is pissed off they had to use htmx. So let's try and capture why htmx didn't work well for this particular use case. And so I asked him "Hey, would you be willing to write an essay on it?" And he agreed, and sort of went through, and like "Here's what we've found."
+
+I would say the way they were using htmx -- it was an SPA developer using it. It was kind of a square peg, round hole situation... And so they weren't using htmx the way I would use it. But at the same time, that's going to be the same experience that a lot of SPA developers have. And so they just -- they're really familiar with the SPA tools, and they make things work a certain way, and they're used to that... And so if they come in and -- I think the big meta lesson in my mind from that essay is like if you're going to take the htmx build, you really have to take it. You have to understand hypermedia, and think in terms of hypermedia. And if you're trying to do these really fiddly little interactions, it's going to get ugly, and not be super-effective.
+
+I saw WebDev Cody as well. He never wrote an essay. I asked him to, but he just was like -- I think what he tried to do is he tried to move to Go. He was a JavaScript guy and he tried to do Go and htmx at the same time. And I think that was a mistake, because now you're fighting with the new infrastructure of Go, and you don't know where -- you know, "Where's my wrench? Where's my hammer?" Like, everything's in a different place... And at the same time, you're also trying to make this pretty. And there is a really big difference between building hypermedia-based applications and building SPA style, or RPC style applications. And I've seen that, and not just his case, but in many cases. If you try and bite off too much, it can, I think, end up in a pretty bad spot.
+
+I always recommend that people try htmx out first on for example like an internal tool. Maybe you've got like some underserved internal tool that is used just by your co-workers support, or whatever it is, and it's just like an old-school Perl app, or PHP, or whatever... And maybe try sprinkling some htmx in there, and see if you can get some quick wins with it, and then that'll sort of show you what the strengths are. Rather than "Okay, someone on Twitter said hypermedia is good, so we're going to rewrite the whole app in hypermedia, and we're going to try and make it all work exactly the same."
+
+**Jerod Santo:** Yeah... So one of the reasons that I've found -- they did list like five, I think, reasons why, bullet points. And one of them I've found interesting, which is a little bit outside of your control, and it was AI and tooling support. And this is probably outdated at this point, because even -- I've had this similar situation with AI tools with Elixir; the earlier versions just didn't really know Elixir very well. It just was wasn't very good at them. And now as I've reached I think Gemini 2.5, and specifically Clawed 4, it's just better. Like, it just knows more, and so like that problem kind of went away, to a certain degree.
+
+\[36:14\] But there is kind of that old saying, "Nobody goes to that restaurant anymore. It's too crowded." It's like, Next.js sucks all the air out of the room, and so people are just going to pick it because it's what they know, it's what everybody else does, it's where the jobs are, where the money is... And so to a certain extent, some of the reasons why people make choices aren't actually based on the merit or anything you can control at all. They're just kind of like the way \[unintelligible 00:36:37.14\]
+
+**Carson Gross:** Yeah. Well, popularity is its own merit. Nobody ever got fired for buying IBM. It's maybe frustrating to people who are more philosophically inclined, but it's just the reality of life. Technology is very much a winner take all sort of situation, in many cases... Although we do see -- Microsoft, for example, was the big dog for a long time, and then was down for a little bit, and now it's maybe back... So these things that seem inevitable -- there are these ebbs and flows.
+
+And one thing I say in that talk that I gave at Big Sky Software is if you're a little guy like me - it's just me and a couple of other guys working on htmx - really the important thing, I think, is hanging around, and waiting for the winds to shift in a direction that is favorable to your technology. And the best way to do that, in my experience, is to stay unrelentingly positive and funny, and just have fun. And so that's one reason why is -- like I told Prime when I was on a show once, "Being funny is no joking matter." You have to like keep it fun, or you're going to quit, because it's frustrating, and these people are just picking that because Facebook said so, or whatever... And you just have to be okay with it, and like "Hey, we're going to go over here, we're going to have our fun, and we're going to make some funny jokes." The technology is serious, there's good ideas behind it, and I'm happy to talk to people about those ideas, but I'm not going to let the fact that so many people pick Next.js by default, and pick the SPA approach by default knock me out of a game. Because I'm not changing that anytime soon. I think we made a dent, and we made a dent because we're just hanging around, and having fun, and saying some interesting things.
+
+**Jerod Santo:** Yeah. Persistence absolutely pays off.
+
+**Carson Gross:** Yeah.
+
+**Jerod Santo:** Especially if you can stay positive and interesting for long enough, people eventually pay attention to you. So Next.js - of course, lots of money behind that sucker. htmx - do you have a goal in mind? How do you make your money? Is this just like something you do because you're a consultant, and you use it yourself, or how do you make your living, and how does it play into that, if at all?
+
+**Carson Gross:** I teach at Montana State University. So I'm a professor. That's how I make my salary, and healthcare, and all that stuff. htmx does have a fair number of sponsors; many more than I ever expected, which I'm very appreciative of. But I try to treat that money as not real, because it could evaporate tomorrow. And htmx is BSD Zero licensed, so there's not money to be made.
+
+**Jerod Santo:** There's no plan there.
+
+**Carson Gross:** There's no plan. Yeah. Do I look like a guy with a plan to you?
+
+**Jerod Santo:** \[laughs\] I wasn't going to say it. You look like a St. Louis fan. That's all I got. But you're out there in Montana, so how do you get to be a St. Louis fan?
+
+**Carson Gross:** It's a funny story... I've actually never been to St. Louis in my life. So I coach youth baseball. That's like my other cape that I wear. And I was in California for a long time; that's where I grew up. And in California, it was impossible to be the Giants, or the A's, or the Dodgers, or the Padres, or the Mariners. Like, you just could never be any of those teams.
+
+**Jerod Santo:** Oh, they're all taken by the time you submit your name?
+
+**Carson Gross:** Yeah, exactly. The cool coaches get all those. And I got sick of fighting over it. I grew up a Giants fan, but...
+
+**Jerod Santo:** Okay.
+
+**Carson Gross:** \[40:13\] But the Cardinals were always available, and I was like "I don't know, I like the hat. Kinda cool." I respect their program. And so we just had all this Cardinals gear laying around, because that's what my little league teams always were... And then MLB.TV came out and they wouldn't let me watch the Giants, because it was in market.
+
+**Jerod Santo:** Oh, like regional stuff?
+
+**Carson Gross:** Yeah, exactly. And so I was like "Well, I got the hat. I might as well start watching the Cards, I guess..." And so I started watching the Cardinals, and it just kind of spiraled from there. And now I like the Cardinals more than I like the Giants, so...
+
+**Jerod Santo:** That's hilarious. I had a similar experience in my childhood. For some reason -- it's just based on the markets. For some reason, I'm here in Nebraska and I don't know why it's this way, but when I was growing up, in the '90s on TBS, I think it was, every weekend it was just Atlanta Braves.
+
+**Carson Gross:** Oh, Braves.
+
+**Jerod Santo:** Which - why would it be Atlanta? I guess probably because in local markets --
+
+**Carson Gross:** Turner.
+
+**Jerod Santo:** Yeah, there you go.
+
+**Carson Gross:** Turner's out of there.
+
+**Jerod Santo:** Yeah. So I don't know why, but the Braves was on, and I watched the Braves like religiously as a kid, and I just loved them, and they actually were a good team back in the '90s. Good pitchers etc. They had some world series runs, and so I became a huge Braves fan. Now I'm not so much anymore, but it just seemed like for me, a kid in Omaha, I'm like "Why do you like the Braves?" Like, that's the only thing on TV, you know?
+
+**Carson Gross:** Yeah, exactly. That's basically my story, too. Actually, I did the -- especially when it was super-hot, about like six or eight months ago, a guy in the Cardinals organization saw me, and saw me wearing a hat... Like a technical guy, a programmer. And he reached out and was like "Hey, do you ever want to come down and get a tour of the stadium, let me know." I'm like "Man, I should do that."
+
+**Jerod Santo:** You should totally do that. Or at least go to a conference or something in St. Louis. Although it's kind of cool to be able to say "I've never been there." You'd ruin that part of it, but... Speaking of cool venues, we were at Microsoft Build up there in Seattle recently, and they had their after party in the Seahawks Stadium. I can't remember the name of the place. Lumen Field?
+
+**Carson Gross:** Yeah.
+
+**Jerod Santo:** So like Microsoft bought out Lumen Field for the night, and they just had their after party on the field. Like, you're on the field, they'd let you kick field goals, and stuff... That was pretty rad, just to be able to be there on the 50-yard line.
+
+**Carson Gross:** Yeah, that is pretty cool. I'll never forget -- what's his name? Beast mode... Marshawn Lynch, that run that --
+
+**Jerod Santo:** Oh, man...
+
+**Carson Gross:** Do you remember that run of his?
+
+**Jerod Santo:** Of course I do. Yeah.
+
+**Carson Gross:** Where he just throws that guy off... So I went to Berkeley for undergrad, and I was there when Marshawn Lynch was the runningback, and he was just always such an unbelievable football player... I'm not a Seahawks fan --
+
+**Jerod Santo:** Same.
+
+**Carson Gross:** ...but I watch that video every once in a while, because I'm like "Man, I need to see the greatest run by a Cal Berkeley alumni ever."
+
+**Jerod Santo:** Unbelievable. Yeah, absolutely unbelievable. Alright, well, now we're talking sports. Let's get back onto your essays. No, I apologize. Vendoring or Codin' Dirty - I want to talk through a few of your essays, just because I think you do a good job of... Now I know you're a teacher, and that makes more sense... Of elucidating your thoughts and kind of your style of programming.
+
+Let's go to this one, which I think you and I both seem to, which is the "Codin' dirty" one. I think you wrote this last year... And it's a bit of a response to clean code; not specifically, but there's this idea of how you're supposed to write software. Of course, Bob Martin's Clean Code book, which is a both highly regarded and also hated, kind of a piece of work... And you wrote "Codin' dirty", which is contrarian to a certain extent, but I also find myself nodding my head. So talk me through some of the thoughts here on your coding style.
+
+**Carson Gross:** \[44:04\] Well, Bob Martin and Clean Code is definitely a dominant idea. It's a dominant sort of ideology, I would call it maybe, in coding. And there are aspects of it that I don't hate, or that I don't disagree with entirely anyways, but there are aspects of it that I don't agree with. And I definitely code very differently, and I've had a pretty successful programming career... And so I wanted to communicate, particularly to younger developers, "Don't be intimidated by these things." It's one thing I dislike - and I Bob and I have kind of gotten into it on Twitter a little bit, jokingly though. He's been a very good sport and doesn't take it seriously, so I really appreciate that. We don't agree with one another, but we can make a joke out of it.
+
+So I really think there is a -- what do I want to say...? I think younger developers can be intimidated by these terms. Clean code is one of these sort of -- like, how could you be in favor of dirty code? How could you be opposed to this? And then you look at the actual ideas and you're like "Wait a second..." There's the headline, and then there's the actuality of it. And the headline - of course, everyone wants their code to be clean, but then the realities of the recommendations I think don't necessarily hold up. So now that I've had some success as a programmer, one of the things I'm trying to do is to go out and say "Look, guys, I do things pretty differently than most of these -- particularly sort of the consulting-oriented, clean code, agile, that whole world; the way they say to do things, I do things pretty differently and I've had a pretty successful career."
+
+So that was the inspiration of this essay... And I think the three sort of things that I went at - so I don't believe in short methods for the sake of being short. I think long methods are okay. And I don't unit-test my code. Like, I don't do TDD. I do test my code, but I don't necessarily do unit testing, and sort of what I would call ideological approach, of like always writing a test before you write any code.
+
+And then the last thing that I take issue with, or that I sort of say - this is how I do things different, is I tend to write code with classes that are much denser and bigger than clean code advocates would recommend. So I tend to mix concerns. I have classes that do more than one thing, which is something that clean code people would object to. So those are sort of the three things that I picked out in that essay to focus on, and I'm happy to go into each one individually.
+
+**Jerod Santo:** Yeah, let's talk about big functions, because I've definitely done it both ways. Early on in my career -- early on, you don't really have opinions, because you don't know what's good or bad. You're just like "I need to make a thing, and so I'm going to try to make the thing." And so I was very open to advice. And I remember early on, in the Ruby community, I took Sandi Metz's advice on many things. And one of the things she said -- of course, she's very level-setting in her ideas... But she'll say "Follow the rules until you know better, and then break the rules all you want, as long as you know why you're breaking the rules."
+
+One of the rules that she had was short functions. And I think she even says, like, five-line functions. Just give yourself five lines, and if it's more, just have another function. And just do that for a while, and eventually you'll know why, and you'll know when not to, and you'll break that rule. But there's a rule. I think it was five lines, which in Ruby is very small, because two of them -- we're not using curly braces, so like two of them is the def and then the end. And so you're like - what, three lines? I tried to follow that rule for a while, and I've found it incredibly constraining.
+
+\[47:57\] And eventually, I broke out of it, but it always gave me this sense of like -- you know, if a thing starts to get too big, I start to at least think about a refactor. I may not do it... But I didn't really consider back then this idea of function size matching kind of the importance of the code... And I definitely realized it with classes. I mean, you have like your user class, for instance. You can go look at any app and see its user class, or whatever. It's functionally equivalent to that --
+
+**Carson Gross:** Massive.
+
+**Jerod Santo:** And you're like "Yeah, okay, here's the app." Or there's usually like two other classes that are like "Here's where the majority of the app is." I think that's good. I think it's good to have this sort of like matching up, which I think is some of your point. But when you say long functions, or big functions, are you talking like 10 lines, 20 lines, 500 lines, 2000 lines? What are your thresholds?
+
+**Carson Gross:** I don't really have a threshold. I definitely will start feeling antsy when a function has more than, say, 100 lines of code in it... But if it's an important function and if there isn't reusability to be achieved by pulling the code out, I still won't refactor it, just because it's too long. I try to make being too long not the reason I do things. If there's code reuse, then I'll do it. And the reason for that is -- so you touched on it already. I think important functions should be big. Important things in general should be large, and unimportant things should be small. And so when you go into a codebase, if there's 100 five-line codes, it's kind of hard to get a handle on what is happening there. And obviously, it depends on exactly what style of code that is. Is it like a library, like a list thing? Or is it like domain logic? Or is it control logic, or whatever.
+
+**Jerod Santo:** Sure.
+
+**Carson Gross:** So it really depends on what style of code that is. But I prefer to go into a codebase and be like "Okay, here's the five functions that are really important for me to understand. Let's go through them and read them top to bottom", and not have to be navigating down five levels deep, and then coming back up, and then going back down... You know, doing that sort of treewalk that you have to do if you have too much decomposition in a codebase. And I use IntelliJ and all the JetBrains products, and I'm good with navigation... So I can move around, but I still just find myself preferring just like "Just tell me what you're doing." Show me, like, "You do this, then you do this, then you do this", and fine. It doesn't all need to be refactored out if there's no code reviews to be achieved with it. So I think that's important.
+
+I think another important observation... And this is -- a bit of wisdom in computer science is that there's two hard problems - naming things and cache invalidation, and off by one errors.
+
+**Jerod Santo:** Right.
+
+**Carson Gross:** So naming things is one of the two hard things in computer science. And whenever you extract a function, you've got to name things. You're like committing to the parameters, the signature... Like, if you accidentally make it public, someone else suddenly starts relying on it... There's a lot of danger in pulling code out into a separate method, that I think is unacknowledged by people that act as if pulling code out to another function is just a free operation, or in fact will just make things better just because you get it. That name changes, you're sort of committing yourself to that particular bit of logic now, because it's somewhere else, so it's harder to go and change, and there's just a lot of -- like, things like that, where you end up constraining yourself because you've committed to this sort of signature now, and name for this thing.
+
+So I definitely -- like, I start to get itchy when a function's 100 lines of code, but I'll also try to be disciplined and like "Okay, if there's no code reviews to be found here, I'm not going to pull it out just to pull it out." Maybe for testing, maybe... But then that gets to the testing thing, where I tend to prefer end to end, and what are sometimes called integration tests, rather than unit tests of like functions. Like, I don't like unit testing as much as I like those other styles of tests... So that I think sort of dovetails with my opinion on longer functions.
+
+\[52:13\] There has been some academic work, some actual empirical work on this, believe it or not, and the empirical work that I've seen indicates that actually, until -- I think it was like 200 lines of code, you actually end up with fewer bugs per line of code if you're willing to allow functions to get larger. And then there was a modern -- someone kind of like tried to fact-check me with a more modern paper on it... And again, that paper didn't show strong association with longer methods and bugs per function. So it doesn't seem like there's much empirical evidence that long functions are bad. It's mostly just an assertion based on what, as far as I can tell, are aesthetics.
+
+**Jerod Santo:** Right. So here's something that I will do sometimes... And I wonder if you would do it or not. It sounds like probably not, if code reuse is your indicator for extraction. Let's say I have a module that does one major thing, and it's kind of complicated. Like, let's just call it like a checkout of an e-commerce. You know, at checkout time there's five or six things that you have to do. And so there's your function - whatever, call it checkout. And then it's going to have five or six steps. And each of those steps -- I might even leave myself a comment, like "Here's my steps - one, two, three, four, five", and I'll describe them. And then I will either have to decide, "Do I just code those up right here? Or is each one of those the name of a subroutine and I code them up in separate routines?" I've done it both ways. I honestly don't have a strong opinion on it being better or worse... But would you just like code all five steps right there in the function most likely, unless there's code reuse?
+
+**Carson Gross:** Yeah.
+
+**Jerod Santo:** You would.
+
+**Carson Gross:** I'd probably put all five right there, maybe with comments indicating "Step one, step two, step three, step four, step five." And in particular, what I would look for is is there shared state between those things? Because if there is, if there's more than one thing that's kind of crossing boundaries for those five steps, now extracting that out can get dangerous, because do you have to encapsulate that state, and are there callbacks, or like some crazy -- you don't want some crazy callbacks... But I would be inclined to write it the linear way, with just ifs and loops, and then if that proves to be unstable or I need to test them individually or something like that, then maybe I would extract them at that point. But I definitely wouldn't extract them -- I would write it for sure the first time just a straight line code --
+
+**Jerod Santo:** Straight through.
+
+**Carson Gross:** ...and then if that wasn't good enough for whatever reason, then I would start -- and I would look for opportunities to like pull out utility methods, if I see kind of the same thing in a couple of places, but... I wouldn't feel it necessary to structure something where it is. That's an important thing for someone to see, that there's five things going on here, and sort of in a linear set of steps... And so I wouldn't try and hide that from them too much.
+
+**Jerod Santo:** Right. But for instance, if step two was calculate shipping - well, there might be some other time you're going to calculate shipping. You'd probably wait until you're actually doing it elsewhere, then you might pull that one out, call it from there, and then also call it from somewhere else. That makes a lot of sense to me.
+
+**Carson Gross:** Yeah. That's the way I would do it.
+
+**Jerod Santo:** I do like unit tests, not for any other reason but they're easier to write, and they're faster to run, and so I feel more productive. Integration tests are like such a drag, because you have to do everything... And I know that's like what the app does, but it makes me not like them. They're slow, they're expensive, a lot of times they include some sort of a runner, which may have to fire up Chromium, or something...
+
+**Carson Gross:** \[55:49\] Right. Yeah, so the language here is always difficult, because like, what is a unit test? And this is one thing that infuriates me whenever I'm dealing with people online. Test advocates are often like -- they'll argue definitions, rather than concepts. It's like "Well, that's not a unit. That's not really--" I was like "Okay, sure \[unintelligible 00:56:07.03\]
+
+**Jerod Santo:** Right. "Not useful." Yeah.
+
+**Carson Gross:** So to me, a unit test is a test where you have one function, that maybe doesn't call any other parts of your system, or you mock those calls out; that's a very common practice in unit testing. And so you're really just testing the logic in one method at a time. Whereas what I call an integration test is anything sort of higher level than that. And I think even testing the public API of a class, for example, is not unit testing necessarily, because the internals of that class might be pretty sophisticated; they may hit a database, whatever. And so that's the level that I like to work at.
+
+My criticism of unit tests is that they often test implementation details, and they're not high-level enough. And so yes, setting up a good integration test suite can be some work, and can take some time to build out all the utilities that you need, and so forth; often you have to build data objects, or whatever... But my experience is that those tests tend to survive much longer, because they express higher level truths about your codebase. Whereas unit tests tend to express more ephemeral, like, "This is true for now, but then if I go into a refactor, then they're all going to break." And the real hard thing about that, in my opinion, and one reason to maybe be a little bit more careful with your unit tests, is that when you get a large unit test suite and you've got thousands of tests, that size has its own quality of like leaning on you. Like, you can't do a big refactor, because you're going to break all these tests. And which of these tests actually -- like, which of these tests that are broken are broken because we did the refactor and we want them to break, and which are tests that we actually broke something important? And that can be really hard to tell when you get -- I've worked at companies with 50,000 tests that take 10 minutes to run. And it's like "I'm not going to refactor anything, because if I do, if I try and clean it up, all this stuff is going to break and I'm going to spend four weeks trying to figure out what is really broken and what just went away, or whatever."
+
+So that again, I think that's one reason to try to get the assertions that you make about your system up to higher levels, sort of away from the details of the implementation, which unit testing -- and again, you can argue about language, and all the rest of it, but I think unit tests tend to test implementation details too much. But I acknowledge a hundred percent what you said, which is that the higher level the tests get, the harder they can be to run. They take longer to run, because you've got to set up maybe some state, and so forth... And then when you get to end-to-end tests, like you're talking about browser tests - it's so far away from the code that like, "Okay, did this test break because a label changed, or an ID of an element changed? Or did something actually break down in the mod?"
+
+So I love end-to-end tests, but I think you have to be really careful there too, for exactly that reason. You need to keep that end-to-end test suite relatively small, otherwise people just start ignoring it, because it breaks all the time, and no one can figure out what's going on with it. And that's why, again, I really focus on what I would call integration tests, for lack of a better term; just like not unit tests, and not end-to-end tests. So what are those? I don't know. Integration tests.
+
+**Jerod Santo:** Right. Somewhere in the middle.
+
+**Carson Gross:** Those in-betweeners. Hopefully, they can run fast enough that they're still not incredibly painful, and then when something breaks, it's more obvious than in the end-to-end case. So things break less frequently than unit tests, and more obviously than end-to-end tests. So that's why I speak up for what I call integration tests, but... I don't know.
+
+**Jerod Santo:** Right. Yeah, there's certainly a give and take with a test suite, where it's like on one hand it's kind of an asset that you have, because it gives you assurances that your software continues to operate the way it was designed. Of course, with the asterisk in there that maybe you tested it wrong, and stuff like that. But on the other hand, if you value that asset too highly, now it's actually holding you back from making progress.
+
+\[01:00:18.19\] So in smaller codebases, especially ones where I'm the primary author, I have no problem either refactoring the tests right along the code, or deleting the tests, sometimes after I write the unit test; it's already done its job, I know the function works... Tests can be gone. But in large projects, with multiple developers, you didn't write those tests in the first place... You don't know if that test is passing because you changed this interface in a way that makes sense, or if you actually broke something unbeknownst to you. And so it becomes a -- it can become a hindrance to progress in that sense.
+
+**Carson Gross:** Yeah, and I think this is similar to the case with extracting functions. People act like adding a test is cost-free, or like always a net positive. Like there's no cost associated with it. And that's just not true. Like, extracting a method, there's a cost associated with that. Adding another test that isn't going to -- even if it's testing correct behavior today, if it's super-low level and testing like a particular implementation, that can hold you back from changing that implementation. And people say, "Oh, well, you just refactor the tests, too." And that's fine to say, but as you point out, when you get into a big development, or even if you can't remember why you added the test, which - I'm old now, so that happens... Like, "What does this test even do? I don't know." Then it's not as easy to just delete tests, as people say. So there are costs associated.
+
+I think that's one meta point I would make to anyone who's listening - and particularly younger developers - is that all these things have costs associated with them. Like, closures are another great example. I love closures. I love doing data structure manipulation with sort of functional style maps, and all that stuff... But closures can get super out of control, and it can turn into a callback hell, like you have in JavaScript. So... So many of these tools that we have - it's just Aristotle again, right? Like, there's underuse, and overuse, and then there's the golden mean. And figuring out what that golden mean is is sort of wisdom in development.
+
+So I think anytime you get someone who's like "X is the way", it's like "Well, okay, but there's probably some trade-offs that are associated with that." And I like to talk about those trade-offs, and not present any particular tool or approach, or certainly the way I do things as like always unilaterally just the best way to do things.
+
+**Jerod Santo:** You also prefer a few classes, you don't mind god objects, and the old fashioned FLs versus polymorphism...
+
+**Carson Gross:** \[laughs\]
+
+**Jerod Santo:** So I'm with you on some of those. I don't have a problem with god objects. I know that's a pejorative. People say that as usually a bad thing. But I think you said that systems have certain more important parts than other parts, and if the user and the -- imagine GitHub, the repo... Imagine how much code is in their repo objects at GitHub. I'm sure it's more than is in their issue objects, for instance. And that's indicative of like it's about repos, you know? And yeah, it can get unwieldy. But I think we are too afraid of that, too early, and hedging against something that may never happen.
+
+**Carson Gross:** Right. We're prone to over decomposition, is the way that I would say it. "If we just break everything up, it'll be --" and it just doesn't... Again, decomposing a system is not cost-free. And yeah, cohesion is the idea that things that are related, belong together. So you probably don't want your user, for example, being also where you send all your emails from. On the other hand, you're always sending an email to a user.
+
+\[01:03:58.25\] So to me, put some methods on the user to send emails and that's alright. Like, it's okay if that functionalities in there. And my experience is the same as yours, the user object tends to be this massive thing... And I just don't think it adds much to break that up a whole bunch. In fact, again, I would say there's a cost associated with that. If you have everything in one file, if you have everything in one class, then it's easier to communicate across... You don't introduce additional lines to your UML diagram, right? I'm not a big UML guy, but I do know that more lines is bad/badder in UML. So the insight there, I think, is to be careful with your decomposition at the class level, not just at the function level. So I think it's just, you know, the same thing.
+
+**Jerod Santo:** Well, one place I'll break from you is that I've in my career moved from Ruby over to Elixir about 10 years ago, and the pattern matching properties in Elixir allowing me to match on function invocations, and have multiple functions of the same name, with different arguments coming in have really simplified, in my experience, my maintenance of branches... What I call like inconfident code; defensive code, that's what I used to call it. It's like, so much of your code, especially at the top of a function, is like making sure you didn't end up with a thing that you didn't want to end up with, you know? And of course, all the TypeScript people are yelling right now, but... I like that more than if/else. But what are your thoughts on that? Have you worked in languages that have pattern matching at the parameter level?
+
+**Carson Gross:** I haven't worked in a language that uses pattern matching as the primary dispatch mechanism. It looks a little weird to me. I'm just a Java guy.
+
+**Jerod Santo:** It did look weird at first, yeah.
+
+**Carson Gross:** Yeah. And so I'm just a Java guy... I don't have a problem with it, though. I think it's probably very effective. And when I say ifs and loops, I think that's more of a spiritual ifs and loops than like the actual ifs and loops... \[laughs\] You know, this idea of just like keeping it simple. And in the Java world, which is where I spend most of my time, not falling prey to this temptation of a complicated object model, when you can just have an if statement that does the same thing...
+
+**Jerod Santo:** Gotcha.
+
+**Carson Gross:** ...you know, like five lines of code with an if statement, instead of like three classes with an interface, and doing dynamic dispatch... Which, again, Clean Code would recommend -- they say explicitly, if you see an if statement, think to yourself "Could this be dynamic dispatch?" And I would recommend the opposite of that. I would say if you see dynamic dispatch, think to yourself "Could this be an if statement?" Because you'll reduce the total cognitive load by getting rid of all these additional crappy, small classes that you've got, that are just there basically to do an if statement. I've never worked with Elixir, or Haskell, or these languages that really prioritize pattern matching as a dispatch mechanism, so I can't have a strong opinion on it.
+
+**Jerod Santo:** Yeah, fair.
+
+**Carson Gross:** But I know a lot of very smart people who love it, so... That seems indicative that there's something very good there.
+
+**Jerod Santo:** Yeah. For me, it's like the only way I've actually felt a lot of my conditionals melt away without huge cost. Because everything's still right there, the function name is the same, and you just have like a list of handlers, effectively.
+
+**Carson Gross:** Right. It is an if statement, at some level.
+
+**Jerod Santo:** It is.
+
+**Carson Gross:** It's just a higher level way of doing an if statement...
+
+**Jerod Santo:** Right.
+
+**Carson Gross:** ...which I can appreciate.
+
+**Jerod Santo:** But it's not passing you off to some new class that just has like a -- you're just instantiating a new object and then calling this thing on it, which is basically what the Java story is. So maybe that's just the -- maybe at the function level is like the happy middle ground there. I don't know.
+
+**Break**: \[01:07:54.09\]
+
+**Jerod Santo:** Let's talk vendoring, the last one here, and I'll let you go...
+
+**Carson Gross:** Sure.
+
+**Jerod Santo:** But this is a topic that's near and dear to my heart, just because so much of what we do is dealing with other people's code, and how to do that well, and then failing to, or when to do it, how to do it etc. Vendoring, of course, is the concept of -- a pretty simple concept that was just like actually copying other people's code into your project, versus statically or dynamically linking to it.. And you wrote on this as well. Your thoughts on vendoring?
+
+**Carson Gross:** Well, I really like vendoring. I really like when libraries go out of their way to have no dependencies... And particularly when you talk about JavaScript libraries, like htmx... htmx is designed so that it can be added to a website just with a script tag. You don't need to adopt a crazy build tool. I've heard very good things about \[unintelligible 01:09:50.01\] but still, your average person is -- like, I just want to put a script tag in the header and be done with it.
+
+**Jerod Santo:** Right.
+
+**Carson Gross:** And I think some of the programming dependency management was an important advancement in computer science. I don't want to downplay it. But I do think we've maybe gone a little overboard, and I think some communities have gone more overboard than others, and the JavaScript community is one of them... Where you add a dependency and then a thousand other dependencies get dragged in, and now you have to have a pretty sophisticated build system in order to make that software work.
+
+And in my experience, for example, for some reason, this seems like such an easy problem, but I've used three or four different static site generators in my life, and they're just like, whenever I go back to these old sites and try to upgrade, it's just - everything's broken, always. The cross dependencies, and... Boy, I've just had so many problems with upgrades like that. And so to me, I think library developers would be -- they would do their users a favor if they weren't... Libraries in particular. Apps are sort of their own thing. Like, you've got your dependencies and it is what it is. But if you're building a library, I think it behooves you to try as best you can to have as few, or ideally no dependencies beyond just -- and if you've got to pull in some code and write it yourself or whatever, then just write the tests for it. And obviously, that's not always realistic. Like, you're not going to do your own SSL library, or something like that, but...
+
+**Jerod Santo:** Right. Do not hold your own crypto.
+
+**Carson Gross:** Right. But typically, that stuff's provided by the system, right?
+
+**Jerod Santo:** Yeah, modern systems.
+
+**Carson Gross:** Like, most of the stuff you shouldn't write is typically provided just as a base library by the system. So particularly in the JavaScript world, I think a lot of libraries could -- and I forget which library it was, but there was a guy on Twitter talking about how each release of his software, he had fewer and fewer dependencies. And in release five, he got this zero, and it's like, that's so awesome, because now your software can be used without all this other complicated stuff. And yeah, the JavaScript community is comfortable with \[unintelligible 01:12:07.21\] or whatever, but there's a lot of people that just want to do web development. They don't want to deal with all that.
+
+\[01:12:12.28\] And I actually think -- so you talk about vendoring... The idea of vendoring is just copying the code into your codebase.
+
+And I think there's actually a big opportunity right now for a vendor-first dependency management system that keeps this idea of transitive dependencies. There are going to be times when a library is going to need to depend on another library, and figuring that out is difficult, and that's a big value add that these dependency systems do. But I also think that you could do that, and instead of having like end products, like compressed JavaScript that's unreadable, or whatever, or in my case just Java jars, whatever, that you have to go and download the source code from somewhere else...
+
+**Jerod Santo:** Sure.
+
+**Carson Gross:** Instead, what you could do is you could have a system like that that downloads the source into, say, a lib directory, and uses some naming convention to keep everything, but just has the source of the libraries there. And then when you compile your project, you compile your project, and you compile that source all at once into the final product. And what's really nice about that is then you can check that source in, that can be part of your project. People don't have to have that dependency management system on their system in order to just check out and build your project, and run it.
+
+And the other important thing - and I think this will become more important over time - is now you can see all the code that is in your project due to transitive dependencies. And so it's going to be a lot easier for security analysis, and trusting -- you'll be able to build your app and then ship it to production without this - what to me is pretty crazy... Like, when people do builds in production, and they're actually doing dependency resolution in production, and then shipping that out, it's like, you're trusting a remote system at the very last minute to give you the exact same code that it gave you previously. And that seems like a really bad security and just general stability idea to me.
+
+To me, I'd love to -- I wish Java had this, where I could just say "Okay, these are libraries that I depend on. Figure out all the transitive dependencies, and then put all that code in my codebase, say in a lib directory." And then if I've got questions about how something works, the code is just right there, and I can do an audit of it, and make sure there's no inappropriate file system usage, or anything like that, very easily.
+
+**Jerod Santo:** Sure.
+
+**Carson Gross:** So I think there's a big opportunity right now in what I would call vendor-oriented dependency managers, where you have a dependency manager that focuses on delivering source to the project, instead of sort of final products, like jars, or compressed JavaScript, or whatever.
+
+**Jerod Santo:** Right.
+
+**Carson Gross:** Sorry, that's a little bit of a rant, but I've been thinking about that...
+
+**Jerod Santo:** No, that's interesting. Ruby on Rails used to do that. I wonder if it still does. It wasn't the happy path, but like you can literally -- I think that's where the term vendor came from, is you can literally type like "rails vendor" and then something. And instead of putting them in your gem file and then having a gemfile.lock, and downloading them to wherever they are on a system, it would actually unpack all that source code into your vendor directory. And that was good for deploying it, because it was all right there, but it also has some drawbacks. Like, if you're going to check it into your history and stuff, every time you update your dependencies you're having these huge splats of like files updated, and it can be a little bit gnarly in that sense. So I think there's pros and cons, but I do think it's an interesting way of going about it.
+
+**Carson Gross:** Yeah. Well, I remember the Rails way of doing it, and I remember it being a little janky, and not working...
+
+**Jerod Santo:** It was.
+
+**Carson Gross:** Because it just wasn't like the focus of it. And then the other thing I'd say is computers are pretty fast... So big, janky check-ins are okay at this point.
+
+**Jerod Santo:** Sure.
+
+**Carson Gross:** I don't know, I feel like I would rather see what changed with my dependencies, and have someone who's in charge of managing that, and maybe is in charge of the dependencies and doing that work on their own machine... Than the situation now, where I just like get new random jars, and I don't know what's in them, or what's changed, or anything like that.
+
+\[01:16:17.15\] I do think there are a lot of trade-offs. I think there's an opportunity there. You'd have to do it right, and you'd have to be smart about it, for sure, but I think there's an opportunity there. And I would couple that with this general plea to library developers to reduce the number of transitive dependencies that they bring into projects as well.
+
+**Jerod Santo:** I think in the JavaScript community that's a growing trend over the last 5 or 10 years, where I've seen especially -- you know, people pitch us a lot of libraries for coverage and conversation and stuff... And dependency-free is now a marketing term, which - I appreciate that one. There's some marketing terms, which I think are silly. Modern... Everybody puts "modern" on there. It just means they just got finished writing it... But dependency-free has actually become a term of art that people will put with their pitch. Like "Hey man, no dependencies." And to me, that makes it more exciting. I'm like "Cool, I can more easily recommend this tool, just because I know I'm not dragging something else along for the ride unbeknownst to me", you know?
+
+**Carson Gross:** Yeah, exactly. And I have to point out again; this is one of those situations where it's like we had bad vendoring where you had to manage everything yourself, and then we went to transitive dependencies, and everyone's like "This is great. I can depend on this", and then it spidered out of control. And now maybe there's this moment when we can sort of go back in between those two and say "Can we have some of the dependency management system, but then some of the vendoring approach as well?" I think that's something that's available to us now. We have the technology for it and the experience to do something like that, so I hope someone builds that. I do not have time to do it, but I hope someone does.
+
+**Jerod Santo:** I was going to tell you, write your pitch deck, Carson. Get that pitch out there. Somebody will fund it.
+
+**Carson Gross:** \[laughs\] Nope.
+
+**Jerod Santo:** So real quick, htmx 2.0... Was there anything to that? Was it a big thing? I remember emailing you when it happened and you're like "Well, it's not like brand new or different." I mean, there's some things, but it wasn't a big deal. Is that the story?
+
+**Carson Gross:** Yeah, we cleaned some stuff up and, and changed basically some defaults in it. And maybe we didn't do quite enough of that. There's a couple of things I look at and I'm like "Man, I wish I had changed that in 2.0", but...
+
+**Jerod Santo:** So 3.0 coming soon? \[laughs\]
+
+**Carson Gross:** No, no, we're not doing 3.0. You know, there's an essay up on The Future of Htmx, on the website, and I talk about how I think htmx - there are mistakes that I made in it, for sure. There are aspects of it that I would change if I could go back in time. But it's not wrong enough that I'm willing to break backwards compatibility to fix those things. Like, they're just small, ugly issues or implementation issues that like "Uh, this would be simpler if I used fetch instead of xhr, or something like that; that'd be a little bit cleaner."
+
+And so I really just -- I think htmx, the conceptual idea of it is strong, and the little grunkiness that exists around the edges is not worth breaking everyone who's using htmx stuff just to clean up for philosophical reasons. And so we've thrown a lot of config options in there to turn on and off things. Like, that's one way that I try to deal with things if I don't like how something behaves. But it's the default, I'll put in the right way, but then put it behind a config flag, so people have to opt into that behavior. So that makes htmx a little more grunky than it would be if I was more willing to break backwards compatibility, but I think backwards compatibility is a big feature. And so the jump from htmx one to htmx two is very gentle for most people.
+
+There were two or three config options that you could flip, and basically have identical behavior with 1.0. That's the philosophy and sort of vibe of htmx. Like, okay, we're generalizing hypermedia controls, this is what it does... We're not going to try and reinvent that. There's no htmx server components or anything like that \[unintelligible 01:20:19.00\] It does what it does, and that's it. So I think that it's a different way of approaching software for sure, but I think there's advantages to it as well.
+
+**Jerod Santo:** Right on. Well, if you want more from Carson, "Hypermedia.systems", the new book. The revolutionary ideas that empowered the web. Read it for free online, buy the hardcover to support the authors. Very cool.
+
+**Carson Gross:** It just came out in Japanese, by the way.
+
+**Jerod Santo:** Oh, nice.
+
+**Carson Gross:** \[unintelligible 01:20:46.16\]
+
+**Jerod Santo:** There you go. Anything else, Carson?
+
+**Carson Gross:** For all your Japanese listeners.
+
+**Jerod Santo:** Excellent.
+
+**Carson Gross:** No, I think that's it. htmx is the big thing. The hypermedia systems, those are the big things right now, so check them out. Oh, gregbrain.dev if you want my humorous take on software development in like a caveman voice. You can check out gregbrain.dev.
+
+**Jerod Santo:** Alright. All the links to all those things are in your show notes, so check there, click through, and connect with Carson. Good stuff, appreciate your thoughts, appreciate your work, and putting it out there for free for all to have. That's always awesome. Carson, I appreciate you coming on the show.
+
+**Carson Gross:** Thanks a bunch for having me on. I appreciate it.

The Web Development Engine (Interview)_transcript.txt

@@ -0,0 +1,487 @@
+**Jerod Santo:** Today, we're joined by Andreas Møller from Nordcraft. Welcome to the show.
+
+**Andreas Møller:** Thank you very much.
+
+**Jerod Santo:** Happy to have you. I first found you from this blog post called "They Lied to You. Building software is really hard." I found this to be very relatable... And one of the things that you said in that post I'd love for you to expand on for us, is you said "If you're looking for one trick that lets you get ahead and jumpstart your career", your advice to the reader is "Don't choose the path of least resistance." Can you unpack that for us? Why? Why not? What does it mean?
+
+**Andreas Møller:** Yeah, absolutely. A little bit of context... Obviously, if you have a job and you've been tasked with building something, you don't always have to go the hardest possible way of doing everything, but I think there's something to be said for leaning into the challenges when they present themselves. So if your solution to every problem is "Is there a library that does that?", without looking into how to actually solve the problem. Or if you're the kind of person who likes to hang around Stack Overflow and see if there's an existing solution to everything. Like, it's good to look it up, but I always would say give it a try first.
+
+Throughout my career, especially with a lot of junior developers, there's this sort of attitude that "Oh, I'll go and look here, I'll go and find it here", and a lot of the times, the part where you really learn something is when you sit down and take your time. So actually leaning into it, saying "This is a hard problem. It's going to be difficult to solve, but I'm going to try and figure it out. And I'm not going to look for the solution until I've at least analyzed the problem and feel like I have a good idea."
+
+**Jerod Santo:** And you're here now today, building a thing called Nordcraft. How did you find yourself here? And you talked about your experience... What does your experience look like, how did you begin building this, et cetera?
+
+**Andreas Møller:** So Nordcraft sort of came about a little bit out of frustration. I've been a software developer for a little over 20 years now, almost my whole career focusing on web development, because I really like the web. I think it's the greatest application platform ever built. And I've really enjoyed building anything from small websites to massive SaaS platforms. And through my career, I've sort of gotten to see -- like, we've gone from there were no frameworks, to gone through a few of them... React has stuck around for a long, long time, and now we're seeing the new generation of framework coming out and improving on React. There's a lot of talk about how things change every six months in web development... And yes, something changed, sure, if you're looking for it. But on the whole, for the last 10 years, things really haven't changed that much. React has been the dominant framework... There's a little bit of modification, a lot of the details, but overall, the way we build apps are very much the same.
+
+\[07:41\] And so one of the things I noticed was this process when you're sitting in a product team, where a designer will -- usually, a product manager will come and say "We need this kind of feature, because..." For some reason. A designer will go and do a layout or a design of what that needs to look like, and then it will get handed off to the developer, which was me. And the handoff is something along the line of "Here's a picture of what you need to do. Please do it exactly the same way." We usually call it pixel perfect, right? And pixel perfect means don't add anything of yourself into this task whatsoever. Just do the exact thing that's been shown you, but in code. And I like my job, but that particular part was the least favorite part, because that's the bit where I don't really matter. I just need to repeat the task someone else did. My opinions don't really factor into this. Now, in actuality they do, because the design is never complete, but I always felt that part was kind of pointless. Like, if the designer already built it, why are we doing it twice? And so I wanted to play around with this idea of saying "Well, why do we duplicate this step?"
+
+And I've seen tools like Webflow, and you can build visual tools for building websites, but there wasn't anything built for applications. There's sort of a limit, and like the second you get above a certain complexity, or you need some flexibility, some dynamic content, then we're going back to everything is done in code. And so I wanted to figure out why is that? If the UI can be done visually, why can't we build something like this, that would work for the kind of application I'm building? ...which was mainly SaaS applications.
+
+**Jerod Santo:** That sounds an awful lot like Drew Wilson's thing he's working on, Adam. I just listened to your conversation with Drew Wilson, and he mentioned he's working on something similar, where the design manifests in DOM nodes, or something. The design tool is actually producing the DOM, versus --
+
+**Adam Stacoviak:** That's the big thing he wants to work on. Yeah.
+
+**Andreas Møller:** I don't know if I know -- what is it called?
+
+**Adam Stacoviak:** It's not named, I don't believe. I don't think it has a name yet.
+
+**Jerod Santo:** I think he said something like Opacity, or something like this...? Anyways, it's not like out there yet, so you're way ahead of him, Andreas, don't worry about it. But...
+
+**Adam Stacoviak:** But he is a formidable opponent.
+
+**Jerod Santo:** He is a formidable opponent, that is for sure...
+
+**Andreas Møller:** There are a lot of tools in this sort of broader space, of sort of trying to make programming visible, visual. There's a lot of tools. What I thought was really interesting was that this sort of idea is not new at all. The broad idea of visual programming is very old. The oldest thing I know of is like 1967. So before we had personal computers, someone was drawing with a light pen on like an old -- like, what's called the \[unintelligible 00:10:35.11\] And they were drawing those together and programming by drawing on a... I mean, a tablet is what they called it. It was a very large monitor, and it'd take up most of the room.
+
+**Jerod Santo:** Right. There was no iPad.
+
+**Andreas Møller:** And it's been tried a lot of times before, but it never -- like, Visual Basic worked and took off a bit, but then it hasn't really taken off. It's always died down again. Except in one industry, which is gaming. And I thought "That is the weirdest thing in the whole world." So if you're building a AAA game that takes 10 years, and takes some of the brightest minds of our engineering industry, then you can use a visual tool to build most of the game. But if you do it in 2D, and it's just a website, then that doesn't work. And I'm like "That doesn't make any sense. There must be something I'm missing here. There must be something I'm not seeing", for why it is that that's the only industry where not only is it possible, it is an absolute necessity. Without those tools, you cannot compete.
+
+**Adam Stacoviak:** Yeah, that's interesting. I've never developed a game before, so I don't know that for sure and how it works, but that does make sense. I mean, as a designer -- so I came up in design, came up in user experience design, product management... I was like a product owner for the company, leading and... I was that person between the business itself and the engineering to make sure the business goals were met, and engineering was able to accomplish the goal... A lot of my life.
+
+\[12:15\] And I would say, on the flip side, it was a struggle as a designer to have to do all this in this pixel perfect way, in pixels, literally, and not in the web, and not have the tooling available. Sure, you can HTML code it, and you can use CSS, but it was just, you could make more progress, more efficiently, in higher fidelity, the oldest of day, in Photoshop, lately Figma, or even Sketch if you're not a Figma person... It was always a challenge to design something like that, to go through all the process, to go through the customer and the user stories, and talk to people, and talk to customers, and like gather all that feedback, to then create an artifact that is static, and attempt to hand it over.
+
+**Jerod Santo:** And hope it looks like that...
+
+**Adam Stacoviak:** And honestly, because you'd done all that work, because you've done all the thinking behind, you're like "Don't change anything, okay?! This is a brittle thing I'm handing you. Please don't change anything, because we've gotten signoff from here, we've gotten business concern handle there, we've got colors established here..." Hopefully, it was that thought through. But for the most part, it was pretty good and thought through, "Don't change it, please" and "Please, can we make this happen?" And so I've been on that side, and it's such a... It's tough for both sides, really. I think we all want better tools.
+
+**Andreas Møller:** It really is. And the interesting thing is my co-founder - his name is Kasper - he is a designer. And his motivation for joining -- like, when I showed him this the first time, he almost to the word told me the same thing. He was saying, like, the frustration... Because it's not just that it's limiting, but it's also the frustration of then seeing what you've done interpreted through someone who thinks blue is blue.
+
+**Adam Stacoviak:** Right.
+
+**Andreas Møller:** And there is white, and I'm not sure what eggshell is. Right? It's just white, isn't it? And then see someone like that take your design and then say "Look, I made it." And then "No. No, that's not what I did." And having to wait, and then come back, and then we have to do a whole iteration again... I have to go back, critique all the things, and lastly say "That's the color, right?" And yeah, it's super-weird.
+
+I'm a big fan of all the JavaScript frameworks. Actually, all of them. I think they're all great, to be honest. And in many ways, we've moved forward, because we were able to build more complex things. But one thing we really lost is it kind of put the final nail in the coffin of "Should designers code?" Before then, there were designers who worked in HTML and CSS. That's more or less not possible today. Because it's not just that you need to know HTML and CSS, you need to know Vite, you need to know how to install that that npm module that never wants to work, or figure out whatever happened to the build system since last when we cloned things from GitHub... The complexity of our builds, of the setup around building an app means that before you get to touch anything design-related, you're like three or four days in.
+
+So we really sort of -- with the technology stacks we use today, we've moved even further away from the idea of a designer contributing anything directly to the application. I think that's probably the worst thing that happened with modern frameworks.
+
+**Jerod Santo:** \[15:39\] Yeah, that does make it tough. I'm curious about -- perhaps a pivot that I'm sniffing out here, because when I first linked to you, which was March 27th, it was on toddle.dev. Now it's Nordcraft. I remember Toddle talking about a game engine for the web, or something like this... And Nordcraft itself is THE web development engine for web apps that delight your users. You talk about creating web experiences. A web game, a web experience - these things are very similar things... So I can see it being just as maybe a pivot of sorts, or perhaps just a rebrand. Can you tell the story of Nordcraft? We're kind of branding nerds. We think about naming, we rename things... And so it's more just intellectual curiosity there than anything else.
+
+**Andreas Møller:** It is a rebrand, it's not a pivot.
+
+**Jerod Santo:** Okay.
+
+**Andreas Møller:** But it's a quite thorough rebrand.
+
+**Andreas Møller:** Yeah.
+
+**Andreas Møller:** The product, we always kind of knew what that was going to be. There was never much discussion on that. I invented this for myself. My co-founder, coming in, knew what he wanted to use it for... And the point was to build something for professional product teams. What we looked at and saw - there was this whole no-code community. And no-code tools are generally -- they're somewhat similar, but the big difference is that they're focused on people who don't come from a developer background. And generally don't want to be developers, in like they don't want to spend time learning to code. They are not interested in getting too deep into the technical part of software development. They kind of want a tool that takes care of most of it for them. A lot of them are very interested in vibe-coding right now, for the same reasons. It's about getting a product out, but the process of building is less interesting. A lot of developers often are actually drawn sometimes even more to the process than to the final product, right?
+
+**Jerod Santo:** Sure.
+
+**Andreas Møller:** And the way that happens in no-code is you try and simplify. You try and figure out "How can we take a process and something that's very complex and how can we simplify it down to a point where people can get started earlier?" And that generally just happens by creating high-level abstractions. What we would normally as developers call massively over-abstracting. It's just like, that's what you're building. And that comes with a certain set of limitations. You can build this kind of table component, but you can't build that kind of table component. You can make it behave like that, but you can't quite make it behave like that. And as long as you're okay with those, as long as you're saying "I'm solving a business goal; the specific details of what it looks and feels like aren't important as long as the business goal is solved", those tools are really great. But that's not what we wanted to build. We thought there was a space for us in that market of saying "Well, we can be in the professional segment of that." Once you hit a certain limit with those tools, we're right there for you to move up. And what we didn't quite realize was that's not that interesting for people, again, because most of them either already moved up and hit the limit, and then their next step is learning code, or they're just not interested in taking that extra step. And at the same time, I think we underestimated just how much a lot of software developers do not like no code, both as a term and as an industry.
+
+**Adam Stacoviak:** Like NoSQL.
+
+**Andreas Møller:** Yeah.
+
+**Adam Stacoviak:** Right?!
+
+**Andreas Møller:** Almost as much as they dislike NoSQL.
+
+**Adam Stacoviak:** I mean, that was a big kerfuffle back then, too.
+
+**Jerod Santo:** Oh, for sure.
+
+**Andreas Møller:** Yes. And in many ways, honestly, it's not that unlike the whole Mongo story of "Look, we don't need SQL, because look how simple this is", and then you're like, now you need a transaction, and then you go "Now I have to invent transactions."
+
+**Jerod Santo:** Right.
+
+**Andreas Møller:** So yeah, for sure, there are certainly similarities. And to some degree, that's what the blog post is about. When you always seek the simplest and the easiest solution to every problem, what you end up doing a lot of time is you put yourself into a corner where it's really hard to get out of. So I've done that. I've chosen Mongo for a project that really needed a transactional and a relational database... And then gotten to a point where I need to figure out how we kind of build our own transactions, so we could roll back at the application layer. And I think we see the same thing in the no code space a lot of times, and I think a lot of developers will have experienced that.
+
+**Jerod Santo:** \[20:18\] Sure. And all the vibe coders will certainly experience that, assuming that their idea is good, and they get past that first phase, which is -- I think it's great for getting that proof of concept, something to show somebody, get your idea out on paper... But extreme limitations, at least at this phase, the current state of the art.
+
+So Toddle to Nordcraft. This was just a rebrand. What was Toddle holding back that Nordcraft unleashes?
+
+**Andreas Møller:** So the name itself was a bit problematic, because Toddle is sort of a relation to toddlers and toddling... Which was actually a joke name my wife came up with in the very early phases. Because when I started working on this, I had no idea if it even could work. There was a part of me that thought "Since we don't have these kinds of tools, there's probably a really good reason why, and I just don't know what it is." But I had to try and figure out, so I kept working on it. And to some degree, when it came up, I'm like "That's kind of what I'm doing. Right now I'm trying something, and then I'm falling over, and then I'm trying something else..." I'm like "Yeah, okay, let's call it that." And then as it started working and we kept referring it to Toddle, and we kept going, eventually then VC investors came in, and then we had to talk to all of those, and that was a bad time to change the name... There was never a time to change it after that. We're not going to sit down and come up with the best name, because it's like, it was fine.
+
+And there's a certain truth to the fact that a name is whatever meaning you give it, eventually. So had we kept the name, eventually Toddle would just mean our app, and it wouldn't matter, because that would be the ideas people had. But we could feel it was holding us back a bit from anecdotes from customers. And at the same time, we couldn't get a dot com... It's hard to search for, because we get so many other results... So there's a lot of small things about the name that wasn't so good.
+
+So when we realized that we needed to make a branding pivot away from being a no code tool, and a more professional tool for beginners kind of thing - which is also a weird phrase to say out loud - we said "Well, let's take the name away as well." Because that at least -- everything will still redirect to the new site. But when people, say, think like "This is what I think about, Toddle", versus this tool that you really have no reason to compare us to, because we serve completely different sets of users... And a lot of that is sort of existing conversations we could actually a little bit leave behind. And that was very intentional. So we wanted a brand that sort of highlighted that this is meant to be for professional teams, it's meant to be for developers who know what they're doing.
+
+**Jerod Santo:** Well, Adam seems to like it. I think I heard you say it's beautiful. Is that what you said, Adam? I think Andreas missed it.
+
+**Adam Stacoviak:** The rebrand looks beautiful. I mean, the application looks beautiful too, and I was only able to find it because, thankfully, you were able to clone an application. So I'm hanging out at toddle.dev/project slash a bunch of slugs. Nordcraft is the next one after that, and so I'm actually like hanging out in the UI. It's all beautiful. It looks really nice.
+
+**Andreas Møller:** Thank you. Thank you. That's my co-founder, who's the biggest -- well, I mean, it's not just him anymore. Right now we have a whole team. We have designers that are looking at that. But that's not my fault. There are old versions, pictures of the first versions, and that looks very much not like it does today. But yeah, I'm really happy about the work they put in there. I think it looks great. And again, the new design has a certain -- it has a little bit more gravitas to it. It looks a bit more serious.
+
+\[24:07\] The other problem we sort of ran into is that we are not very serious people, my co-founder and I, and we love this part of startup life, when you can do ridiculous things for attention grabbing, and like send out content that is ridiculously silly, and people enjoy it... But when you stack that on top of a childish brand name, and everything being bright, childish colors, it becomes too much. So when we have a serious brand, actually being a little bit silly as founders isn't the problem. So I think that all aligns quite well with what we're trying to do.
+
+**Adam Stacoviak:** Yeah, I'd actually -- I heard about Toddle by way of a friend through friends, I would say... Salma Alam-Naylor, I'm sure you know her.
+
+**Andreas Møller:** Yes.
+
+**Adam Stacoviak:** Amazing DevRel, amazing all the things... I admire her so much. She's so good at speaking, and so good at demoing, and just thinking how users feel, and like that intersection there... She's really just a star. So I'm sure you're glad to have her, but... I was like "Man, Toddle sounds pretty cool", but the name didn't really make me feel like "Oh, I should --" I don't know, it didn't spark my interest, let's just say.
+
+**Andreas Møller:** It doesn't make you go "Oh, I need to use that for the next tool that I maybe focus my career around in the future."
+
+**Adam Stacoviak:** It seemed like a toy. It seemed like whatever it is, it's cool, it's groundbreaking, but it's almost silly, just by name. That's it. The application - obviously, different scenario. Now, if I'm being honest though, Nordcraft is almost as ambiguous. Almost.
+
+**Andreas Møller:** Oh yeah, but that's kind of on purpose. We don't want it to mean anything, really.
+
+**Adam Stacoviak:** Well, I mean, Nordcraft sounds like buildings...
+
+**Jerod Santo:** Nordcraft?
+
+**Adam Stacoviak:** Yeah.
+
+**Jerod Santo:** I mean, the crafting part makes me feel like this is for crafters. This is for builders.
+
+**Adam Stacoviak:** Maybe.
+
+**Jerod Santo:** I also think of Minecraft of course, because it's such a strong brand in the crafting space...
+
+**Adam Stacoviak:** True.
+
+**Jerod Santo:** So there's a bit of fun there. Nord - I wonder what it is. It sounds cold, Nordic. I don't know. I've been to the website, so I've seen some of the trees and stuff, so you do have like the forest thing going on... I like it. I think it's better than Toddle. I agree that it leaves -- if you just said Nordcraft to me, I wouldn't know what it is.. But you know, back when you first said Google, I didn't know what it was, because I didn't know what a Googleplex was back then.
+
+**Andreas Møller:** One of the things we've found was that trying to explain what it is was extremely difficult, because everyone has -- especially when you're doing marketing content, people have a desperate need to find a bucket and put you in it.
+
+**Jerod Santo:** Yeah.
+
+**Andreas Møller:** And if you don't give them one, they'll take one. So I once had an entire 20-minute conversation with an investor, that never understood that we were not like Airtable, even though we do not have a backend or any database associated directly with it whatsoever. But he just couldn't -- he couldn't let go of that idea, because once he sort of tried to represent \[unintelligible 00:27:27.00\] it was locked. And when we say "visual editor", if you think it's kind of like Webflow, it's technically not wrong, but it doesn't quite communicate the important part, which is this is actually a tool that teams will use to build very complex applications. And Webflow is for building mostly marketing sites, and it's predominantly not used by developers.
+
+\[27:57\] So even though there's a lot of similarities, it doesn't quite capture the point. And what we've found is actually, we think the best thing we've come to yet is this idea of a web development engine. It's essentially like "What if we had game engines, but designed for building web apps?" Not web games, but that same idea, that same concept of tooling that you have in the game world - how would that look if we built that for the web?
+
+**Jerod Santo:** So go ahead, explain the tech to us, exactly what it is, how it works. We're all developers here, so... Hopefully, we don't get completely locked into something that is not.
+
+**Andreas Møller:** No. So, I mean, probably the simplest way of -- again, we use a lot this analogy to the game engine. A game engine is sort of a big box of a lot of different things. It has rendering engines, physics engines, audio engines, animations engines... All these things. And then that's all tied into a central framework for building games, that sort of powers the game loop and everything. And then on top of that, they have this suite of visual tools. And most of the people working on them are like a mix of some developers, a lot of designers, but they're all working the same tools. And these tools essentially program the engine.
+
+And we're not just talking like building characters. It's not just that they're using visual tooling for like 3D modeling etc. Most of the game logic is built in visual scripting. So in Unity it's called blueprint -- no, in Unity it's called... I think it's just called visual scripting. In Unreal engine it's called blueprint. And I didn't realize this at the time when we started, but actually, the AAA games today, most of the game logic is built with visual programming... Which for me was such a mindblowing moment, because we just rejected this concept, every other industry. We're going "Okay, that's cute. That's for Sketch. That's for children."
+
+**Jerod Santo:** Right.
+
+**Andreas Møller:** But when we look at these big, massive games we play, they've been built this way. Like, there is a lot of code involved, there's a lot of really talented engineers working there, but a lot of the logic is built using these visual tools. And so the basic idea was "What if we build web apps the same way?"
+
+So the best way of describing it on a technical level is probably that it is, in effect, a programming language with a built-in framework, but the biggest difference between -- like, if you're familiar with Elm, that's actually a really good analogy as a programming language.
+
+So Elm was this programming language with a built-in framework. The Elm architecture was the inspiration for Redux, when that came up... But it was a whole language built just around the idea of building frontend apps. And Nordcraft is actually the same, but instead of the code part -- normally, when you have a programming language, you have the code, that gets passed into an AST, which is this abstract representation of your program, and then that gets compiled to either code, or JavaScript, or whatever it is the engine you need to run it in. And all we did is we sort of cut off the code part, and we were saying "What if we just design the big syntax tree of what our program is going to be, and then we store that in a database? And then we build an editor for manipulating that tree."
+
+And what that means is a lot of the things -- and that editor is where normally a visual program is like a big board of like a million nodes. We have some of that too, but what we've done is saying "Different problems have different needs in terms of how you build a tooling for it." So if you're building UI, obviously the editor is going to look like a UI editor. You're going to drag HTML elements, and you're going to click them, you're going to apply styling. And because we wanted to build for professionals, very early on we set on this path of saying "We're not going to try and reinvent new abstractions here", because that's really hard, and it usually takes a long time. And there's a reason why we haven't seen new abstractions since probably React, or Angular time. All the new frameworks are the same abstraction.
+
+\[32:15\] So we're not going to try and invent a new abstraction. We are just going to create a very nice UI for working with the web platform. HTML, CSS, everything you set in our style panel maps to a CSS property. And we even in our tool tips show you what the CSS property is, and what the CSS is going to be rendered as. Because we want you to have full control.
+
+So the HTML you see inside Nordcraft's visual editor is going to be the HTML we render for every component, on every page. The CSS you see is the CSS we render. We do hash some class names, but that's about -- other than that, you basically see exactly what it is.
+
+And then different problems have different solutions. We have a workflow editor for like "When I click a button, what do I actually want to happen?" That becomes almost a flowchart. And it turns out that -- and we sort of separated, we made a distinction between what we call actions, which is I do something that's like I call an API, I update a variable, I trigger an event... Those kinds of things. Things that we in functional programming say have side effects. And pure functions. So pure computation, we call that formula, anything that has a side effect \[unintelligible 00:33:35.15\] And those are actually two different editors. And that allows us to sort of visually separate the two, which turns out is really nice, because you get the high level structure of what is it this thing does, and then you can dive into the details much easier. So it becomes a naturally nice way to organize your code, of saying "Take all the computations and actually hide them initially, and just show me what are the steps. And then if I want to dive into why you did that step or why you did that other step, then I can go and do that." And I think it's not far from how people often organize code anyway. But each of those kind of have their own UI. They're a separate problem. The editor for doing pure formulas, pure computation looks very different than the editor for doing a flowchart. And completely different from any of the UI stuff.
+
+**Adam Stacoviak:** When I'm poking around the UI, I see the option to add formulas and variables. Is that what you're talking about when you say flowcharts, and stuff like that? That's pretty interesting, because you've got different types, and integers, and data coming from different places that connect via nodes. Is that what you're talking about?
+
+**Andreas Møller:** So when I say flowcharts, obviously it's a bit abstract concept to apply directly, but I'm mainly referring to what we call the workflows. So if you select an element, there's an Events tab, and there you can click on the "click event", for example. And that sort of shows up this dialogue that says "Okay, what do you want to happen when someone clicks this element?" And there you can add nodes. And that's a different UI. It looks a little bit like the formula UI, but it's actually tipped on a different axis to sort of make it distinct.
+
+And so we separate, like, what do you want to happen when I click? Okay, so I want to update this variable. But the value you set on the variable - that's probably something you compute, most of the time. And that computation is a formula. So in that, you open the formula. And by having those two separate, you actually - you get to a point where it's very easy to get a high-level overview of what's going on, and very easy to then dive in to exactly the point you want. And one of the benefits we see with that is also that when you're working in the same tool, both engineers and designers, designers can sort of opt into how much they want to get into. When my co-founder joined the project, he only touched the UI. He didn't touch any of the logic, none of the data fetching, none of that. It's just, "I'm only looking at the UI, I'm styling it", and he was super-happy about it.
+
+\[36:17\] And it worked really well, because effectively we worked twice as fast as we would normally do, because we weren't doing all the UI work twice. He would just design directly in Nordcraft, I would just actually wire up the functionality. A lot of times I would build it first and not style it at all. Just add like the minimum amount of UI to make it functionally work. And then he would take over afterwards, and actually go on shipping it. But as a designer, you can choose "Do I care about how the formulas work? Do I care about the workflow? Do I need to know?" And initially, the answer is probably no, and over time, what we often see is that you can sort of get a high-level idea. "Oh, okay, so this is where it updates the variable. This is the button that eventually makes us fetch the data from the server." That kind of thing.
+
+Or if you want to, you can go really into "Why is it that decision? Why did it do exactly that thing? And why was this the value that was set?" But if you're not a programmer, and you're just coming into this, you sort of have the option of saying "I don't really need to deal with that yet", but still get an idea.
+
+**Jerod Santo:** Well, I also know how the editor open and I'm clicking around as we talk... It's very interesting. It reminds me of -- well, like Pixelmator, or Figma, with more things in there for like... Like actions, for instance. Is it baked? I mean, is this thing ready to rock? Are people using it? Is it -- how long you've been working on it? And is it -- I mean, nothing's done, but is it like the 1.0, is it ready?
+
+**Andreas Møller:** It is. It is super-ready. We've launched a year and a half ago.
+
+**Jerod Santo:** Okay.
+
+**Andreas Møller:** And we've got quite a few -- obviously, we don't have any massive enterprises that built the whole stack in Nordcraft yet.
+
+**Jerod Santo:** Right.
+
+**Andreas Møller:** Because a year and a half is not quite enough for people to discover a tool and then build on it for several years. But we have some really interesting projects that's built in Nordcraft. One of my personal favorites is a guy in Malaysia, who built -- it was essentially a competitor to Shopify. It's not feature-complete like Shopify, but in Malaysia, apparently, the way Shopify does card payments means that it's almost impossible to use. So he built a e-commerce store platform like Shopify in Nordcraft. And it has its own visual editor, where you can build your shop, which has a nice sort of inception concept to it. And I think he's already processed orders for over $10 million in that project. And I think on the first day he launched, within the first three days, he processed the first $10,000. So like immediately successful, and he's done a ton with that. And we have a couple of projects like that, that's really successful.
+
+A lot of people are building hobby projects and interesting things, things for fun to try out. Most of our users are always on the freemium plan, and just building interesting things. And of course, the biggest project, the whole editor that you're using right now - we're building it in Nordcraft as well.
+
+**Jerod Santo:** Gotcha. So it's like self-hosted in that way. What's the language? I mean, you said you created a language and a framework together...? Would I learn the language as I build things, or would I not have to know...?
+
+**Andreas Møller:** So language in this case... We haven't given it a name, because we don't really refer to it in that way. It's just the thing underneath, that powers the whole thing.
+
+**Jerod Santo:** But you never have to poke under the covers.
+
+**Andreas Møller:** No, we haven't invented a syntax for it. Like, the whole thing is just stored as ASTs and JSON.
+
+**Jerod Santo:** \[40:11\] So what about when I'm hooking up my actions? When you click on this, here's what I want to do. Go hit this API, get the results back, display them in a table. Am I coding in the language then? Or am I clicking on things? What am I doing?
+
+**Andreas Møller:** I think that's almost a philosophical question, isn't it?
+
+**Jerod Santo:** \[laughs\] It almost is, yeah.
+
+**Andreas Møller:** Because especially -- I used to say that it's not coding, but it's programming, just because coding for me was like code...
+
+**Jerod Santo:** Sure.
+
+**Andreas Møller:** But people use the word "code" a lot to refer to what they've done in Nordcraft.
+
+**Jerod Santo:** Yeah.
+
+**Andreas Møller:** And I'm -- I mean, I'm not sure it's wrong. But there is no textual representation of it.
+
+**Jerod Santo:** Okay. So I am clicking on things and hooking them together.
+
+**Andreas Møller:** Yeah.
+
+**Jerod Santo:** I'm never going to be like --
+
+**Andreas Møller:** So again, if you were writing a JavaScript function that says "Add two numbers together", what the browser actually does when it sees that initially, it starts passing the actual text, and then it builds up this tree structure in memory. And like the first node is going to say "This is a function node." And then it's going to have some attribute that says the name of it, probably whether it's async... I can't remember what the standard AST for JavaScript looks like, but something like that. And then it's going to say "And there's some parameters, or argument." That's going to be an array. And if you have numbers as input, it's going to show that you have parameter A, that's of type number, and parameter B of type number. And then it's going to say "And there's a block."
+
+So it basically takes your code line for line, or bit for bit, and try and represent it as a large tree structure. And every compiler does this; that's the parsing part, right? And then if you're using Babel, what Babel does is it just applies a bunch of different transforms on this AST. If you're doing TypeScript, that's again - like, TypeScript runs type checking on this data structure, so that you can run all these different processes against it. And then the final bit of that is if your language is a compiled language, it turns that into a bunch of ones and zeros; a complex process that I feel like we can leave out of the scope.
+
+**Jerod Santo:** Sure.
+
+**Andreas Møller:** Also, I don't really know anything about it, because I've never built a compiler. Or if it's an interpreted language, there's literally a program that goes through that tree and then performs all the different actions that matches what that is. So if you're saying, okay, if this node is a function call, and this is the name of the function, it's going to go "Well, then I need to go look at my function stack and find that function and apply it." And that's the same thing that we do. So the only difference between, in this case Nordcraft conceptually as a language and any other one, is that we don't start out with text. We cut that off. And this AST that the parser normally generates is actually our source of truth. And that's what we store in our database. And the editor loads that in, and has that model in memory when you're working with it, and whenever you edit anything, it goes directly into that AST and saying "I'm going to change this property." So if you rename a function, it's going to go find that function in the AST and say "I'm going to change the name property."
+
+**Jerod Santo:** So when you're building a formula, you are clicking, you were saying "add an if condition, here's a type of variable, it's a boolean..." You're like connecting things together. And so it's very visual, even when you're building formulae... Formulas? I'm not sure how --
+
+**Andreas Møller:** Formulas.
+
+**Jerod Santo:** Yeah, formulas.
+
+**Andreas Møller:** Or maybe we should do formulae.
+
+**Jerod Santo:** I think formulae might be the --
+
+**Andreas Møller:** I kind of like it.
+
+**Jerod Santo:** ...formal way of saying it. I know that because in Homebrew they have formulae, I think.
+
+**Andreas Møller:** Oh.
+
+**Jerod Santo:** And I always thought "Well, that's fancy..."
+
+**Andreas Møller:** Maybe we should start saying that.
+
+**Jerod Santo:** \[44:01\] Yeah, it's kind of cool. Anyways, I do find there's a button down there, because I'm building a formula, a test formula right now... And as a software developer, this is cool, but at a certain point you're like "I don't want to click and drag and do stuff necessarily..." And there's a button, "Convert to custom code." And now I'm looking at a JavaScript editor, with a JavaScript function. Is this kind of your solution for that situation, where you're like "Yeah, this thing is limiting me. I need to do something outside of the ordinary", and I can bypass the editor? Or tell me what's going on here when I click this "Convert to custom code", because there is an asterisk there about server side rendering.
+
+**Andreas Møller:** Yeah. So the whole idea of custom code - we call them custom actions, custom formulas.
+
+**Jerod Santo:** Okay.
+
+**Andreas Møller:** It's basically that if you're building a new language, you are missing a lot, no matter what you're doing. There's a lot of things you can do in Nordcraft without ever touching any code, but not everything. I often use the example of like - I wanted to build something that I could use... I wanted to try the Bluetooth API, right? And so I made a simple example where I could use a PlayStation controller to browse a website. And that Bluetooth was not very high on our priority list of things that Nordcraft needed to do, because I've never actually used before in my 20 years... To be fair, it hasn't been there for 20 years, right? But it's not the most used web API. But you might still need it for your app. It might be that the thing you're building needs Bluetooth connectivity. And so there always has to be a way to solve your problems. And the way every programming language ever, I think, has done this is through a foreign function interface.
+
+So when Node came out, it could do a lot of things, but most of the libraries were just bindings to C. It was just "Here is the JavaScript interface. I'm actually calling some C code underneath." And for a long time -- over time, it kind of changed, and a lot of things actually got written in JavaScript, and it stayed within the same ecosystem. But initially, it was all a different language, and it was just an interface for it. And that's the same thing we've done here, saying "Well, if you want to add to something that Nordcraft can't do yet, you can create this custom action." And it's essentially a foreign interface. You go and say "Well, the action is going to take this kind of input", it can emit some events back, and whatever it does in between, that's entirely up to you; you can write whatever code you want. And it allows you to effectively say -- there is no limit. Even if you hit a really weird edge case that your app needs to do, that's possible. If you want to bring in a third party library that does weird graph rendering things, because someone made that and you want to use that. Or if you really want to use AG Grid, you can use AG Grid. You just go and say "I'm going to define in Nordcraft and saying this div is where Nordcraft--" Just like you're doing in React, right? "This is the div where Nordcraft stops touching the content inside, and then I'm going to use JavaScript to render the rest." And that's perfectly possible. You just build an interface so that you can work with that inside of Nordcraft. Again, this is not us coming up with something fantastically new. This is the same thing React does whenever you're opting out of React for rendering, and it's the same thing every programming language has done for an interface.
+
+There are also cases where it's just -- we ran into an example the other day with a markdown parser, where it's like, yeah, I mean, just use the JavaScript one. And we had an engineer who definitely felt like he wanted to build it natively into Nordcraft. But yeah, it's not a great tool for building a markdown parser. That's not the point. JavaScript can do that just fine.
+
+**Jerod Santo:** \[48:03\] But when you're building Nordcraft with Nordcraft, do you find yourself ejecting often, the way that you guys have set this up, or not very often?
+
+**Andreas Møller:** Ironically, almost never.
+
+**Jerod Santo:** Almost never.
+
+**Andreas Møller:** If you look at the project -- and we are we're in the process of taking it open source, so in the coming weeks we're going to open up the editor project, so anyone can dive in and see how we build it. And in there, there's a lot of custom code, but almost all of it is because we made it before -- like, maybe a couple years ago, before Nordcraft could actually do that thing. So it's really not a lot of times. It's quite rare we go in and do that. The most recent example I had is we rebuilt our whole style panel, and because we want to support the full CSS spec - so even if we don't have a UI for a specific type of style, you can still pop into... There's a CSS panel on the side where you can go and look at the actual CSS. And you can write CSS in there, and we want to support you doing that, whatever it is. But that also means we actually have to parse all that CSS, so we can figure out how to properly render it in the style panel. And so we had to sort of build one, sort of half open source, half built ourself parser... And that was a classic example of "Yeah, you can definitely do that better in code." That's not what Nordcraft is built for at all. But it is quite rare. Most things, it's just easier and works better to build it directly in Nordcraft.
+
+**Adam Stacoviak:** It's interesting to build it inside of itself... How does it work?
+
+**Andreas Møller:** I mean, it's a nice experience now, but I started doing that before we had version control. That meant if you made a mistake, you broke the tool you're going to use to fix it.
+
+**Adam Stacoviak:** Oh, dang...
+
+**Andreas Møller:** Yeah. So I had a few times where I had to like check out JSON from a big JSON tree from the database, and try and search through and figure out what did I do. How do I fix this.
+
+**Adam Stacoviak:** So you're literally building the engine inside of Nordcraft? Is that what I'm hearing? What part are you building inside of Nordcraft?
+
+**Andreas Møller:** Editors. Just the editors.
+
+**Adam Stacoviak:** Okay, so the visual UI. All the UI you see is being built with Nordcraft.
+
+**Andreas Møller:** Yeah. So the runtime, as we call it, the one that actually executes and renders HTML and CSS into the browser, that is built in TypeScript. And that is all open source, and you can see it on our GitHub repo.
+
+**Jerod Santo:** And you're in the process of open sourcing more things. Or everything? More things?
+
+**Andreas Møller:** Yes. So we've open sourced the runtime and a bunch of sort of utility functions, and a few examples that basically allows you to run your whole application. Like, anything you build in Nordcraft, you can run it by yourself, on your own infrastructure. And when you do that, all our code is -- what is it? GPL.
+
+**Jerod Santo:** And your investors like this?
+
+**Andreas Møller:** They were actually fine with it. They didn't have a problem. I mean, obviously, I had built a story before I went to them.
+
+**Jerod Santo:** What was the story? Tell us the story.
+
+**Andreas Møller:** So the basic story was saying like -- we never thought we were going to build a big company out of the no-code community... Because first of all, it's not nearly as big as the developer community. And it's, per definition, never going to be big enterprise customers. Because that's the big one. So if we want to take this thing where we think we can go, we need to get developers on board. And we had some, and they were interested, and all of them were asking "So is this going to be open source?"
+
+**Jerod Santo:** Right.
+
+**Andreas Møller:** And basically saying if we want companies to trust us with the kind of projects they're building, and that size, we've got to show them something. We've got to say "Well, actually, we're not gonna rugpull you halfway through and say "Hah! Surprise, now it's going to be twice as expensive."
+
+**Jerod Santo:** \[52:10\] Lock-in is something that we're all thinking about... And if you want to build something serious on top of there, even your user in Malaysia, who's got now a platform that he's making money off of, and he needs some assurances that that's not going to dramatically change, or somehow lock him out, or whatever happens... Having open source is peace of mind.
+
+**Andreas Møller:** Exactly. And personally, I'm not sure I would -- like, I would have that in mind. There's a limit to what I would build on a platform that literally -- because it's not just some of it, it's like all of it. It's literally your whole app that you're building. So if you don't have options, that would worry me.
+
+**Jerod Santo:** For sure.
+
+**Andreas Møller:** And I think it worries other people in the same way. But we didn't build like that from day one, mainly because - well, my experience and my understanding was that open source doesn't necessarily make you go faster.
+
+**Jerod Santo:** No, not necessarily.
+
+**Andreas Møller:** And especially in the early days, we wanted to go fast, so we didn't want to open source it there. We already had in mind that this was going to be something we would do... It wasn't something that came up a lot in the early days, from users... So therefore, it was like "Okay, we're going to keep it closed. We're going to keep going." And now there's a bit of process in sort of separating first "What is the platform code that's our hosting code?" And that part of the platform that's related to our hosting service and the actual editor. So we need to separate those two apart. And we also need to sort of have an open source backend for it. And that's not going to be 100% the same as the one we have that's custom built for Cloudflare at the moment.
+
+**Jerod Santo:** That all makes sense. It's for web apps... Is it bad for websites, or just not built with that in mind? Is it just as good of a choice, or is Webflow the way to go?
+
+**Andreas Møller:** I mean, almost every time when people ask "What should I use it for?" I say that the answer is the same as you would say for React. I think that for websites, we might have a little bit of a leg up, because it's quite easy to build a website. Like, React is a lot overkill, and so is Nordcraft. But you don't experience the overkill, the extra stuff in Nordcraft to build a website. It's kind of like building in Webflow.
+
+So it totally works for websites, but you have options. There's a lot of other tools that can do websites as well as us. Where it begins to change and where we really sort of land as being our sweet spot is when you start pulling in data from different data sources or backends, and start making it more interactive and dynamic.
+
+**Break**: \[55:00\]
+
+**Adam Stacoviak:** What about things like authentication, authorization, when you get into those scenarios? I know it's an editor at this point, it's a visual interface to it, but how do you -- is that simply just part of the programmatic ways it has behind the scenes? You can enable that? How do you work that kind of stuff in?
+
+**Andreas Møller:** So because most of authentication is a backend issue, we don't handle most of it, because we don't have a backend, we just work with whatever your backend is... Whether you coded your own, or whether you use like Supabase directly as a backend, or whatever you're going with. There's also some -- like, Xano is one of that more low code/no code style backend; it works really well with that as well. So from our point of view, again, we're frontend, so we're just doing HTTP requests to a server. So what we've done around authentication is say "What matters for us is how we store the token, in essence." Whatever strategy you have for authentication is possible.
+
+What we sort of nudge people towards is storing your token in an HTTP-only cookie. Because that's generally the most secure way of handling it. And what we've done is we have an API proxy. So whenever you have an API from the frontend, if you run it through our proxy, we can actually authenticate that request. So let's say you need to send an authorization header with your user token to an API - that's the most common approach - well, you can actually set up "Well, that's what I want to send", and you don't actually have the token on the client, because that's stored in an HTTP-only cookie from when you authenticate it, but we actually put that token in before that request goes to the server. And that way, even if you have some script injection attacks, etc. they can't actually steal your token. That's not going to fit for everyone. Sometimes if you want to do -- especially if you're doing like real time WebSockets and stuff, that's not always an option, and you do need to have the token on the client. So whatever your authentication strategy is, is possible. And it sort of speaks to the general approach...
+
+I sometimes say "We didn't really invent that much." And by that, I mean almost everything works the way you would expect it if you stopped looking at it as a visual tool, and started looking at it as a JavaScript framework. So when people ask "How do you do authentication in React?", it's sort of almost a half abstract question, because it's like "Well, however you like." And similar, "How do you store things on the client?" Well, there's local storage, there's session storage, there's IndexedDB, there's all these options... Because it's a web browser, right? And we are a framework, and we just build visual tools on top of that. But we didn't invent whole new abstracts. The web's pretty good. The HTML, CSS - it's the best we've come up with yet, I think... So we're not reinventing it.
+
+Obviously, it's not running JavaScript, but it's very much stealing everything it can in terms of function names and making sure that everything is very familiar to someone coming from that world. And again, all the browser APIs, they're the same.
+
+**Adam Stacoviak:** No need to install modules and bundles and things like that?
+
+**Andreas Møller:** Well, there's always a question of like "How much do you want to build yourself, and how much do you want to use other people's code?" So we have packages, is what we call them, and you can install them. And basically -- so by default, when you open Nordcraft, I think people are often surprised that there's nothing there in terms of like pre-built stuff for you. So if you open it up, what can you add to your page? Well, you can add a header tag, or a H1, or you can add a button, or links... All those great HTML elements you can add. But there's no tabs. There's no complex media player or carousel. You can install those as packages, just like you would do in any framework, but that's not a built-in feature. You can build them yourself or you can use something made from the community.
+
+**Adam Stacoviak:** What is deployment like then? I suppose, if it's the visual side, is it -- it's not hosted. I'm still trying to figure out what exactly -- where it lives, I suppose. Tell me more. I don't know how to ask this question, but are you deploying something? Is it living somewhere? You mentioned ASTs, and your database, but how does this application -- how do you deploy this, whatever this is? Or is it always deployed?
+
+**Andreas Møller:** It is always deployed.
+
+**Adam Stacoviak:** So you're live in production when you're building, or staging, or branching.
+
+**Andreas Møller:** So how detailed do you want to go?
+
+**Adam Stacoviak:** \[01:02:08.01\] As deep as you want.
+
+**Andreas Møller:** Okay, so there's basically two --
+
+**Adam Stacoviak:** Whatever a developer is gonna care about, I would say. If a developer is thinking "Okay, I get it. Adam doesn't, but I get it. React, but different." Take us there. What would a developer \[unintelligible 01:02:20.07\]
+
+**Andreas Møller:** Okay. Well, we're dipping a bit into the developer space then. So we built our version, our hosting platform on top of Cloudflare, because Cloudflare kind of uniquely has a bunch of services that are very interesting for us. One of them is something called Durable Objects. And what a Durable Object is, is essentially a serverless worker with state. And so you can basically save things in memory, and it'll still be there next time you ask it. And it has this weird property where there can only be one instance of it deployed worldwide at the same time. And that is for most things really bad, because it doesn't scale. But it is essentially how we -- we kind of treat it like a shared file system.
+
+So whenever you create a branch - and we have version control that's not technically Git, but we copied this as much as we could. So whenever you create a new branch, each branch has its own Durable Object associated with it. And the way these work, which is quite cool, is that they are always close to you. So if I connect to a branch, that Durable Object is going to be in a data center near me, whatever closest to me. Cloudflare has like 250, or something like that. If you connect to the same branch, or if you connect to the same branch later, it's going to move that instance closer to you, to make sure you have better rewrite performance. But there's always going to be only one. So whenever I'm working on a branch and I'm updating my data and sending that to the Durable Object, that instance always has the latest version of my project for that branch. So whenever I go and say "Actually, I wanna preview this one. I wanna see what it is live", whenever I open that, it connects to the same instance when you're on a branch, when it's not deployed.
+
+So in real time, you can always see what is there. Because the second I make a change, that actually gets sent; it takes about seven to 16 milliseconds. Then that's saved, and then your application will see that. So branch deployment, those live preview environments, that is below 100 milliseconds from I make a save to you see it live.
+
+For the actual main we don't do this, because it's actually a good thing to have a lot of different instances all over the world for that... So there, we just have the main version of that project, and we use that as the cache infrastructure to make sure that every single data center has a version of that. And we are rendering -- whenever you make a request to a page, we are rendering that whole page based on our code. And we actually looked into caching, and the funny thing is it didn't do much, because it renders it really quickly on this worker... So actually caching the final page for a static page didn't make it faster. Maybe on average three or four milliseconds, but it really didn't change much. When you start adding data, when that page needs data in order to render, obviously that's a different question.
+
+**Jerod Santo:** Right.
+
+**Andreas Møller:** Does that answer it? Does that make sense?
+
+**Adam Stacoviak:** It does, yeah.
+
+**Andreas Møller:** \[01:06:02.08\] But even our deployment, even our deployment is just saying -- all we're really doing is saying "This version that we have in the database, that is now the live version." So we are moving a pointer for when we're deploying. So deployments happen really fast as well.
+
+**Jerod Santo:** How are you approaching education?
+
+**Andreas Møller:** We hired a really great person for education, a common friend, I believe, to help us with that. Salma Alam-Naylor.
+
+**Jerod Santo:** Was that Salma? That's her role?
+
+**Andreas Møller:** Yeah.
+
+**Jerod Santo:** Okay.
+
+**Andreas Møller:** She's our head of education. So education is a big part of this, because even though a lot of this is actually familiar to developers, once you get past the interface you kind of have -- like, there is a little bit where it has to click. We've seen a lot of people where they're like "How do I submit a form?" And like, it's a form, right? You know how to submit a form. You just have to see that because the visual is -- you thought there was all sorts of other things, but it's the same way you think. There's a button inside, click the button, submits the form.
+
+So there's those kinds of things that people often a little bit have to get over, but after that, it actually is very familiar, if you've used a framework before. But that being said, education is still a big part of it, because it is conceptually a new thing. Like, it isn't quite -- even the tools it looks like, it isn't quite that, and you wanna think about it a little bit different.
+
+**Jerod Santo:** Yeah.
+
+**Andreas Møller:** So it's something we've put a lot of focus on and effort into, both as a sort of "How do we approach it?" We just worked with an agency -- it's a Nordcraft-only agency. They only work in Nordcraft because they're kind of picky. They're like "We don't wanna do projects that aren't Nordcraft." So they're very focused on that. We just work with them and they actually helped us rebuild our whole documentation side, as well as a lot of the technical documentation, because they are incredibly good at it. They really understand the platform. And I think if you're looking into Nordcraft, go and check out the documentation. It's some of the best I've seen. There's a high level attention to detail. They did a fantastic job on this one. But also a lot of video content, a lot of sort of trying to "How do we try different medias, figure out what people learn?"
+
+We also have these challenges we added when you first sign up, and the basic idea is -- I kind of hate these step-by-step tutorial, where it's just "Move the mouse to the right and click the button." Like, "Okay, yeah..." We wanted something a little bit more fun, but that still kind of held your hand. And so our approach is challenges where it kind of sets your challenge -- like, you're gonna build a weather app, and you need to add an API, fetch some weather data, and you need to add all that data to your UI, and this and this and this. And it shows you the steps; there's always a chance of it showing you, but it kind of like lets you try first, and saying, you know, "Give it a shot, and then if you're not sure, come back." So we're trying a lot of different things to sort of figure out what is it that kind of gets people to get it the fastest.
+
+**Jerod Santo:** I noticed in your welcome email you allow people to book a personalized onboarding session, which reminded me of, I think, Paul Graham's essay, "Do things that don't scale." Something in this one's not gonna scale as you get popular... Is that with yourself? Is that with Salma? Is that with your co-founder?
+
+**Andreas Møller:** No, it's literally with me. And the worst thing is, I think it might scale.
+
+**Jerod Santo:** Because no one clicks it, or what?
+
+**Andreas Møller:** \[01:09:55.01\] Because I don't get a lot. \[laughter\] And it's one of my favorite things. Obviously, at some point we can't do that, and it's probably not that far away. But I think I have two a week, of like 15-minute calls... It's not that bad. And it's not because there aren't people. I think it's just this idea of like --
+
+**Jerod Santo:** Kind of intimidating.
+
+**Andreas Møller:** ...going through a call with a -- for some people it's a little bit "Ah, I don't know." But developers aren't booking calls for new tools. That's \[unintelligible 01:10:23.02\] So we had it in because if someone wants do that, generally it's been a good thing. But actually, it's not -- like, yeah, we don't get a lot of people doing it. Most people very much prefer learning on their own.
+
+We have a Discord server, and we really recommend everyone going there because it's such a great place to get help, and it's a fantastic community. But there's also a lot of people who don't wanna join a Discord server for a new tool. They wanna see it first, and figure it out, and then we'll see about community afterwards. That's fair enough.
+
+**Jerod Santo:** Well, it's a cool tool. I mean, I'm impressed. Very polished... It seems like you thought through a lot of the things that would stop me as a developer from using something like this, such as a way of ejecting, such as a way of self-hosting... You're obviously getting there in case Andreas and co no longer exist, or get sold or whatever, and things change...
+
+**Andreas Møller:** Or get really greedy. I mean, there's always an option saying "We just want more money."
+
+**Jerod Santo:** Yeah, or you get greedy. Yeah, what's it costing right now? How does it work? What's the pricing look like today?
+
+**Andreas Møller:** So we think it's very reasonable. You can build a lot for free. The only time we really start charging you is... Well, so by default, we call it the open source plan. And it basically says, if you're doing a public project that anyone can see and clone, then you don't have to pay for it... As long as like -- we have some traffic limits etc. And if you want a custom domain for your application, because you're starting to brand it, or if you want it to be a private app that other people can't see... Because maybe you don't want to build an open source product, right? That's fine. Then you need to go on a paid plan. And they start at 36 for monthly, dollars per month per developer, and then once you move up to a scale-up plan, which is either when you sort of move past the traffic limitations \[unintelligible 01:12:20.14\] or you get more than three seats, then it's $100 per seat.
+
+**Jerod Santo:** And is that like collaborative? I guess we haven't talked about collaboration at all, but I assume those seats are -- like, it's multiplayer; you can be doing the same thing, doing different stuff, seeing each other move around...
+
+**Andreas Møller:** Yes. You can do that. I would say that generally doesn't work as well as people think, because being in the same branch... It's kind of like being in the same branch in GitHub. It's rare that really that's how you want to work. You generally want to work in separate branches.
+
+So you can, absolutely. And the way this works - if you go into the same branch, you'll see the update someone else makes almost as fast as they do... Assuming you're in a similar location. But most of the time people work async. So just like in Git, we check out our own branch, that's what we're working on. When we're done doing what we're doing, we'll commit that branch and publish it. If someone else did that since, we are merging their changes and our changes, and we can see what we're doing. And again, the workflow is exactly like if you're working in a codebase with Git. The big difference is that the way we actually diff the projects - because there's no code, therefore there is technically no Git, because that's file-based. So we had to change some of the implementation details, but the process and the workflows are the same.
+
+**Jerod Santo:** Cool stuff. Adam, any other questions for Andreas before we let him go?
+
+**Adam Stacoviak:** I don't think so. I don't think so. I'm excited about it. I think you're onto something good here. Don't stop.
+
+**Andreas Møller:** \[01:13:59.01\] We want. We are going to keep going, because excitement is only building. We've got some really cool things coming. One thing I'm really excited about is that one of our engineers has been working on an animation editor. And I've always felt like animations on the web -- we really technically had the tools for doing great animations for like a decade... Like keyframe animations - you can do a lot just with keyframe animations and transitions. They're quite powerful. But if you go and look at people's web apps, especially SaaS apps, they're all super-static. They're all kind of boring. And part of it is that while we've had the tools, finally adjusting a keyframe animations in CSS... It's pretty painful. You have to go and say "Oh, actually, the timing is a bit off, so now I need to go and move keyframe number three. Am I going to move that 2% or 3%? Let's see", and then I go back, and then I'm waiting for live reload. That kind of work, adjusting animation to feel nice - that is fine-tuning work, and code is not great for that, because the feedback cycle is so slow.
+
+So one of the really nice things is that once you have a visual tool with a fast upload, you can start actually building an animation keyframe editor, and you can visualize your keyframes as in the good old Flash stage. We lost something when that went away. We also gained something... But you can start adding in those kinds of keyframes, and actually really animate every aspect of something, and adjust them and fine-tune them. It's the same thing we're seeing now, just something like a gradient. Just having a gradient in it means that you will actually use gradients... Because nobody knows how to type gradients into CSS. You forget it the second you've done it.
+
+**Jerod Santo:** Totally.
+
+**Adam Stacoviak:** Every time.
+
+**Andreas Møller:** So having a tool for those kinds of things means that you actually use them, and it means that the results, the output actually gets more interesting, and more sort of delightful, I think is the word \[unintelligible 01:16:07.08\] So I'm really excited about the animation. He's done such an incredible job, and we're going to have a ton of content and tutorials and stuff coming out when that launches.
+
+**Adam Stacoviak:** How do you know you're on the right track? What is the feedback loop? I know you mentioned the other project, how they're doing some cool stuff... I didn't hear a big name yet necessarily. How do you know you're approaching or near product-market fit? What is that for you?
+
+**Andreas Møller:** So the hardest part with product-market fit is actually for us -- because I think we have what we call a product-customer fit. Like, our customers really love it. They go and make videos about how much they enjoy this, they go and talk about it, sometimes ad nauseam to other people. We get complained about people going "Look, we get it, he likes your tool, but... Can we talk about something else sometimes?" I had a call with someone from an agency that's like "I kind of just took this because he wouldn't shut up about it." \[laughs\] Okay, fair enough. Maybe they'll try it, maybe they'll like it.
+
+So we have a very strong fit. The number one guiding thing for me is that we are working in this tool every single day. Like, the whole team. Obviously, some of our engineers are working on the backend and different aspects, and we write code as well as part of it... But the vast majority of the work we do every day in the whole team is inside this tool. And because we're still a small team - we're only 10 people at the moment building this thing - it's a very advanced, very experienced team. So these are people who've been building -- I think our most junior has been doing it for 10 years. So "most junior" is the wrong way of phrasing it. But these are people who've been doing this for a long time, and they know how to build these things. They're exactly \[unintelligible 01:18:11.28\] of users we're looking for as well. And we know how this works. We dogfood, as you say, every single day.
+
+\[01:18:22.15\] And once you get to that bit where it clicks and you start working with other people and you figure out how all these things work, that dynamic of designer and developer working together, the idea of letting go of that afterwards - that feels awful. The idea of me going back to having a design handoff with a Figma design that I had to go and replicate... I mean, that feels so backwards right now. And personally for me, I have like a couple of hundred apps now of projects I've built in Nordcraft. Because every time I have a new idea, that's always gonna be my go-to. Not just because I start a company, but because it's just so much easier to get started, it's so much easier to get your ideas out, it's so much easier to build UI, visually, than it is to sit and write code.
+
+And a lot of the other bit, I think -- I find that most people I talk to... Because it is a big sell. It is a big difference. It's a big shift compared to what people are used to. It is quite a radical idea. Even though, again, when you think of the development engine, like gaming etc. it probably shouldn't be so radical... But it is a big shift. And it took me two years to convince myself it was a good idea. I didn't believe in it for the longest time. I've been the hardest one to convince, because - why wasn't it done? What was I missing? What was the thing -- okay, it looks fine now, but what about six months down the line? Is it still gonna be good? And that was the motivation for starting to rebuild the editor in itself, is because I needed a project that was big enough that I got that feel of "What is it like at scale?" And now understanding -- obviously, 10 people, you can debate whether that's scale. But it's enough to give you the right idea.
+
+And now building this tool -- like, if we did that in anything else... I think doing that in React would slow us down a lot. We wouldn't be able to move at the same speed at all. Just the fact that 2 out of our 10 team members wouldn't be contributing directly, but just feeding in designs, right? Well, that's like a lot already. So our whole team, we're using it, we're building all sorts of stuff in it, we're constantly drafting... We also know all the bad parts, all the things that aren't as good. There's always "This thing is not as good as I want it to be", or "This thing annoys me, but we'll see when we get to fix it." We also get to feel all the bad parts way earlier than anybody else. But really being inside the project, having it be your daily driver for everything you do - it gives you a very good idea of what it's capable of.
+
+**Adam Stacoviak:** I'm thinking -- I'm just kind of camping out on like this transitional period. So you've got this very uphill battle when it comes to marketing. That's what Jerod asked about education, because he's like "How do you convince people they should make this change, that they have a problem?" We all know there's a -- we've just defined the problem. Your co-founder had the problem, I had the problem, you've had the problem in the past, of like designers and developers working in the same tool, and not double-working or having this scenario. But I'm just thinking like \[unintelligible 01:21:35.25\] but who is coming into this thing initially? Is it developers? Because you've spoken about this allergy of no code, and things like that... So who is it that's coming in or seeing the light, so to speak, this a-ha moment, that is the early adopter?
+
+**Andreas Møller:** \[01:21:59.29\] It's a good question, and it is a split. There's not a one clear... Most people who come in, when we ask them "What do you identify as? Are you a developer, a designer, or something else?", most say developer. And then that's like half, around half. And then there's a 50/50 split between designer and none of the above. One of the things that's really interesting is that I think there's more to win as a designer early on. Because as a developer, what we're basically saying is "This is a slightly nicer way of working, and you'll probably work a little bit faster." And also, if you -- again, a big part of the value prop is the fact that building UI is very valuable in a visual environment.
+
+I think the logic itself, the reason why that's visual programming is not necessarily because -- it's not because there's anything wrong with code. It's just because context switching quickly gets annoying. So my initial draft, all the logic was done with code, and only the UI was built visually. And then I sort of realized that "Okay, but if I'm just binding a value to an attribute here, it's kind of annoying having to switch to a code editor to do that. Why can't I just do it right there?" And then over time that kind of kept growing, into a point where I'm saying "Actually, I kind of want the full power of a programming language, without having to context switch out of the app to a code block somewhere." But that wasn't really the point. The point was building UI. And that's where I think most of the value still is. And for that reason, as a designer, you go from "I make a Figma design, or maybe I can make a Webflow site" to "I can actually build an app. I can build probably a symbol to start with app." So that's a massive value change in what you can certainly do.
+
+For developers it's a little bit less, because you could always -- like, anything you can build in Nordcraft, you could probably already build the same thing. But at the same time, the barrier of entry is lower. It's easier for developers to understand, because we see people who don't come from a developer background - it's not necessarily that they have a hard time using the tool, it's that they hit those developer problems, of saying "Well, I don't actually know how to solve this problem." And it wouldn't really matter what the medium was, it's just how do you solve the problem.
+
+Let's say I have an array of 25 likes, and I need to figure out if I like this post. So you have to search through that list of likes to figure out the one that you made. Okay, how do I do that? And the answer is abstractly the same, whether you're doing it in Nordcraft or JavaScript. But if you've never done either, that doesn't help.
+
+**Adam Stacoviak:** It certainly seems like a big win for designers who are consistently interfacing with their developers, building applications; not obviously building out other things that are not applications, because that's what this is for.
+
+**Andreas Møller:** Absolutely.
+
+**Adam Stacoviak:** Because you don't have the risk of your research and your baby and all your work not getting translated well... And you get to work directly in a tool that is essentially as real-time as real-time can be to your production interface. You can branch, you can do all these fun things... And you don't have to double the work. Your team doesn't have to double the work.
+
+So from a business standpoint, it's like "Well, we could save a lot of money." And the tool is actually -- the one tool is the one tool. It's not "This tool, and that tool", and some sort of translation period. It's in this static design pixel-perfect world that has to get translated. So it sounds like designers are probably your first customers, and developers, engineers following suit because they're being pulled... Maybe happily, I'm not sure. I'm not sure.
+
+**Andreas Møller:** \[01:26:03.18\] Whenever we talk to companies, that's how we go about it. We're saying "This is a tool for your designers that brings your developers in." And that's very much our strategy. Of course, when it's just people trying it out, people come in for whatever that interests them. But one of the interesting things about the design profession is that, as I said -- well, with frameworks we definitely push them out of the browser. They're not working in that anymore. But we sort of pigeonhole designers a little bit... And I think design tools are maybe a bit to blame as well. Because we have UI/UX designers working in Figma, but you cannot do UX design in Figma. There's nothing for you there. It can do a semi-responsive website, but what about keyboard interaction? Those things have to be designed. There's so many aspects of UX which is until you touch and feel the application, that is the user experience.
+
+So we've kind of cut a lot of that roll off, and say "Actually--" So today, in a standard dev team, frontend developers -- there's more UX than UX designers, because they're the one who are actually hands on. I remember talking to my partner and saying -- I talked to him about tabindex, and he's like "What is tabindex?" Because -- I mean, he didn't know. And it's the -- well, that's basically how you decide what order you're going to go through the interface when you press the Tab key. And which things can be focused, which things can't. That is a big part of UX. But there's none of that in Figma. And I would probably guess that a lot of especially younger UX designers today have no idea... Because they don't have the tools to work with that. That is -- by the time that becomes relevant, that is way into the developer domain.
+
+**Adam Stacoviak:** You also get to fast-forward a lot too, because you make -- I've made design choices that were incorrect. And it was evident immediately, as soon as you use the real application. It's like "Oh my gosh, we actually implemented it the way I wanted to, but it wasn't -- what I designed was static. It wasn't really usable." In modern tooling you can do a lot of the stepthroughs, things like that, but you can sort of define some of the UX, but not all of it. And so until you see the real thing \[unintelligible 01:28:25.20\] you're kind of making some educated guesses and hoping that your selections are right, and they weren't so wrong that you've got to do a bunch of rework.
+
+**Andreas Møller:** Yeah. And we actually usually do what we sort of dubbed, jokingly, the reverse handoff, which is a lot of times I will build something hideous - and we actually almost made a sport out of making UI that's as ugly as possible before we hand it off to a designer... Because then the designer would actually do the UI. We just technically needed to get the data on the screen. Because when they're designing, they're designing with real data, real application right there. Everything they're doing. I mean, they'll be rearranging the stuff around, and making sure it's laid out right, everything. But it's the real app.
+
+\[01:29:13.13\] And the funny thing is when we talk responsive design, we always talk about changing the width of the browser... But there's a lot more to it, because the content changes. So everyone, every developer has at some point received the design that looked great, and then come back and say "What do I do when the title is 120 characters?" And then it's like "Well, then actually, unfortunately, the whole design falls apart, because it can't break line, there's not enough room, and it can't overflow, and "Ah, do we do a tool tip?" Like, these kinds of things - what happens when there are no tags on this element, because then it just leaves this massive white space in the middle? All these things -- will design change based on data when you're building dynamic applications? So if you can't see the data, you don't know the cases.
+
+**Jerod Santo:** Right.
+
+**Andreas Møller:** And that's why we have this sort of like back and forth in our teams all the time, of like "Well, what about this case?"
+
+**Jerod Santo:** Well, this has certainly been probably the most expensive part of most buildouts in the history of web development, is the glue between design and dev, and implementation, frontand/backend... All the seams where we come together and collaborate, it always gets very expensive, and you can save a lot of money if you have teams that are good at collaborating around those things, and catching stuff early. But that takes time and experience, and so having a tool where you can meet in the middle, develop in the middle, and design in the middle, and get that quick feedback, certainly I think needs to exist. So I'm excited that y'all are building one, and that's going to be open source, and that it's polished and ready to be used. So I'm going to give this thing a shot. I'm going to try to build a really ugly UI, and I'm going to pass it to Adam and see if he can make it look better.
+
+**Andreas Møller:** Yeah. I mean, I find that all the named HTML colors are really good in this case.
+
+**Jerod Santo:** Oh, yes...
+
+**Andreas Møller:** There's a lot of -- yeah. There's a lot of good ones there. You don't need to go into any of this hex business. Just pick one that has a name... That's kind of the best ones. That's my strategy before I hand off to a designer.
+
+**Adam Stacoviak:** This hex business. \[laughter\]
+
+**Jerod Santo:** Very cool. Andreas, thanks for sharing with us, man.
+
+**Andreas Møller:** Well, thank you very much for having me. It's been a pleasure. I've been listening to the show for so long... So it's really exciting to be on it.
+
+**Adam Stacoviak:** Very cool. Glad to have you here.
+
+**Jerod Santo:** Absolutely.

The era of durable execution (Interview)_transcript.txt

@@ -0,0 +1,507 @@
+**Jerod Santo:** We are joined today by Stephan Ewen from Restate.dev. Stephan, welcome to the Changelog.
+
+**Stephan Ewen:** Hey, thanks for having me. It's a pleasure.
+
+**Jerod Santo:** Pleasure to have you as well. Adam, how are you doing, man?
+
+**Adam Stacoviak:** So good, Jerod. How about you?
+
+**Jerod Santo:** I'm doing well. Always excited at the beginning of a conversation to dig into something new, something different... And something called Restate. This is supposed to be the simplest way to build resilient applications. This was a requested show, Stephan. So we do take episode requests. This listener would like to remain anonymous.
+
+**Adam Stacoviak:** Ooh. Secret.
+
+**Jerod Santo:** However, they say that Restate is a super-exciting approach to managing distributed systems. And they say that we should get you on the show, and so we just take orders around here, and our listeners often get what they want... And so that's how we found you. It was through a listener request.
+
+**Stephan Ewen:** Awesome. That's very cool to hear. It's that open source communities at work, and all that.
+
+**Jerod Santo:** That's right. So Restate. Let's not get into Restate itself at first. Let's talk about resilient apps first, because this is called -- tagline, "The simplest way to build resilient applications." Let's talk about that. What is exactly a resilient application in your estimation?
+
+**Stephan Ewen:** Okay. Yeah, so the way we think of it in the context of Restate, we're talking mostly about the backend of the application, the coordination and orchestration logic. A resilient application would be an application that doesn't accidentally drop your order, that doesn't accidentally place it twice if you hit F5 at the wrong point when you're in the browser, that doesn't accidentally book your Uber for two people instead of for one, that doesn't just disconnect you from your chatbot, lose the history, make you start over... All these kinds of things. This is what we're talking about when we mean resilient apps in the context of Restate. So basically, apps that tolerate all sorts of hiccups, errors in the infrastructure, unavailable endpoints, network failures, process failures, temporary outages, but also types of programming glitches that cause requests to fall through, and having to be retried in order to process reliably, but then not get duplicated, but the system understanding how to idempotently treat them as a retry, and not as a second request. This is sort of the bigger picture of what we mean here with resilient applications.
+
+**Jerod Santo:** Can you take a moment to demystify that term that you just said, idempotently, so everyone's on the same page? What does that mean, idempotently?
+
+**Stephan Ewen:** Yeah, I think idempotency, you can think of it as just understanding that a repeated request is actually not a new request, but it's the same request again. You're just sending it again because you were maybe disconnected from the original request, or you got an error back. It's the thing that if you don't do it correctly, that is actually accidentally placing an order twice when you did try to place it once.
+
+**Jerod Santo:** Right.
+
+**Stephan Ewen:** It's the thing that many applications don't get right, and that's why you still see too many websites saying "Don't hit F5 while this is showing."
+
+**Jerod Santo:** "Do not reload while we finish this transaction."
+
+**Stephan Ewen:** Exactly, because it doesn't understand that you submit the thing again, and this is actually supposed to be the same thing. It's just another submission of the same request. It has no way of identifying that, so it might accidentally treat it as a second one, or it has just a very rough way of doing this. It's a surprisingly complicated problem, and there's lots of applications that don't get it correct, and there are lots of weird ways applications work around this.
+
+\[08:12\] As a fun fact, I think the first bank where I was a customer, they did only allow you to wire to a certain recipient, a certain amount once per day. If you were trying to wire the same amount to the same person a second time for the day, they just wouldn't allow it, because they didn't know if that was a retry from your browser, or if that was something stuck in a queue at some point in time... That's just how they duplicated it.
+
+**Adam Stacoviak:** Interesting.
+
+**Stephan Ewen:** So generally, idempotency is deduplication in a meaningful way, understanding retries versus new requests.
+
+**Jerod Santo:** Yeah. That's called overkill, I think, when they did that. They're like "How can we just use a blunt tool to solve this tiny little problem? Let's just not let you do more than one per day." Not ideal, surely, for lots of uses. Okay, so same request twice; operates the first time, won't operate the second time. Generally speaking, how do you achieve this? I mean, you could just limit to one request per day, but if you're not going to do that, how do people usually implement or ensure idempotency in their applications?
+
+**Stephan Ewen:** Yeah, so in a way, you basically have to find a way to anchor the identity of requests all the way through. There's different standards for doing that. The HTTP standard has actually defined a header where you can put an idempotency key, and the servers, when they support this, they're supposed to understand if that key is set. A previous request with the same key has come, and the same parameters, so this is a duplicate request for the same operation. But then, down the road, you basically just try to anchor requests and the processing of different steps in each other.
+
+When you do message queues, you place correlation IDs. When you're working with databases, maybe you try to address primary keys, or transaction IDs, or \[unintelligible 00:10:14.22\] or tokens... There's tons and tons of tricks people do, but it's ultimately still a very hard problem to do if you want to do it end to end.
+
+**Jerod Santo:** It's kind of a mindset. You have to set out to do it, don't you? Otherwise...
+
+**Stephan Ewen:** I think you have to design from the start for it. The weakest link in the chain that breaks, it's one of these things. Yes.
+
+**Jerod Santo:** Or you can just say, "Do not reload your page while you're hitting this API."
+
+**Stephan Ewen:** Exactly.
+
+**Jerod Santo:** \[laughs\]
+
+**Stephan Ewen:** It's the simplest way to do it. Just make it the user's problem and don't handle it in your infrastructure, yeah.
+
+**Jerod Santo:** Also, as an engineer, when I've achieved idempotency, I know it feels very good when you're like "Okay, I'm for sure not going to do double execution in this particular code path." That feels good. And as an end user, I'm also happy when I know that I'm not going to get charged twice, for instance.
+
+**Stephan Ewen:** That's right. So you can actually see good APIs design this in from the start. I think the first API where I came across that where I thought that was really, really well done was Stripe's payment API. You can really see -- I think that's also why it got so insanely popular so quickly. The way they made that handling just seamless for folks that embedded this into their code to really understand "How do I make sure I set this request once?" and deal with all these things... That was really stellar.
+
+**Jerod Santo:** Alright, so here's a harder question. Where does the word "idempotency" come from, and why do we use that to describe this thing? It seems unnecessarily verbose and jargony. Do you know?
+
+**Adam Stacoviak:** You should spell it first, Jerod.
+
+**Jerod Santo:** I-D-E-M-P-O-T-E-N-C-Y would be the adjective. Do you know? If you don't know, you can just say "I have no idea." \[unintelligible 00:12:00.20\]
+
+**Stephan Ewen:** I don't really know, but my guess is it's a Latin word. It comes from Latin originally.
+
+**Jerod Santo:** It does sound like that, yeah.
+
+**Stephan Ewen:** I don't know.
+
+**Jerod Santo:** \[12:12\] It's all Greek to me... Maybe we can get a real time look up for that and follow up on it from some sort of LLM. Just prompting Adam behind the scenes to prompt his favorite LLM.
+
+**Adam Stacoviak:** Yeah, I'm on Wikipedia as we speak...
+
+**Jerod Santo:** Okay, good. We're stalling for you.
+
+**Adam Stacoviak:** I don't really have the details here for you, Jerod. I'm sorry I can't LLM quickly enough for you, but it says "Idempotency is the property of certain operations in mathematics."
+
+**Jerod Santo:** Alright. I went straight to the LLM and I got the answer to my question.
+
+**Adam Stacoviak:** There you go.
+
+**Jerod Santo:** So "The term idempotent comes from Latin roots." So Stephan, excellent call there. "Idem meaning the same, and potent meaning having power or being able to. Put together, idempotent roughly means having the power to remain the same."
+
+**Stephan Ewen:** Mm-hm. Wow.
+
+**Jerod Santo:** So there's the actual word. And then, yes, mathematics, blah, blah, blah. I've stopped reading now, so hopefully that wasn't a hallucination and we can all move on happily.
+
+**Adam Stacoviak:** Oh, yeah.
+
+**Stephan Ewen:** No, that sounds about right. Like, if it's an hallucination, it's a --
+
+**Jerod Santo:** It's a good one.
+
+**Stephan Ewen:** ...close one.
+
+**Adam Stacoviak:** \[unintelligible 00:13:12.29\]
+
+**Jerod Santo:** \[laughs\]
+
+**Adam Stacoviak:** Yeah, I-D-E-M plus potency - same plus power.
+
+**Jerod Santo:** There you go. It's the same power. Very cool. Well, thank you for scratching my itch, my curious itch there both of you, and ChatGPT, I suppose, pitched in on that one. What else? So we're talking resiliency... I'm curious -- obviously, to have a resilient app just good... Like, who wouldn't want these things? And just taking one of those things, idempotency, and realizing it's hard to achieve on your own, throughout especially more complicated applications... Why did y'all set out to solve this problem for folks, to build Restate, and why did you feel like "I'm the guy for the job"?
+
+**Stephan Ewen:** Yeah, that has a long answer and a short answer. I can start with a short answer.
+
+**Jerod Santo:** Okay.
+
+**Stephan Ewen:** The short answer is it's really, I would say, the state of the art to build backends that are supposed to -- backends that do any non-trivial state management and coordination, it's completely unsustainable, I think, the way we're building this today. Just to give you an example... Or let's actually start with an example. Let's stay with LLM, because we just talked about it, right? So let's say you're building a chatbot; you're submitting something, like a message there... This thing in the end has to reach the LLM, but it has to look up the context in which that chat happened before, it has to make the call, it has to go back, store the context... You don't want it to just lose everything if you lose your connection in the middle of -- let's go with the F5 thing again... You don't want it to actually trigger the same request twice, or lose the entire session, make you start over. So you're probably just putting this as an asynchronous request that runs in the background, that you're sending from your from your chat session, from your browser, but it's a separate asynchronous request that runs, that talks to the LLM. You want it to be actually retrying in case something fails, or is overloaded and it's throttling your \[unintelligible 00:15:19.22\] And then you want to be able to reconnect to that task or request in case something goes wrong in your browser, or you accidentally hit the Back button, or whatever. Just implementing this is a surprisingly complicated thing, where you start to stitch together probably a queue, a database, and a bunch of tasks to manage that.
+
+To give you another example - we just talked about Stripe, right? So let's say you're sending a request there for a payment, and sometimes they tell you "Look, this is good or bad." Like, they accepted it or didn't. Sometimes they tell you "I don't really know." "Our fraud detector is still running." Or "We have some weird thing in the background that we're still asking, and it hasn't told us, so I'm going to send you a webhook in a moment to tell you whether this went through or not."
+
+\[16:02\] And now you have like a synchronous request there, and then somewhere else is an asynchronous request coming up. You just want to make those two reliably meet. Even if this one fails, you want it to sort of like recover somewhere, understand where to reconnect with the webhook that you're awaiting... And this little piece - it's really just one case handling in the backend, where Stripe says "Okay, I'm processing", instead of yes or no. This is actually many days of work, to make that look reliable, to make that work reliably. And it's like lots and lots and lots of things like this, that just get in the way, with so many moving pieces, so many... Yeah, so many APIs to talk to, so much work. So much more work than originally happening asynchronously, and like in separate requests, than just in the synchronous user interaction. Just gluing this all together has become such a complicated thing that we felt "This does need a better solution."
+
+This is like the motivation more from, let's say, the use case side. I can give you a motivation more from like why we actually ended up doing this. I think this is a motivation that probably lots of folks stumbled across ultimately, and so like "Okay, this needs a better solution." I think there's different projects approaching that problem. Why are we approaching it the way we're doing? This has to do with where we come from. Before we worked on Restate, we were building Apache Flink. This is a different system. It's a stream processing framework. It's basically events and analytics. So you have these events coming in, often through a message queue, and you want to aggregate them, join them. A few examples where this is used is like fraud detection in banks. Some payment events go in, you aggregate feature vectors of that through a fraud model, or... Things like the TikTok recommender use Flink to actually join information from users and interactions together in real time, and understand how to update the features that will go into the recommendation model. I think companies like Uber use it to determine pricing and traffic models and ETA.
+
+So it's whenever you have events and you want to analyze them in a way that you aggregate them into some sort of -- yeah, typically statistical value, or a materialized view. This is what we were building before. So it's an analytical framework. What did actually happen then is at some point in time we saw folks were using that thing to solve the distributed transaction processing. The types of things where you would say -- let's assume an order processing service that takes the event, checkout that order, and it has to do a bunch of steps. Let's say update the inventory, trigger payment, call the service to prepare logistics, maybe call another service to put this in the user's history... Maybe more steps, and so on. And we started to see folks using Flink for that, because it had this interesting property that it had sort of this baked in way of reliable communication and state management. It was all built for analytical use cases. But they found this is such an interesting property that they started to apply this to the transactional use cases as well, like order processing, just because they've found that this is otherwise way too complicated to build... Way too easy to build it in such a way that it is brittle and not scalable, and it has corner cases for when it violates a lot of these properties that we've just said.
+
+\[19:48\] And when this started happening repeatedly, we thought "Okay, but apparently there isn't really a good tool out there yet." And apparently, this property of like correct, stateful coordination is something people really appreciate. They feel like it makes their life easier to build this type of frameworks. And then we set out to build a solution for that, and that became Restate.
+
+It's in many ways, actually, from the way it approaches things, or from its architecture, it's inspired by our work on Apache Flink...But it's almost a complete mirror image implementation of it. It takes almost the opposite design choice in most aspects, because it's really optimized for low-density transactional processing, rather than high throughput analytical processing, which was Flink. But what we retained from this idea is - yes, stateful orchestration, and event-driven foundation and so on, this is something we should build and we should be working on. And yeah, that became Restate.
+
+**Adam Stacoviak:** What's the timing in all this? When did Flink start? How mature was it, or is it? And when did Restate be born out of the idea? Give us a context in time.
+
+**Jerod Santo:** Yeah, so this context is measured in decades... \[laughs\]
+
+**Adam Stacoviak:** Okay, that's useful.
+
+**Jerod Santo:** So yeah, I mean, Flink was officially founded in 2014, but the work that became Flink is back to 2010, when I was still in university. So it was like four years of academic work, and then Flink started 2014. I worked on it until 2022, so eight years after it became an open source project. And then I left Flink, because I needed to work on something else for a change. And then we started working on Restate... End of '22, early '23. So Restate is a bit over two years old now. Flink has had its 10th anniversary last year. Flink is super-mature. I think it's used by thousands and thousands of companies at absolutely insane scale. The probably largest installation of Flink that I know must be Alibaba, who runs tens of thousands of cores for like a single processing pipeline, to live-compute their e-commerce search and recommender continuously. Restate is not quite as old and quite as mature as Flink...
+
+**Adam Stacoviak:** Yeah, two years.
+
+**Jerod Santo:** It's been two years. But we've released our stable 1.0 last summer, and we've recently released our first distributed, replicated, highly available architecture... And I have a bunch of folks that actually productively use that, so I would say we're on course to get there.
+
+**Adam Stacoviak:** Nice. Was leaving Flink difficult? Sad? Joyous? Like, what was that like when you left? Because that's a long time to work on one thing, and then to move on to that thing... I mean, you become pretty attached, don't you?
+
+**Jerod Santo:** Yeah, absolutely. So it is definitely a difficult thing... And it was not just leaving Flink. So we created Flink as an open source project, but we also built a company around that, and the company went through an acquisition, but we actually still stayed there, started building and growing the team... So I was simultaneously leaving the open source project and the company I'd built, and everything... And absolutely, it's a difficult thing, because that becomes like your baby. It's actually like two babies, in a way. The project and the company. But yes, I feel -- at some point I felt like there's this... I don't know how to say this in English, but I felt I was getting this tunnel vision on problems. I've been working on this for so long, and I've seen so many repeated things that like whenever I heard a problem, I was just like putting it into this category that I knew. And it started happening that I did that with things, and I later realized "Oh no, that wasn't actually the right thing for that problem. That might've been for the last nine, but this one should actually have been different." So you get this kind of... If you work too long on the same thing, you're starting to not see the forest for the trees, and I felt I was reaching that point... So it sounded like, yeah, I should probably start doing something new.
+
+**Adam Stacoviak:** \[24:20\] It's kind of like familiar grooves in your brain. If you just go to the same thing over and over again, it's like your brain gets new grooves, and then different problems fall into those familiar grooves, and they just slide in there... Sometimes when they don't even fit, or they aren't a good fit. I certainly understand that. I think when you've focused on one thing for a very long time, it is hard to think outside that groove.
+
+**Jerod Santo:** For sure.
+
+**Adam Stacoviak:** And it's interesting that you're building a similar system, spiritually similar, but I guess you could say radically different architecture. Is that the way you described it? Like, it's inspired by Flink, but it seems like it takes the opposite approach to the world.
+
+**Jerod Santo:** Yes, I think you could call it like that. It is under the hood still an event-driven system, as Flink is, but it is built for completely different \[unintelligible 00:25:16.02\] Just to give you an example, the core of Flink is the way it is this exactly one stateful processing, so it basically has these data streams that it keeps moving, operations that do stateful stuff with the events, count, joined, and so on... And then there's this asynchronous process running in the background that takes these consistent snapshots. So if something goes down, you can restore the state from the snapshot and just start the flow from there. And it has this clever way to do this in a way that maintains consistency across all the parallel machines, and it does that efficiently, incrementally, frequently, and so on. But it's a very throughput-optimized thing. So it stays off the critical path, if you wish; it runs in the background. So yeah, it's really good for throughput, but it does this persistence operation once every couple of seconds. If you tune it to be very frequent -- most people actually run it more in the order of minutes. So when something goes down in a pipeline, you just replay the last minute of data, which typically doesn't quite take a minute, because it replays faster than the original, the data rate at which the events are produced... But still, it takes you back a certain amount in time, which is usually okay for analytics. The worst thing that happens is, okay, this feature here in that vector that goes into that traffic model or that recommendation is maybe a few seconds older than it would have otherwise been. It's not such a big deal, typically.
+
+On the transactional processing side, imagine you have this multi-step process. You want a really fast checkout process, and you say "I do want to just start the next step after I know the payment has gone through. Before that, I'm not updating inventory. I'm not kicking off any of the other processes." Then you really need actually a persistent step to be recorded, ideally in milliseconds. Maybe it's not that critical for the processing, but we're building this for even more low latency use cases, like payment processing and settlement and so on. And there, you really just want to kick off the next step after -- you know, the previous step is persisted, possibly in a multi-data center replication way... And only then do I start the next thing. So you have to really design this completely differently. It's completely optimized for low latency, transactional durability, rather than analytical throughput. So yeah, it's a completely different design, even though both ultimately are event-driven architectures.
+
+**Adam Stacoviak:** Right. But your atomic unit now is compute, right? Like, it's logic. And perhaps data that comes from that. It's a transactional step. And so you can't just skip a transactional step, because there's a workflow here, and certain things rely on other things... And so the way that you think about durability as opposed to analytical data is, like I said earlier, radically different. That makes sense to me.
+
+**Jerod Santo:** Yeah, exactly. The atomic step in Restate is extremely fine-grained. We're really building this in such a way that you should feel comfortable in a program that you do to use Restate to persist fine-grained steps, state updates... It uses internally actually this durable mechanism for leader election to understand that it can lock and fence off different retries... And this is such a fine-grained nature. What's really important is that recording a durable step has the lowest possible latency, whereas in Flink, the atomic step is like a couple of million events being aggregated together in some state distributed over 10 machines. So that's one atomic step. It's completely different, yeah.
+
+**Break**: \[29:08\]
+
+**Adam Stacoviak:** This word durability is being used a lot... Durable execution, durability... What exactly is durability? Doesn't fall down, doesn't break, always good?
+
+**Stephan Ewen:** Yeah, yeah, yeah. Yeah, something like that. I think durability is probably the same as persistence maybe, with a bit of a stronger emphasis on it really doesn't get lost after it happens. So durability is the D in ACID when it comes to databases. Databases say "We're giving you atomicity, consistency, isolation, and durability. Once you do an update, we're not going to lose it." No matter what crashes, the database has a mechanism to bring that change to the database back. If I told you I've recorded that row, I've recorded that change, it will be there, no matter what. And in the context of Restate, that doesn't mean -- for example, the core building block of Restate is a stateful, durable function. You can think of it like that. And the stateful, durable function, when you schedule an invocation for that, or as you go through the code of that stateful, durable function, it has like multiple steps, recording a step. Whenever you go beyond a step that you asked Restate to treat as durable, you know that no matter what happens, you will never re-execute that step. You'll never come up with a different value. Like, if your machine goes down, the Restate server goes down, if you deploy it across availability zones, the data center goes down, the network gets partitioned, whatever, you'll never ever go back and re-execute that step if it once told you that it's done it. TThat's sort of the meaning of durability. Once it says it's there, it's always going to be there. And I think this is, in a way, almost one of the magic ingredients.
+
+The way Restate looks at making distributed application development simple... I'd say there's two core pieces that you need to think about. One of them is the durability; make durability extremely fine-grained, and extremely cheap. Because if you can apply durability in fine-grained steps, you always have to worry about very little after a failure. Let's say your durability is coarse-grained. Let's say the other workflow is one durable step, and it crashes in the middle. It gets retried. It's up to you to figure out "Well, did I actually process the payment already or not?" Maybe there's a way to just like assume "Okay, it's idempotent. I can send it again." Or I might even not be able to ask the service "Did I do that or not? Did I actually decrement the available kind of product already or not?" Maybe I have a way to, again, make this durable, or not. I don't know. These things tend to be harder than one thinks, because sometimes the API gives you -- you know, it might've given you an error back the first time, and you thought "I didn't do this" and followed some control path flow... And then the next time you actually get, not an error, but the real result, and then you follow a different path... So people mess up this all the time. It's really hard to reconcile if you have these multiple steps as \[unintelligible 00:35:08.20\] atomic unit. "What did I do? How did I do it the last time? How do I recover from this?" But if you have extremely fine-grained durability, if you're recording every individual step as durable in the system, and when it comes back, it can tell you exactly like "This was the last step that you recorded", then you just have a very small amount of uncertainty. "Okay, here's this one thing that I might have tried already. I have to just worry about that bit", instead of the whole history, and possible control flow, and all the choices, how I might have ended up here that I need to reconstruct in order to proceed consistently from there. So just like very fine-grained durability is extremely powerful in simplifying things.
+
+I'd say the second magic ingredient is then how do you anchor this in the whole retrying and resolving potentially inconsistent situations with partitions, with timeouts, with zombie processes, and so on, so that there's always a very consistent view of what the last durable step was. I think that's the second sort of ingredient of Restate. It's not just durability, it's actually durability and consensus, and giving you a very, very crystal clear view on where you left off, where you need to continue from. I think if you take those two things in conceptually, you've simplified the problem massively, and the rest is almost API sugar that you built on top of that.
+
+\[36:27\] That's, I would say, the magic that happens in the Restate runtime. It's a very low latency, durable consensus log that fuses queuing, state management, locking, fencing, creating futures, resolving futures... All these kind of operations that tend to be part of a distributed coordination process.
+
+**Jerod Santo:** When you say the Restate runtime, what can you liken that to, for those of us who don't know what a Restate runtime is? Is it like a Node.js thing? Is it like a database? Is it like a -- what is that?
+
+**Stephan Ewen:** Yes. So using Restate is a bit like, I would say -- somewhere in between using a database and using a message broker.
+
+**Jerod Santo:** Okay.
+
+**Stephan Ewen:** So you write your program pretty much as code, how you would write it before, but you're using the Restate SDK. Think of it a bit like your database driver, in order to sort of wrap certain operations as "Okay, this operation here should be recorded as a durable step", or attach this state to the invocation transactionally, or create this future, complete this future, and so on. So you do these operations through the Restate SDK.
+
+Restate itself is then -- maybe message broker is the best comparison. It's on the level of the message broker. So when you invoke your code, you're not calling it to that directly. You're actually calling it to Restate, which makes the invocation of your function on behalf of you. The programming model that we try to provide is you're writing a service that looks like an RPC service, like you're writing handlers, RPC handlers... And then Restate almost looks like a reverse proxy for you. So the other services, instead of calling the code directly, they call it indirectly through Restate, Restate proxying the call. And it puts itself in the middle with its durable consensus log, and when it forwards the request to the service, it just isn't forwarded naively as an HTTP request, but it actually uses an invocation protocol; it uses HTTP/2 or another type of streaming connection, holds onto that connection, allows the service to sort of synchronize fine-grained steps. When it forwards an invocation, for example, it will tell it exactly what the supposed state of the world should be, as in "Here's the steps I know you should treat as completed. Here's where you should continue." And it will then allow to use that connection, that sort of lifeline to let the application create durable actions. So yeah, it's like on the level of a broker or database. It looks like a reverse proxy to the invoker, it looks maybe almost like a database to the service that uses it.
+
+**Jerod Santo:** And when would somebody reach for this? Now, you said distributed systems, but some people think every network attached system is a distributed system.
+
+**Stephan Ewen:** Yeah.
+
+**Jerod Santo:** I mean, if I'm building a web application - let's call it a monolith that answers HTTP requests, and has a database backend; a Ruby on Rails, or a Phoenix, Django... Insert your backend framework here. Are those folks pulling in Restate and using it for certain aspects of their workflows? Or is that not necessary for them, because they are kind of a monolith? Like, do I have to be building a services-based architecture? Like, where does it fit in?
+
+**Stephan Ewen:** I would very much be with you on like almost any system we've built as a distributed system.
+
+**Jerod Santo:** Yeah, completely.
+
+**Stephan Ewen:** \[40:09\] So I think it becomes useful very, very quickly. Maybe one way to think about this is -- your backend, where you maintain state and update it and run operations and changes, it usually has a database that has the sort of core business state. Some operations go just like purely straight to the database. That's all they do. That's fine. But any time you have to do something that's not straight against your core database, but something that goes against, like different APIs, something that runs in the background... Yeah, something that is asynchronous work that goes beyond just touching the database, I think you already are at the point where it's starting to become useful.
+
+Then, if the only thing you're doing is maybe forwarding one call - yes, maybe it's overkill. But I think the usefulness starts much, much sooner than lots of folks realize. I would say every time you think about pulling in a message queue, you should probably start to think about pulling in something like Restate... Because it gives you a way to do the things you're probably trying to do with the message queue, but in a more high level, in a more well-defined, concrete way. You're not reading events, but you're dealing with stateful, durable invocations, stateful, durable functions all of a sudden, which is very often what you really want. If you're putting something off the synchronous path with a queue, you very often want to say "Okay, here's something where I really care about that this happens. It shouldn't get lost." That's why I'm putting it in a queue. And then you probably care about this thing happening once, having reliable retries... You quickly reach the state where the processing of this operation is actually multiple steps, and then you're again in the "How do I do reconciliation of multiple steps if it failed somewhere in the middle and I don't know what I already completed or not?"
+
+So I would say the moment you started pulling in a message queue, you probably should think about something like we said. The point comes very quickly. That's the simplest use case. I would say the most complicated ones that we see people build with this right now is using this to replace a complex choreography of like multiple Kafka topics, and RabbitMQ queues, and session servers, and workers, and so on. Or even like a distributed sort of payment ledger keeping system. So it's really a very broad spectrum. I would say in a way you could think like all the type of work you do in the backend that's not the central database that keeps your business state, I think it's ultimately where Restate comes in.
+
+**Adam Stacoviak:** Yeah. What about a scenario where it's publishing? I'm thinking like TikTok, or YouTube, for example... As a creator, we will upload videos to YouTube. There's a process that happens. There's a certain orchestration that happens. It has to be compressed, it has to go through certain filters... Maybe there's even a content filter it has to go through, a copyright filter... Is that an example of where you would use something like Restate, where you want it to go -- you want the user to be able to upload properly, and your server capture the data, and all the good things... But then you've got to run it through a process of saying, okay, this is now content that can be seen by what we call the world, because it's been blessed by the copyright filter etc. Is that a scenario where it makes sense?
+
+**Stephan Ewen:** \[43:57\] Yeah, absolutely. This is basically a workflow, again, if you think about it. You're uploading the video, let's say; maybe the upload first puts it into some cloud storage. But then, as you said, you first pass it to the content filter, then you have maybe a few steps that even run in parallel, like recoding it for different resolutions, optimizing it to be served to the CDN, and so on. Then you're, I don't know, running it through a system that tries to figure out what's the best sort of title frame to display, and all these different steps that you do. And they take potentially a long time, so it's a long-running process. There's a fair chance that the container goes down and the \[unintelligible 00:44:32.06\] wants to be migrated. And when it comes back up, you really want this to understand "Where did I leave off? What are the processes I should reconnect to, that are doing the encoding or the analysis?" This is exactly the orchestration of that processes where Restate would come in. You wouldn't feed the video frames to the system. That's like overkill. You don't need to feed the video frames to a transactional log. You just put them in whatever cloud storage or so, but the orchestration of the process, of the workflow, of the pipeline that does that - that's a very good Restate use case, actually, yes.
+
+**Jerod Santo:** That's a great example, Adam, because it definitely makes it easy to think through. I guess as people who upload to YouTube, we are intimately familiar with all the different steps.
+
+**Adam Stacoviak:** That's right, we are. That's why I enumerated very well.
+
+**Jerod Santo:** And it's asynchronous, because you can go about doing the other things while it's working on the long running tasks, for instance. And somebody coded up some nice orchestration behind that sucker to keep that thing running.
+
+**Stephan Ewen:** Yeah. And Google has the engineers to code up reliable orchestration flows, even in a way that they're nicely observable. You can reconnect to them, they're efficient, they know how to parallelize steps and synchronize steps, and so on. It's a much harder thing to do for many companies who don't hire the same type of engineers as Google does, and I think for those reasons it actually makes these type of things much more achievable than if you try to embark on that journey without...
+
+**Adam Stacoviak:** Yeah. Even when you were sharing how it worked, early on you were saying that it seemed - at least from my perspective, it seemed like it was user born every time. Every new application, every new scenario, every new job, every new what have you... Maybe even in your boring scenario where you were sort of focused for a bit there... You know, you keep recreating this durable invocation world over and over and over again. And why not turn it into - like you have done here, with a server and a client and SDKs for different languages, and a flow that every developer can grab? Is that kind of what landed you to this point here, was that frustration of the repetition and repeating and rebuilding every single time you build an application?
+
+**Stephan Ewen:** I think that's a great way of putting it. Yes. I would say the number one alternative to Restate that people do or use is roll your own. Absolutely. And it's a very repetitive process. And most of the time, I would say folks don't realize really all the edge cases that existed, what they do. They just maybe don't even solve them. So this is basically half baked, roll your own. And it's - yeah, every time, again and again. And it's very often very similar problems that you're solving. Let's say you're taking a message queue to say an action that I triggered, it should run asynchronously, and it should \[unintelligible 00:47:31.04\] retries. And then I'm pulling in another store like Redis, or other key-value store to record different steps. Then I might be pulling in something like Zookeeper or Etcd to place a lock on certain operations. So I don't know, they don't happen concurrently... Like, that shouldn't be updating certain -- I don't know, a retry shouldn't work on the same payment ID if the lock is still being held by the original process, or so.
+
+\[48:01\] And then you're trying to sort of - going back to idempotency, trying to make an update to another system, and understand how do you actually anchor the ID of that processing forward into the update to that other system. And you're recreating exactly, you're recreating that type of pattern over and over and over again... And I think this is where in a way workflow engines were originally born, if you wish; enterprise workflow engines, which try to say "Let's try to define a flow where we can have steps following a certain predefined control flow graph, and we have the workflow engine giving you the guarantees that step B, that follows step A really only starts after step A is done, and step A is transactionally persistent before B starts", and so on. They tend to be extremely heavyweight and flexible... They break all the tools and everything when you want to interact with them... And what Restate does and the durable execution space there, what they're trying to do in general is kind of bring this level of guarantees in a very, very lightweight way, into like almost arbitrary programs, because it's just such a useful power to have to kind of define these durable steps... Especially if you don't have to branch out in a different domain-specific language or graphical way to define them. But if you could just like write your regular code, but have it treated, have it executed with the same sort of guarantees as if it was an enterprise workflow.
+
+**Jerod Santo:** It seems like you might have a large education challenge in front of you, because there's so much thought that has to go into this kind of architecture. The fact that your largest competitor is roll your own means most people don't know. We kind of all discover this pattern slowly over time, inside of our own daily work... And so I'm just curious how you think you can attack that, or are there other people... Is there a common thread or movement that you can attach to, or create, in which people are like "Yeah, here's this new style." I thought of this because you mentioned workflow... And I think that's sort of in the ballpark of what Restate is. Message queue... I mean, is there like a simple idea or concept pattern of which Restate could be one? Or maybe Restate is the brand... But have you thought through this? Because you have a marketing problem here, or a challenge, I should call it.
+
+**Stephan Ewen:** Yeah, so I think that is very true in many ways. I think there's lots of layers of answers to that. I would say in the simplest way you can actually explain it reasonably simple, if you just start from it's stateful, durable functions, which have guarantees that they execute, they run to the end, they're able to record steps, they're able to basically do these sort of asynchronous building blocks that you have in your usual programs, calling other functions, creating promises, resolving them, updating state, making calls and so on, just in a fine-grained, persistent way that knows how to recover. This is sort of the basic building block of stateful, durable functions. Now, the harder thing is actually in a way making people realize that they should be using something like this rather than roll your own. It's not uncommon that -- it's mostly on the junior side of engineers that you talk to them and they say "I don't get it. Like, I know how to write a retry loop. What is this?" And it's a journey from there.
+
+Interestingly, I think the most enthusiastic audience is often the engineers that have been burned before. That know, "Okay, I know how to build distributed systems, but holy cow, I know how hard it is, and even though I'm really good at this, I tend to overlook still 2 out of 10 corner cases, and I get paged Sunday night or so..." Those are the ones that often go "Yes, I know why I want to use this, because I know how much time I would otherwise spend on solving all these things if I have to do it myself."
+
+**Jerod Santo:** \[52:14\] Right.
+
+**Stephan Ewen:** So I guess that's right. There's definitely an education challenge there. And I would say a very sort of like in your face example of this is if we look at the AI space and agents right now. I think every AI company is reinventing workflows in the context of agents and agentic workflows.
+
+**Jerod Santo:** Good point, yeah.
+
+**Stephan Ewen:** Everybody's building, and I would say slowly rediscovering all those things. There's been an entire sort of industry that has been working on this for like -- I mean, we've been working on this for two years, but if you ask IBM, they've been working on this for probably 30 years, or something like that. I mean, in a very different way, right? But still. And I feel like all the AI companies are kind of like bit by bit rediscovering this. And when you start talking to them, I think some of them understand, okay, even if you're building agents, if you deploy them, how they ultimately end up having to solve these problems. Again, imagine you have a chatbot that does your flight booking. It's like, there's something you have to do to make it not rebook your flight twice if the agent just crashes on the wrong point.
+
+**Jerod Santo:** For sure.
+
+**Stephan Ewen:** They're ultimately going to the same problems. They actually have a perfect foundation to build on with these systems being built today... But yes, I think they're not aware yet that this is something they'll run eventually into.
+
+**Jerod Santo:** Right.
+
+**Stephan Ewen:** So yeah, I think you can see this in many places, that the industry is rediscovering work in different sort of like subfields, that other fields have been done, just because our information flow isn't perfect.
+
+**Jerod Santo:** Minor off topic rant... Why are all of the AI agent Hello World examples - why are they all booking flights for us? It's like, do you want some undeterministic, half-baked language model booking your flight? That's like a very difficult thing to roll back. I just don't -- that's gonna be one of my last human out of the loop AI agent moves. Like, can we start with something a little bit less critical? I don't know about you, Adam, but I get serious heart palpitations thinking that someone's gonna book a flight for me...
+
+**Adam Stacoviak:** And you don't get heart palpitations, Jerod. You're a pretty chill dude.
+
+**Jerod Santo:** I am, I'm pretty chill. But I just feel like "Gosh... Do you know how hard it is to roll back a flight?" I mean, come on.
+
+**Adam Stacoviak:** Oh, yeah. Well, I think it depends. I mean, I don't mind -- I think it's the human dream to have somebody or something take that kind of action. That specific action. "Book me a flight."
+
+**Jerod Santo:** Yeah. Let's simplify it.
+
+**Adam Stacoviak:** "Set me a lunch appointment."
+
+**Jerod Santo:** How about you just give me a -- yeah, exactly. A restaurant reservation. Because worst-case scenario, I ghost it and feel bad. But if I don't show up for my flight, I lose my 400 bucks, or whatever, you know...
+
+**Adam Stacoviak:** Yeah...
+
+**Stephan Ewen:** Maybe there's just so much accumulated pain from people waiting in the call centers for airlines...
+
+**Jerod Santo:** Yeah, exactly.
+
+**Stephan Ewen:** \[unintelligible 00:55:03.26\] "Oh, that's a perfect example for a chatbot." Like, people will wanna use it, because they will not want a single other minute to spend on the phone with these calls.
+
+**Jerod Santo:** Yeah, perhaps.
+
+**Break**: \[55:16\]
+
+**Adam Stacoviak:** I'm going to go out on a limb to bring us back into somewhat left of center, but basically center...
+
+**Jerod Santo:** Please do.
+
+**Adam Stacoviak:** ...and I'm going to say that this is the year - 2025 is the year where durable execution of things is more important than it ever has been.
+
+**Jerod Santo:** Oh, really...
+
+**Adam Stacoviak:** It's always been important, but more and more people are leveraging APIs, they're building out this agentic world we keep hearing about... And I think you keep having more and more people program against brittle APIs, brittle latency of networks, databases etc. and you need that promise. I'm going to say that this is the year where the marketing problem that you have, that Jerod alluded to, is still there (I'm sorry), but it's less. And I'll tell you why it's less. Because Render - I just talked to Anurag Goel, CEO/founder of Render. And this is on their radar. So they're building an application for developers. We did a whole show on this... And during that conversation, he mentioned a brand - at least I think I did, actually... I mentioned a brand that sponsors us - not this show, but has been, and I think still is a sponsor into Q2, and maybe Q3... And that brand is Temporal. So I'm going to ask you to sort of help me understand the difference between Temporal, NATS, Synadia, Restate, your open source flavors in your cloud, what Render may be doing for application developers... It seems like this durable execution retry model doesn't live in the language itself. It's something you have to build every single time. That sucks. And it seems like more and more people are trying to solve it. So break down all those for me. Temporal, NATS, Synadia, yourself, what Render's doing, and anything else that may be doing... I mean, Flink, but that's a different world.
+
+**Jerod Santo:** There's another one called Resonate. You know that one, Stephan?
+
+**Stephan Ewen:** Yeah, I know Resonate. I know the guy behind it.
+
+**Jerod Santo:** It's pretty new. But anyways, there is definitely -- like you said, there's other people trying to solve this problem.
+
+**Stephan Ewen:** Yes, exactly. I think this starts from the same observation. The state of how things are built if you don't rely on one of those tools, it's almost unsustainable. It's hard to build, it's hard to hand it over to another person... There's often so much implicit and brittle assumptions in how this works. So folks have been trying to come up with solutions.
+
+From the ones you mentioned, Temporal is absolutely the closest. Maybe -- yeah, between Temporal and Resonate, I would say; those are the closest to Restate. So I would actually focus on those. I would say NATS goes more in the direction of flexible, persistent messaging, together with some state management blended in, and so on... But you can already see folks are trying to just figure out what are the different aspects we need when building applications, and sort of like make them work together with each other in a tighter way.
+
+\[01:01:59.14\] And if you wish, I think this is -- for me, Restate is... There's a couple of things that make it unique, but I would say two things stand out. First, I'd say the model goes a bit further than every other system. So Restate is -- if you look at Temporal, Temporal is workflows. That's really what they implement - workflows and activities. So it's like durable steps, and then with sleeps in there, and signals, and so on. So full-fledged workflows. It's actually fairly flexible if you're a power user and know how to use that.
+
+Restate goes beyond that by saying "We're not just looking at a workflow, like, one durable execution of multiple persistent steps. But we're sort of generalizing this." Almost what Temporal does for workflow, we're trying to do this for a distributed service architecture consisting of like multiple stateful services that interact with each other. And you can see this from the fact that Restate has persistent messaging and RPC built in. It has state built in that lives across a single, durable execution. So again, in Temporal terms, the workflow is done. The workflow is done. It's sort of a self-contained unit. Within the workflow, across the durable steps, it remembers context, but once the workflow is done, it's done.
+
+And then Restate is a stateful model where -- you could almost think of the activities are like decoupled from the workflow. The activities can be stateful services and entities that live for a very long time, and then you have durable functions that interact with them. It's a much more flexible and powerful model to build things like distributed state machines. We have folks that actually start to ditch certain elements of databases to put their state in Restate, because that is transactionally integrated then with the durable steps, and out of the box consistent.
+
+So I'd say number one, think of the Temporal model, but generalize it into distributed services to include long-lived state, include communication between microservices... It makes for a more powerful, more flexible box. That's the one thing. The second thing goes a bit back to what I said earlier. When we started this project, we set out with the following. You can implement durable execution. It's not terribly complicated to implement a durable execution API on top of a database, if you make it very simple. Have a step, write it to a database, on replay, just like query the database, what are the steps that are already in there... It has a lot of holes, but it gets you started. But then, okay, let's talk about the holes. All of a sudden, you have a problem with long-running processes that suspend for a long time, scaling this to zero, you have a problem that you have to worry then in your library, you have to implement your own distributed locking, and mutex in case you have timeouts, and zombie processes, and so on. So when you try to make it a really good experience, you quickly come to the point "Okay, we actually have to go a lot further than building a library on top of database."
+
+Then you start -- you know, maybe we're building a big orchestration server, that still uses a database in the background. And then you really come to the point of, if you want to make durable execution so lightweight that you can use it almost pervasively, how low latency do you have to make these \[unintelligible 01:05:21.16\] What is the best you can actually do if you deploy this across multiple data centers, if you deploy this across multiple regions? And then you come to the point that a distributed database across multiple data centers and regions - there's a lot of coordination back and forth, because the database model, it needs to guarantee integrity, it does a lot of transaction time-stamping, back and forth and round trips.
+
+\[01:05:46.25\] On the other hand, if you built this on a log, on an optimized transaction log, you can get as good as make one flexible quorum write across your different data centers, and you have a step persisted, and you can continue. So Restate is kind of going to the point where it's saying "If we want to make this extremely fast, so low latency that it has--" You can actually start to use it in places where you didn't think you could use durable execution before, because it becomes so cheap, so low latency. How would you have to build a system to do that? And that's where we went. You'd have to build it from first principles, starting with a low latency replicated log, on top of that build it like end-to-end event-driven, so you don't do batch queries on a database, but you do the most low-latency thing you can do, you do fine-grained messaging and event pipelines... And you basically layer from there.
+
+And then, the other thing is, okay, let's not just make it really low latency, but at the same time it has also to be an extremely lightweight thing... Because somebody who -- you know, we just said "What's the simplest use case? When should you actually start looking at Restate? Only when you have a distributed ledger to build? Or do you want to do this --" If the only thing you want to do is like put your asynchronous email sending in the background, but reliable. So the next thing is, how do we actually make this extremely lightweight? What's the most lightweight package we can give that thing? And the most lightweight package is single binary, zero dependencies. Just download that thing. It has its log built in, its orchestration layer, its metadata consensus module... Everything in a single binary. Just download one command, it starts in a second, and you're done. Like, there's literally nothing else to do.
+
+And then you can take this thing actually and start scaling out just by adding more nodes. If you want to migrate it, let it take a snapshot to an object store, start deploying the data, send \[unintelligible 01:07:35.05\] go from there. So what's really the experience that durable execution needs if you want to be able to take it from the point that it's so lightweight you almost want to embed it with almost any application, to this thing powers distributed multi-regional payment processing. What's the architecture needed for that? So that's what we started building with Restate.
+
+So the second thing -- that was a very long way of saying the second thing is Restate is really sort of a durable executions deck built from first principles, for low latency, serverless operations, high throughput, and just like really, really nice operations from the small to the large scale... Rather than saying, "Let's start with whatever database we have." I think in Temporal's case, when it came out of Uber, they started with Cassandra, and said, "Let's build a server that sort of like sits on top of Cassandra, and stores all state that it needs for coordination in there." And then you have like different pieces that you need to scale, you have a database that does actually a lot more than you really need for durable execution, but on the way also sort of sacrifices the potential for optimizing. Those are the differences, I would say.
+
+**Adam Stacoviak:** Gotcha. Built for speed, basically. That's what you're saying.
+
+**Stephan Ewen:** Built to be lightweight...
+
+**Jerod Santo:** Scale down.
+
+**Stephan Ewen:** Yeah. Scale down, scale up... And yeah, lightweight, simple to operate... I usually don't like to do this. I usually like to talk more about what makes Restate great than what makes other systems not great...
+
+**Jerod Santo:** Adam put you on the spot.
+
+**Adam Stacoviak:** Well, I think it's important -- well, if I'm gonna go out on a limb and say "This is the year", then you have to follow me. Okay? You have to follow me and you have to answer my questions...
+
+**Stephan Ewen:** Absolutely. Okay.
+
+**Adam Stacoviak:** ...because I'm reducing your marketing churn for, you just by nature.
+
+**Stephan Ewen:** Yeah. So I'll just say, if you look at the way Restate is built and it allows you to get started, then scale from there... If you say "Okay, I care about self-hosting this, because what I pipe through this is like critical data. It's not something I trust with some managed cloud. It really has to run in my account", I think the experience you get out of Restate is vastly different than what you get from many other systems. And that's because it's just been this very thoughtfully crafted stack from the very beginning, and not sort of incrementally evolved from --
+
+**Adam Stacoviak:** Morphed...
+
+**Stephan Ewen:** Yeah, from this database, and then that server... Yeah.
+
+**Adam Stacoviak:** \[01:10:08.19\] Right. If you're directly comparing to Temporal, which is an incumbent, which was spun off from, as you mentioned, Uber, and has different principles for which it was built on... You went back to first principles and said "Okay, if you want to get to the point where you can put this in almost everywhere you want to, you have to be low latency, you have to be fast..." These first principles you built on have to be there.
+
+**Stephan Ewen:** Yeah. And you can't have the requirement to first install a distributed database before you get started.
+
+**Adam Stacoviak:** What is the requirement? That's where I was going to go to. So it seems like its client, which is an SDK, essentially, inside your codebase, making calls to a server... What is the architecture, the infrastructure required?
+
+**Stephan Ewen:** So the Restate server, which is where the low-latency consensus log lives, and the thing that basically becomes the reverse proxy for your service - that thing has not really any requirement if you wanted to get started. It's a self-contained binary, it embeds its own distributed log, a RocksDB storage engine, its own consensus engine... The only thing, if you want to run it as a single node, is you need to give it a persistent disk. It's almost that -- let's say, if you would want to run SQLite or Postgres, it's a little bit like let's go back to the good old days where you download one binary, I just start it and it's actually running. It's actually good, there's nothing else you need to do. But then at the same point, it's also able to go from that single process that you start with a single binary to actually cluster up and build a distributed cluster... And there's a very interesting architecture in there. We build it basically for the cloud data age, where you would say any system that you run at scale should not really store its own data, but it should just make use of object stores for as much as it can... Because S3 and these systems - they're these bottomless, insanely durable and insanely cheap storage systems. So make use of that as much as you can to put a large chunk of your data.
+
+So that means while you may be working with the data on your individual nodes, you're not really required to safeguard it on the nodes, because you can recover it from S3 or an object store. So what Restate then does is it actually implements its log in such a way that it only uses this to really give you the very low latencies for the durable steps, and then in the background it incrementally moves data to S3, which makes the individual nodes fairly lightweight to operate.
+
+So to go back to your question, like "What are really the requirements when you want to run it?", if you want to run it on a single node, non, or a persistent volume if you actually want to run it like in production - if you want to run it in a distributed setup, give it an S3 bucket. Those are the requirements. If you want to use it from your code, the requirement in your code is to use the SDK, and to basically create a Restate entry point that Restate can connect to and where it can use a sort of durable invocation protocol that understands how to decode that. This sort of entry point is mimicking the popular frameworks; it's relatively close to Express.js if you're talking in the JavaScript world, and the Java world looks more like Spring Boot and so on.
+
+\[01:13:42.14\] And then within an individual durable function, service handler, you need to use the Restate context to say "Okay, I want to run this step and record it as durable. I want to create this as a durable promise for a persistent callback", or so. But otherwise, the structure of your code is very much the same as it used to be. So it's supposed to be as little invasive or to get as little in the way of how you used to do things as it can... Just sort of changing the paradigm as in -- because it has this fine, good durability for these operations, you can get rid of a lot of the sort of unhappy path code. There's still cases you need to treat, but mostly you still need to treat sort of like persistent errors that come from the application, and where you'd say "Okay, you're making a call to an API where you're not authorized. There's not really a way you can recover from this. You're just trying to do something you're not supposed to." And "Handle this." But don't worry about handling process failures, network failures, rate limits that bounce you back. Don't worry about many classes of race conditions, about the state being maintained in the database versus the logic that interacts with it in a function that where you don't know really "Did this go through or not?" If you put the state at the Restate handler, it's just going to be consistent for you. All of those things by keeping the structure of the code like those two, what you used to write.
+
+**Adam Stacoviak:** So I made you talk about things you don't like to talk about, except for maybe the architecture. That seems kind of fun to you. What is it that you do like to talk about when it comes to defining and describing Restate and why developers should consider it?
+
+**Stephan Ewen:** So no, what I don't like to talk about is competitors, in the sense of I don't want to say "Okay, I don't like this about a competitor. I don't like that about another competitor", because number one, I'm not an expert in those systems... I try to be honest. I look at them to the extent I need to look at them, but actually no deeper than I need to, because I've found this very liberating, to not have my judgment sort of clouded or sort of pre-biased by having looked at something. I feel if you look -- for example, if we look in depth at how Temporal would build their APIs and so on, there's a very good chance that "Oh yeah, I get it. This is why they did it. This makes sense", and so on, then there's a good chance you'll probably do it the same way, just because you've sort of seen this example, decoded it, understood it...
+
+**Adam Stacoviak:** Preconditioned.
+
+**Stephan Ewen:** You're preconditioned, exactly. If you don't do this --
+
+**Adam Stacoviak:** Yeah. It's like Schrodinger's Cat. Is the cat alive or dead in the box? We won't know until we look.
+
+**Stephan Ewen:** Maybe, yeah.
+
+**Adam Stacoviak:** It's both dead and alive, in the same moment.
+
+**Stephan Ewen:** But if you don't do that, you actually have a chance to do something -- to come up with your own creativity, and possibly do something better. So that's one of the reasons why I don't like to talk about them so much, because I'm absolutely not an expert. I look as much as I need to, but I usually don't try to go super-deep into these systems.
+
+And the second thing is -- I don't know, there's so much... I'd rather talk about good things than bad things. It's like more of a --
+
+**Adam Stacoviak:** Yeah. Good idea.
+
+**Stephan Ewen:** It's more fun to say nice things than bad things.
+
+**Adam Stacoviak:** I understand your discomfort then now. Definitely, it's always -- it can be tumultuous talking about competitors and what they do and what they don't do... I think in the context and the reason why the question is pertinent is because whenever -- like, to Jerod's point, you have a marketing challenge ahead of you, and I think it's because the idea of durability and idempotency is mostly well known, not always easily implemented, and there's options out there. And so when you sort of look at that challenge, you think, "Well, what could someone reach for? When would they reach for it? When does it make the most sense to reach for it?", and "Does it actually fit whenever they do try to implement it at scale, across different boundaries and whatnot?" And so when you compare that and you look at like "Well, NATS --" NATS is a whole different scenario, but they do similar things. And it's kind of funny because when you mentioned Flink, you were like "Well, it does this in a different way." And then you got to Restate because of your experience there and whatnot. Same thing with NATS. NATS does a lot of the similarity things where you're brokering messages, there's a lot of retries, there's a lot of key-value storing in there... It's like, a lot of that same principles, but it's not about durability, it's not about retries. And then you obviously have Temporal, you have Render who's trying to or going to do something like that in the same platform, which I just had that conversation with Anurag... And then you obviously have Restate and how you went back to first principles, versus being spun out of something, or what have you. So I think from a guide standpoint, you're the best suited guide in this conversation to explore those... Because Jerod and I can't do that for us.
+
+**Stephan Ewen:** \[01:18:24.08\] Yeah, absolutely. So if you want a quick summary, I'm very biased, but I think there's almost no reason to not reach for Restate. I think it really is this solution from first principles with amazing developer experience, with a very powerful abstraction that allows you to build what you can build with workflows and signals, but also so much more. And yeah, just the journey from the beginning, downloading the binary, then migrating, scaling out is an absolutely -- it's a great experience. I mean, the project is newer than other projects, so it will have a rough edge here or there, but it's also moving very quick. It's very good at reacting to community feedback fast... So I think it's a good choice. It has made a lot of users happy so far.
+
+**Adam Stacoviak:** Can we use maybe, Jerod, an example from our own application, to consider how we would pick up Restate? I know we publish episodes, right? We publish episodes, we often will have scenarios where the slug isn't right... We've had different scenarios where we had to do things in prod to fix something. It could be metadata, and you've got different checks before the publish process. Is there a way knowing what you know now about Restate, how you would consider implementing something like that to safeguard publishing episodes in a durable way?
+
+**Jerod Santo:** I've never really used one of these tools before, so it's difficult for me to say that. I do know just at a technical level that I do not believe Restate has an Elixir SDK, so we might be out in the yard..
+
+**Stephan Ewen:** An Elixir one? That's a good ask.
+
+**Jerod Santo:** \[laughs\]
+
+**Stephan Ewen:** Okay, maybe I can help you come up with an example here.
+
+**Jerod Santo:** Sure.
+
+**Stephan Ewen:** Let's say you're recording the episodes and then every time an episode is done -- no, let's do an AI thing here, or so.
+
+So you're building your chat where you can chat with an episode, like "Okay, tell me where did they talk about this", or "Tell me what episodes talked about these topics", and so on. So what you're doing whenever an episode is done - you're feeding it first through a model that transcribes the audio, then you're chunking it up, feed it through embeddings models stored maybe in a vector database, and then you have kind of a RAG style way of... You know, when a query comes, create the embedding, look up the similarity search in your vector database, feed it to the model to get the answer... For something like this, if you -- let's say you started just building the flow in let's say a Node.js application, like in a simpler way; you just said "Okay, here's the episode." I have something, it gets uploaded... Let's say you're uploading it to an S3 bucket and there's an event whenever something gets uploaded to this bucket; you have an event that represents this, and then it starts a Node.js script, or something like this. And this script is of the type that if it fails, somebody would have to restart it. Now let's say you're trying to implement that with Restate. I would say approach it the following way... The first thing is get a handle of Restate itself; there's a cloud service that you can use on our site, which has a free tier. Either go there, or just use one of these ways to run it yourself on a single machine with an EBS volume.
+
+\[01:22:06.14\] Then you have the server there. Then put your Node.js script maybe -- you can actually put it on something like Lambda or ECS; just like use a serverless option to host this. And then use the Restate SDK to define the entrypoint and tell Restate "Okay, hey, here's the service that you now should durably manage." So Restate will then go there and discover this and understand "Okay, hey, there's this (what what do we call it?) video transcriber, or video embedder service." And then Restate knows about this. And then you would go to your Amazon console and say "Okay, for this type of event I want to create a webhook to Restate", so that it makes an invocation to Restate and says "Okay, this thing has been updated." You know, the kind of event that would previously call directly your Node.js process or script, you actually make it an HTTP call to Restate, and Restate would then call your process.
+
+You've already gained one thing right away. You now basically have a reliable queue in front of it, right? Like, just that if you don't do anything special. So when the webhook comes, it's going to be acknowledged back, and Restate has this \[unintelligible 01:23:15.22\] of your process if it crashes, it will retry this... It will actually give you a nice observability; much more than you would get from your average message queue, about like individual retries, configuration about time-offs and back-offs and timelines and so on.
+
+As a next step, you would actually then go into your script and say "Okay, let's actually identify the steps where if something fails in this step, or after that step, I don't want it to go back." Like let's say forking the process that does the transcribing, or like calling the LLM to create the embeddings. You introduce then the Restate context that you get by using the Restate SDK and just say, "Okay let me wrap these API calls just with restate.run." That will capture the results of this durably, and basically turn -- you've now turned it basically into a workflow.
+
+Let's say you want to do something like parallelize the different steps. Maybe just typing this one by one through this embeddings model is a little tricky thing... You want to fan out. You could then go and say "Let me try and do the exact same thing I do in a regular Node process and just make a bunch of function calls record, like remember the promises", sort of a way to promise that all for those in the end join the results, and put those in the database. You can do exactly that in your code. Just, again, anchor this in the Restate context, so you get like this durable parallelization, durable stuff like scatter, gather, and so on. And so you would then incrementally sort rewrite your code to say "Okay, let's actually make this step durable, let's make that step durable, and that step durable."
+
+Say as a next thing maybe one of your folks wants to approve it before it really goes out. So you then possibly -- you "Let's do that in the simplest most possible way", which is we create an awakeable or a durable promise in Restate and say "Okay, somebody needs to complete this actively", like send an event, make an HTTP caller to complete this and say "Okay, this is approved, go through" or "No, this is not approved. Abort." You can then use, for example -- you could put the result, the transcription just in Restate, save it, and somebody could look at it from the UI and then say "Okay, yeah, I'm making an API call here to approve this", and it continues. And so you can then incrementally rebuild your process into durable steps.
+
+As the next thing you could then, for example, take it and migrate it from a long-running process to a Lambda function. Because one of the nice things you have with durable execution is when it's waiting for something else to happen, it can actually just make this thing go away. Because with durable execution, it knows how to recover back to the place where it was by replaying the history of durable steps.
+
+\[01:26:14.28\] So you could then say -- you know, if you're on vacation and you approve it a week later, you don't have some process running and waiting for it. It's just like, it's going to go away, and when the approval finally comes, it's going to come back, use the durable steps to replay back to the point, and then to the remaining steps. And so typically, folks would incrementally then rework their non-durable services; first connect them to Restate to basically get the equivalent of a durable queue, and then incrementally rework it and say "Okay, I want to add durable steps here, maybe parallelization, maybe a signal", and...
+
+**Jerod Santo:** Yeah.
+
+**Stephan Ewen:** ...I think that's typically how you'd approach it.
+
+**Jerod Santo:** That makes a lot of sense. I do see also you have some guides on the website about how to implement certain things... I'm curious about the observability bit. Is that a part of your hosted offering? Is that a part of the open source project? How does the business end fit in, and is observability part of that open core sort of thing?
+
+**Stephan Ewen:** Yeah. So at the moment, what you get in the open source is very broad. You get in the open source, compared to the hosted offering, pretty much everything except the fact that you would self host it, and the whole authentication and API tokens and so on that exists only in the managed offering. But other than that, we've started with an open source first approach, so the open source has pretty much the full suite at the moment. The observability - there's two things about observability and Restate. Number one, it can actually give you an amazing amount of observability itself out of the box, because it funnels all these durable steps through its consensus logs, it has all the information of what happened... And it's not just function call, but like to the granularity of "Here's a step that happened" or "It's actually failed." "This is the last step that completed in this failure, and since then I've retried so and so many times, and this is the last exception I've seen." It has all that information available, because it also is connected to the service and it understands, "Okay, what type of errors are happening? Is this a retriable error or not?" And it gives you access to all that observability data in its own UI. It's actually a fascinating way that this is implemented. If I can show everyone like one or two technical details...
+
+**Jerod Santo:** Sure.
+
+**Stephan Ewen:** So there's this durable log that recalls all the actions, then everything is indexed into RocksDB instances to sort of retain it in a scalable way, and we've built a SQL query engine using the Data Fusion project around this, that allows you to basically do SQL queries against all of that invocation, and transaction journal state, and so on. So what the UI actually does is basically issue SQL queries. It's almost like back to the good old days when all your state was in a single Postgres database, and if you wanted to find out why your application is stuck, you'd just open the SQL shell and start querying. And we kind of lost that, because we went into distributed microservices, and if you want to find out what happened, you now have to do like murder mystery with 20 services.
+
+**Jerod Santo:** Yeah. You're bringing it back.
+
+**Stephan Ewen:** And we're kind of bringing it back, yeah. Like, SQL query for the win, for your distributed application state. So this is one of the things - you get an amazing amount of insights right out of just the Restate journal. The second thing is because all the operations go through there, Restate can also just out of the box generate open telemetry traces and spans for you. So if you give it an OTEL endpoint that it should push those two, it would just give you the traces right away, without you needing to configure anything. You can then extend it and augment it with your own traces. So those are the two things you can do.
+
+**Jerod Santo:** \[01:29:58.16\] Cool. And so the business end is basically cloud hosting for Restate.
+
+**Stephan Ewen:** The business ends going to be a lot more than that, but...
+
+**Jerod Santo:** Okay. What's it gonna look like?
+
+**Stephan Ewen:** I don't think we can go into this yet. Interview us again in six months. It's not ready to be to be announced. But what's available right now is -- most of it, what we have built, is also the open source. So yes, on the on the business side we do currently only hosting.
+
+**Jerod Santo:** For the next six months or so.
+
+**Stephan Ewen:** Maybe.
+
+**Jerod Santo:** \[laughs\] Fair enough. Cool. Well, I think it sounds like a really cool system. I'm excited about this new world of durable execution functions, and some way to slap a name on that that brings all of the junior engineers to the yard, along with us seasoned engineers who have felt these pains for all these years. You know, like the serverless folk did; they just said it's serverless, and then like "Oh, okay, cool. Serverless. Maybe I should try it." I feel like Restate and friends need some sort of a marketing term to just simplify the overall concept of what you all are building. But I do think it's very interesting tech, and very promising. I do like the term resilient apps, so I think maybe something with resiliency involved. But that's all for me. Adam, any other questions from you before we let him go?
+
+**Adam Stacoviak:** I'm almost liking durable, personally. I don't know if you want to wordsmith that a little bit here, but I like durable. I think that seems to have \[unintelligible 01:31:36.29\]
+
+**Stephan Ewen:** Yeah. So if you're interested, we've actually gone through a few iterations. We've started with something, just calling it durable async await, because in many ways, that's what is underneath the hook.
+
+**Jerod Santo:** Yes.
+
+**Stephan Ewen:** It's like durable asynchronous operations. A function invocation as an asynchronous operation made durable \[unintelligible 01:31:56.28\] asynchronous maybe API call, with a durable result...
+
+**Jerod Santo:** Right.
+
+**Stephan Ewen:** And there were some sort of expert programmers that were like "Oh, that's really cool. I get it. It's like distributed durable event loops. That's very cool." But then 90% of the folks did not get that. Then we just went with durable execution, and then it turns out that's a very -- that's a term that's maybe increasingly more recognized, but it also undersells a little bit what we \[unintelligible 01:32:22.22\] because folks actually think "Oh yeah, so it's just the same thing as Temporal", but actually it does a bit more.
+
+So yeah, we're still on the wordsmithing side. Yes, distributed durability, resilient apps, resilient distributed state management... There's so many things on the table. At the moment -- I used this earlier in the talk... It's like stateful durable functions, or something we've used. I think this is maybe increasingly getting recognized because of a lot of the efforts that, let's say, Cloudflare does with its work, with like durable objects. There's a construct in Restate called a virtual object that has a surprising amount of similarities with durable objects, and we would have called it durable object if that term hadn't been trademarked by Cloudflare.
+
+**Jerod Santo:** Hm... Jerks.
+
+**Stephan Ewen:** And Azure Durable Functions is probably even closer to Restate than Temporal. So if you -- I think you can actually think like Temporal, Azure Durable Functions, and so Restate is a bit more than that. I think it combines a bit more of like orchestration and stateful logic, even in a more flexible way than durable functions does. But yeah, so stateful durable functions is where I've currently landed at. But look, it's a journey... I think honestly even Temporal hasn't figured that out after five years.
+
+**Jerod Santo:** Well, that's why I said, it's a challenge that I think maybe Restate shouldn't solve it, but I feel like maybe everybody who's in this category... There's a missing category. It's almost like a style of application, or an architecture, where it's like "Well, what architecture is it?" Well, it's model view controller. Okay, it's MVC. I can build an MVC style app. Where this is like -- I don't know what to call it. I'm missing a word. It's almost it's Restate style, or something like -- maybe you have to term it after yourself if you want to really own the market.
+
+**Stephan Ewen:** \[01:34:21.18\] It's durable function style, or it's...
+
+**Jerod Santo:** Yeah. Durable functions doesn't speak to me personally at all. It sounds really boring. But that's just me. And maybe it's working on other folks... Adam likes durable. Durable to me just sounds like "Cool, it's not going to break."
+
+**Adam Stacoviak:** Stateful durable functions... That's what you said? Is that right?
+
+**Stephan Ewen:** That's what I said earlier, yeah.
+
+**Jerod Santo:** Stateful durable functions.
+
+**Adam Stacoviak:** SDF.
+
+**Jerod Santo:** SDF style.
+
+**Adam Stacoviak:** Make an acronym, or something like that, I don't know. Like TDD. SDF.
+
+**Stephan Ewen:** MVC, yes.
+
+**Jerod Santo:** Yeah. I mean, model view controller doesn't have any sort of appeal to it either on its face... So maybe that wasn't a bad example. Anyways... We could continue to workshop it until we're blue in the face, but obviously you've been working on it longer than Adam and I have.
+
+**Adam Stacoviak:** Hey, listen, this is the year. This is the year of it. I'm just saying.
+
+**Jerod Santo:** The year of the what?
+
+**Adam Stacoviak:** The year of whatever this is...
+
+**Jerod Santo:** Oh, the durable function.
+
+**Adam Stacoviak:** Whatever this is, it's the year.
+
+**Stephan Ewen:** Yeah, I feel -- there's something about durable in itself that's not recognized by lots of folks. I think you actually asked about that earlier as well... Like, "What does durable really mean?" Durable is maybe a stronger emphasis on persistent, or so... I think there's something to be said about resilience.
+
+**Jerod Santo:** I think resilience is a much more attractive word generally speaking, and one that to me calls and says "Is your app resilient?" I'm like "Ooh, I don't know if it is."
+
+**Adam Stacoviak:** "Ooh, I want a resilient application."
+
+**Jerod Santo:** "I want resilience."
+
+**Adam Stacoviak:** Because durability can show somewhere, whereas resilience is like "You know what? No matter what happens, I'm going to succeed. I'm going to try until I --"
+
+**Jerod Santo:** Until I bounce back. Yeah.
+
+**Stephan Ewen:** Yeah, and I think you're right in the sense of durability is a means to an end. It's very implementation detail, if you wish.
+
+**Jerod Santo:** Right.
+
+**Stephan Ewen:** Restate achieves resilience by doing a lot of fine-grained durable operations, which make it easy to bring it back to a consistent state, and that drives resilience.
+
+**Jerod Santo:** There you go. Now you're getting your messaging down. Love it.
+
+**Adam Stacoviak:** Now we're deep... Listen, hey, we have a fun place to hang. It's called Zulip.
+
+**Jerod Santo:** Oh, that's true.
+
+**Adam Stacoviak:** Go to changelog.com/community, join us in there, and then if you have some ideas about this name or this --
+
+**Jerod Santo:** Wordsmith this with us.
+
+**Adam Stacoviak:** ...this world that Stephan is creating, then pile on, share your thoughts, all the good stuff. What else? What's left? Anything left unsaid about this durable, resilient world we're going to live in?
+
+**Stephan Ewen:** What's unsaid about the durable world that we live in? I think it's inevitable to come. The question is mostly in what shape is it coming? I think it's been worked on, actually, from multiple dimensions. They're folks like us that work on this from the Restate side, like "Here's a lightweight durable log that sort of is easy to integrate with other functions." I think that sort of the serverless folks and the Wasm folks are working on that from a different side, saying "Okay, hey, let's compile everything to Wasm and let the system sort of use the Wasm interpreter to snapshot things." And I think there's folks that kind of use container engines to implement this. I think what they all share is just the understanding of it's completely unsustainable to not have anything like this. It's harder. It's hard. It gets increasingly more important the more moving parts and the more asynchronous processes you have, and I think if we all believe what the AI people tell us, that like 80 percent of all this is anyway going to be some agentic stuff in two years... And I think you've just created an even bigger problem and an even bigger need for this type of systems. So I think this is coming, in one shape or the other.
+
+**Adam Stacoviak:** It's inevitable.
+
+**Stephan Ewen:** This is our bet of how to best achieve it, and it's going to be fun to see what happens.
+
+**Adam Stacoviak:** There you go. Alright, Stephan, well, thank you so much for sharing the journey, sharing the love, sharing the things... We appreciate you.
+
+**Stephan Ewen:** Thanks for having me. Cheers.

The power of the button (Interview)_transcript.txt

@@ -0,0 +1,545 @@
+**Jerod Santo:** So we're here with Rachel Plotnick, an associate professor of cinema and media studies. Don't get us talking about movies...
+
+**Adam Stacoviak:** Oh, my gosh...
+
+**Jerod Santo:** ...at Indiana University, and the author of a book called "Power Button: A History of Pleasure, Panic, and the Politics of Pushing." Rachel, we are here to talk buttons with you. Welcome!
+
+**Rachel Plotnick:** Thank you so much for having me.
+
+**Jerod Santo:** Happy to have you. I've found you by way of an IEEE Spectrum article about touchscreens. They're going out. Tactile controls are coming back. The re-buttonization expertise of yours is in demand... How did we get here? It was like, buttons were all we had, and then we went no buttons, and now it's like "Hey, buttons might be cool..."
+
+**Rachel Plotnick:** Yeah, I think there's kind of these interesting pendulum swings that happen with these different interfaces where something becomes really en vogue for a while, then not so much, then again... And the button is really a very old and enduring technology. The touchscreen is obviously much newer... For the last couple of decades I think there became this kind of what I call touchscreen mania, where it was like "Alright, everything's got to be a touchscreen now", and people got really hyped around that. But over time I think we've begun to see some of the cracks and flaws and problems with doing everything that way, and people are starting to maybe dial back a little bit and think about when's the right situation for a touchscreen and when's the right situation for a physical button.
+
+**Jerod Santo:** I think that pendulum swing that you're describing finds its way into all kinds of stuff that we do. And it's like "New technology, or new shiny thing - let's use it everywhere. Let's throw out the old, and let's only do the new. It's going to fix everything." And then we go that route a little while, and start to see the cracks or the misapplications where it's like "Actually, it's not so great in this particular..." And we learn, and we realize that broad brush could be more precise. But then sometimes we pendulum swing back the other way, and throw out the new for the old... But eventually, I think we hone in on what is practical and useful, and not just in push buttons... But in your book, you talk about kind of the history even of the button, and knobs, and switches... Can you give us, you know, not the whole history, but maybe a primer for all of human history for us pushing buttons? I assume it started somewhere around the Industrial Revolution, or I don't know, when we had things that could be on the other side of a button.
+
+**Rachel Plotnick:** Yeah, I'm going to dial that history down to just, you know, 60 seconds.
+
+**Jerod Santo:** \[laughs\] Thank you.
+
+**Rachel Plotnick:** We can condense that history. I think it's a pretty fascinating story. You're right that it dates back to electrification and industrialization in the late 1800s, early 1900s... And at that moment people were starting to deliberate, "How are we going to control our machines in a way that's more effortless, that's more automatic? How do we initiate a circuit, turn it on and off, open and close that circuit?" And there were a variety of answers to that problem, various kinds of switches and pulls and throws and toggles... But the button kind of emerged as one of the standard ways of doing those kinds of activities.
+
+And a lot of that had to do with reducing the effort that it took, because you'd have to pull a heavy cord, or pull a lever, do something that was pretty effortful... But also, I think there was this idea around accessibility that buttons were supposedly easier to use. And one of the most fascinating things was how they were advertised to women and children, because they supposedly were not smart enough or strong enough to operate machines otherwise... So there was kind of this talk of the lowest common denominator, so easy a child could do it; anyone can push a button.
+
+**Adam Stacoviak:** Not cool. Not cool.
+
+**Jerod Santo:** Not cool... \[laughs\]
+
+**Adam Stacoviak:** I do like the simplicity of a button, and I do get the move to that... I'm thinking of things like cars, you know? Obviously, the real answer is like, I've got this button in my vehicle where it auto-turns the engine on and off. So my engine has this where if I stop at a stoplight, it auto-offs, and there's a button to turn that off.
+
+**Jerod Santo:** To turn that feature off.
+
+**Adam Stacoviak:** Yes. And thankfully, it's not behind a screen. But as you were describing that, I'm in this state of like "Well, I don't want a screen, I want a button." But then you're making me think, "Well, that button really is just a physical button that's still just like the same button in the screen." Because if the screen doesn't work right or the operating system of whatever makes the screen work - does that mean that the button can't work anymore? That's the concern I would have.
+
+**Rachel Plotnick:** \[08:02\] Yeah. I think it's a really interesting tension between the flexibility of software, which is versatile, and hey, you can update that button, and today that button does one thing, tomorrow it can do another thing, and the physical button's always going to do that one thing.
+
+**Adam Stacoviak:** Right.
+
+**Rachel Plotnick:** But on the other hand there is kind of that simplicity factor of "I can always go to this button. I know that it has a discrete task that it's meant to operate." And so I do think it becomes a question of "When is the right interface for the right kind of activity?" And as you said, driving in particular I think is one of those cases where we don't want a lot of cognitive overload, and we don't want to have to be looking around, trying to figure out "What am I supposed to press or push or tap?" when we're also trying to do the act of driving.
+
+**Adam Stacoviak:** And don't change the core feature of a car, you know? If the wheel turns, that's what makes the wheels move. Don't make the wheel turn do the turn signal. Like, that would be silly, right? Keep the common feature set of that in place. I'm thinking of the Cybertruck. The Cybertruck is not so much anti this, but the turn -- have either of you ever driven a Cybertruck?
+
+**Jerod Santo:** No.
+
+**Adam Stacoviak:** Okay. I've driven it one time. A friend of mine has one... And it's just so crazy, because it has -- I don't know what the steering's called. I'm not a fan, so that I know all the terms and whatnot... But the steering, the wheel does not turn like a normal wheel, or a traditional wheel, I should say, where you have to turn, turn, turn to make your left-hand or right-hand turn. It is very much just a very little -- and if you can see the camera, I'm just doing a little motion. It's not like big-big.
+
+**Jerod Santo:** Like a race car style.
+
+**Adam Stacoviak:** Yeah, it's very -- and it's adaptive. But to drive it, you have to relearn how to drive it. When you take your foot off the gas - which is a button, right? A pedal's a button. It's not some sort of like computer screen switch. It's literally a foot pedal, and it's a toggle for you to analog how to make the vehicle move forward, or stop, or whatever. And as soon as you take your foot off the accelerator, it begins to decelerate. Not like a vehicle, like a non-Cybertruck; it begins to brake, basically. It's basically the way the engine works. It's a natural brake. So you actually don't have to brake much in a Cybertruck, which is kind of cool, but you have to relearn how to drive. And I guess I'm bringing that up because sometimes you'll change the common feature set of what are typically analog buttons or pedals or whatnot, and you have to relearn how to use them in this new, modern way because somebody decides to change it.
+
+**Jerod Santo:** Yeah. I think that that is some of the tension, like you said, with the flexibility, right? It's like, we can change it because it's now actually disconnected from these previously mechanical things that this button was doing. But should we change it, or should we have it at all? And Rachel, when did we decide -- I would peg it to the iPhone, but maybe there was prior art to be like "Buttons aren't cool" all of a sudden. I do remember the BlackBerry was cool. And this is like 2000 -- pre-iPhone. '06, right? Because the iPhone was '07, I think. And it's like BlackBerry was very cool, and it had all the buttons; it's kind of a power user's thing, and you could type real fast, because you had the whole keyboard there... And then the iPhone came out, zero buttons, or I guess it had the home button, and then like the power toggle on the right. But it was like a single pane of glass, as they say. And that might've been -- okay, now I'm just answering the question for you, which... You're the expert here.
+
+**Rachel Plotnick:** This is great. You're doing my job.
+
+**Jerod Santo:** I'm thinking, is that it? Is that the one?
+
+**Rachel Plotnick:** I can just sit here.
+
+**Jerod Santo:** Yeah, yeah. Is that when it happened, or was there a different one?
+
+**Rachel Plotnick:** No, I think that that's probably right. I think that minimalist aesthetic took hold, where it was like "Alright, everything's going to be smooth, and flat, and glass", and I think the iPhone was really a big factor in that. Probably we were seeing this a little bit earlier in popular culture even. I often point to the movie Minority Report in 2001, where Tom Cruise was gesturing with his hands, and... I think people were starting to get excited about the idea of "How do we manipulate our data in a way that either minimally involves touching it, or doesn't involve touching it at all?" And of course, we've seen that around VR as well, people getting interested in the idea of kind of just waving our hands around in the air. So that concept's been around for a long time. But I do think -- you're right, between 2007 and 2009, that's when this kind of minimalist touchscreen interface took hold. And that's, in fact, when I started working on the topic. Everyone was saying "It's the death of the button. We're not even going to push buttons anymore in the next 10, 20 years." And I was really fascinated by "Is society being gripped by this wholesale shift in the way that we interact with technology?" That seemed like such an interesting question.
+
+**Adam Stacoviak:** \[12:29\] Where did you begin your research then? So you got curious... What was the next step?
+
+**Rachel Plotnick:** Yeah, it kind of took a while. I mean, it wasn't like I went to graduate school and said "You know, I'm going to research push buttons." That's not really my origin story. But I thought that that death of the button was kind of what sparked it, and then I thought "Alright, as I'm starting to think about pushing buttons, I realize that everywhere we turn, we see people doing this activity, whether it's your coffee maker, or going up and down an elevator, driving your car, social media buttons... What are we doing all day but just pushing buttons?" That seemed like such an interesting way to talk about how people interact on a daily basis with their machines. And as I started digging into it, I just kind of became gripped with this question of "How did we get to be this way? Where did this technology come from? What were some of the cultural issues around pushing buttons?" And that led me to kind of this historical approach to say "Okay, I want to understand some of the earliest buttons and societal fantasies and fears around that", and next thing I knew I was down this crazy wormhole trying to figure out all these things about buttons, and it was pretty interesting.
+
+**Adam Stacoviak:** I've gotta imagine color matters, right? Like, certain buttons should be certain colors. Certain buttons should be concealed behind a piece of glass and a key... You must turn the key to push the button, or turn a key to open the thing to be able to turn the key to push the button...
+
+**Jerod Santo:** Right. No Homer Simpson moments...
+
+**Adam Stacoviak:** Yes, exactly. Or other moments, but sure.
+
+**Rachel Plotnick:** Yeah, someone just actually emailed me to tell me about an experience at Universal Studios in Hollywood that involves the Simpsons, and I guess there is a nuclear reactor button there, and they've got like a countdown clock, how long has it been since someone's pushed the button...
+
+**Jerod Santo:** \[laughs\]
+
+**Rachel Plotnick:** And you're totally right, Adam, that I think there's that psychology around button pushing that's so fascinating. As soon as you tell someone you can't push the button, all they want to do is push it, right? And of course, that gets back to these kind of grand fears of nuclear warfare, the big red button... But I think it's also just in our nature that we want to do things that we're told not to do.
+
+**Jerod Santo:** Yeah.
+
+**Adam Stacoviak:** You're so right, though, about a lot of the activity we do in a day is pushing a button. I wonder if it would be a good exercise - or maybe not - for a podcast to imagine most of a day... Like, what's the very first thing you do in the morning that you might push a button? Well, I have a toothbrush that has a button, that begins to, I don't know, vibrate... I don't know what it does. It does something to make my teeth get better. \[laughter\] My doctor recommended it. It's dentist-approved, of course. So that's at least one early morning button. I think maybe the next button might be like the lever from my shower. Like, is that a button? Or is it a lever? Are levers buttons? What do you think?
+
+**Rachel Plotnick:** You know, the whole lever/key/button/switch distinction is very gray. That was a big problem for me. As soon as I started to research buttons, I was like "Well, what really is a button?" We could get pretty philosophical about that.
+
+**Jerod Santo:** You got existential on it, you know?
+
+**Rachel Plotnick:** I decided to just say "Okay, if someone calls it a button, I'm calling it a button. If they call it something else, then I'm not calling it a button."
+
+**Adam Stacoviak:** Let me just say, if I had to push a button to turn my shower on, I would love it if I could just push a button that says - I'm just guessing at a temperature - 100 degrees. Fahrenheit, of course. Or maybe give me 92 degrees. I want like a slightly cooler shower, because I don't want to be so hot... I would love that button, where it's like a precision button, versus like a fine-tuning the where is the -- and I've got one of those ones that begins at 12 o'clock and you turn it to a certain part that's like "Hey, that's your hot temperature."
+
+**Jerod Santo:** \[15:58\] And it's cold, cold, cold, cold, cold, and then it's scalding hot.
+
+**Adam Stacoviak:** That's right. Yes.
+
+**Jerod Santo:** There's a whole like button and switch and design aspect to this with user experience, and the failure states of something like that, where it doesn't have memory. So you have to build in the memory of where exactly -- like how many degrees do you turn that sucker around in order to get exactly the way you like it? Whereas in software - and you can do this in hardware design as well - you can actually create switches or levers or buttons where it can either have helps for you, like clicks. It can click into a certain spot and you can know six clicks is perfect.
+
+**Adam Stacoviak:** Feedback. Haptics.
+
+**Jerod Santo:** Or it can have memory, where you're like "Just go to the six click thing", and it knows to do that. So there's the -- I think affordances is probably the word that designers use in order to help us get what we want out of a particular button. And yeah, I think the shower one is a good example of that. Also toasters... I mean, you can burn toast very easily because you set it to the wrong temp... When it's I like, I like it at this degrees, for this long; just do that every time.
+
+**Adam Stacoviak:** I don't know about you, but my toaster goes to 10, and I never go to 10. I'm always at like three. I'm like "Who goes to 10?"
+
+**Rachel Plotnick:** No.
+
+**Adam Stacoviak:** What are you doing to go to 10? What kind of person are you?
+
+**Rachel Plotnick:** And I think what you're saying - you know, that kind of gets into the fact that I always tell people that buttons seem simple, but actually they're really complex. Because when you think about knowing which button to push, at the right moment, who should be pushing it, especially... You know, imagine going into a DJ booth, or the cockpit of an airplane, or something... It's not going to be super-intuitive to you if you're not really familiar with that system, how to operate any of those things. So I think the button's almost kind of seductive in that it seems really simplistic, but in a lot of situations it's actually pretty complicated.
+
+**Adam Stacoviak:** Especially a cockpit. I mean, you've got toggles, you've got flips and switches, then you've got the push thing... I think that's the one where they've got the throttle, maybe even the yaw, or whatnot. I don't know, I'm not a pilot. Mitchell might know, Jerod, since I was just editing the show with Mitchell...
+
+**Jerod Santo:** Right.
+
+**Adam Stacoviak:** You've got lots of different forms of buttons in a cockpit. Like, there's no -- and then what happens if this thing is no longer a binary on or off toggle? Some of it you might have to go back and forth 10 times to get a new functionality. Like, you've got this button that is not a button, it's more of a toggle, that now the feature for this thing has to be three different options versus two... But you've got to swap the button out. You've got to get an actual button, or a slider, or something different.
+
+**Jerod Santo:** I think that's a case where buttons are actually intimidating. When you look in a cockpit, you're like "This does not look simple." But I think the allure - and Rachel, go back into some of your research as well, because I read an excerpt from your book where you were talking about some of the history of the sales around, like the selling of the idea, this fantasy of like "At the push of a button, all your wildest dreams will come true." And of course, we also have the fear side of "If I push this button, a bomb will go off." There's sometimes hidden complexity behind a button, sometimes it's as simple as switching something off and on... But what did they initially, when buttons became like in vogue, or like, whoever was out there selling buttons, what were they telling us? Like, what were we buying?
+
+**Rachel Plotnick:** Yeah, you're absolutely right that there was a lot of kind of consumer culture around push buttons initially. They were really a way to sell products. So you had Kodak with their amateur cameras, their slogan was "You press the button, we do the rest." It was very much about "Hey, anyone can be a photographer, and now you don't need to worry about how to develop your film, or very much at all", which was very threatening to professional photographers who were used to kind of people not having the skills to be able to access that profession.
+
+Other things like vending machines and elevators were some of the earliest button usages as well. And all of that, really, the rhetoric was around this idea of "Anyone can do it. Button pushing is so easy." And I think, Jerod, to your point, there was the flip side of that then; there was the panic of "Oh my gosh, if anyone can push a button, what does that mean for our society?" Because then you do have that question of skill. We've kind of opened up the ranks of these technologies to everybody...
+
+**Jerod Santo:** Right...
+
+**Rachel Plotnick:** ...but also, what happens if this button falls into the wrong hand, and now -- whose finger's on the button, and do we need to worry about that? So you have kind of this pleasure and panic at work at the same time.
+
+**Jerod Santo:** \[20:13\] Yeah. We've experienced some of that with our children. We take for granted some things that adults know, and that need to be taught, like some obvious things... And one thing that I took for granted with kids - not that they wanted to push buttons, because I know that that's like inherent in us, is like the desire to push a button, and we can talk more about that... But they have no thought of recourse on the other side of a button, and they have no idea what a button does, and that's not going to stop them. As an adult, I think "I don't know what that button does. I shouldn't push it." And I didn't think I'd have to like explicitly teach that to every one of my children. But you do. You have to tell them. Because they're going to push the button. Like, it's just there. And there's an education of like "That button could have so many things behind it that you have no idea what they are. If you don't know what a button does, don't push it." And you've got to teach kids that. I'm just now sharing this for future parents to know - yes, you're going to have to tell your kids not to. And here's the kicker... Even after you tell them not to, they're still going to push it.
+
+**Rachel Plotnick:** Oh, yeah.
+
+**Jerod Santo:** But at least you've done your job. At least you've warned them.
+
+**Rachel Plotnick:** Yeah, it was kind of amazing to me that people were complaining about this even in 1905. You had kids who were running up to people's doorbells and just ringing them when they weren't supposed to, or they'd get behind the wheel of the car and honk the horn... So I guess this has always been a problem, and I have kids and find myself in this same situation, too. And it's an issue of control, really. Like, my son used to get up at four in the morning just to go turn on the TV, so then we started hiding all the television remotes... And then we'd forget where we put them, we're opening drawers... So who's got the button? You know, it's all about power in the house.
+
+**Jerod Santo:** Oh, man. The button. Who's got the button? Is that from a movie? I don't think I know that phrase.
+
+**Adam Stacoviak:** Well, there's a book that you should get and read... It's called "The good for nothing button." I won't tell you the story, but it's a good book for kids.
+
+**Rachel Plotnick:** Nice.
+
+**Adam Stacoviak:** The good for nothing button.
+
+**Jerod Santo:** The good for nothing button. Okay.
+
+**Adam Stacoviak:** It's about pushing buttons.
+
+**Jerod Santo:** Well, they've gotta be good for something. So Rachel, when "The button is dead" narrative began to spread, and you got into this, did you believe it at first? Like, were you -- because now we're starting to think, and we're seeing trends the other way, like "Yeah, we overstepped. We need buttons for some things, namely in a car, for instance..." But when that -- because I kind of bought it. Like, I was ready to be done with wires, buttons... Internal combustion engines - I'm still cool with getting rid of those... But I'm just thinking "Buttons - let's stick them around" now. Did you buy the hype of "It's going to be touchscreens for everything"? Because I kind of bought it, honestly.
+
+**Rachel Plotnick:** I think I must've always been a bit of a skeptic, because I guess I could have gone the other way and research touchscreens, and been like "This is the future. Let's look at this!" But something instinctually, I guess, told me "Hey, let's look at buttons." And I think that certainly bore out. And maybe part of it is nostalgia now, part of it is a society that's become a little bit burnt out on screens, and is tired of having to look at screens all the time... But I think even then I had the sense in 2009 of -- anytime people say the death of something is coming, I'm always a little bit skeptical, because that seems so dramatic. And we've seen with many technologies that they tend to come back over time. Like, look at all the people into vinyl records now. So I'm not convinced that a lot of things really ever die out.
+
+**Jerod Santo:** Yeah, "Video killed the radio star", that kind of thing. It's like nah, we just layer on, don't we? We just add more... We don't necessarily kill things. Of course, mediums do come and go in terms of their popularity and usage...
+
+**Break**: \[23:40\]
+
+**Jerod Santo:** But the touchscreen has a lot going for it. It also has a lot of drawbacks. However, I would say that over the last -- is it 15? ...almost 20 years since the first iPhone, that the touchscreen, like the software button, that's really just not a real button, it's a touchscreen, but it's mimicking a button - it has gotten a lot better. I mean, they've gotten a lot of ways to provide -- what's it called, haptic feedback? It even feels like you're pressing it now... I guess that's the haptic feedback. There's sound... There's lots of things that they've done to kind of simulate a button press that makes it feel more real to us, but I think it doesn't really solve the underlying issue. Do you think that's fair to say? Well, maybe I'll ask you what is the underlying issue. Why do we still need them, in your opinion?
+
+**Rachel Plotnick:** I think that you're right, there are a lot more kind of affordances to what a digital button can do now, and thinking about sound and feel and all of that... But at the end of the day, it's still a visual technology, and that's what I always remind people, is that we call it a touchscreen, but it's a screen; it's meant for looking at. And if you think about trying to operate your phone or your car touchscreen with your eyes closed, you're not going to get that far. And if you're blind, you're definitely not going to get that far.
+
+And so I think there's a real accessibility issue around there, questions of safety, and the fact that at the end of the day those buttons are really only going to make sense if you're staring at them. And when we think about the various kinds of mechanical buttons that we use in our daily lives, a lot of those can be operated just by knowing where they are, feeling around, and finding where the button is. And that lends itself to a certain kind of experience that's quite different than the touchscreen.
+
+The other thing too I think is just the limited sort of tactile experience you have with digital buttons. Like, if I'm looking at clothes, I want to touch them and feel what the fabric feels like. But at the end of the day, you're always just going to be touching glass. Buttons, I think, can give us a little bit more nuance in terms of their textures and their qualities.
+
+**Adam Stacoviak:** I'm thinking about the tried and true DVD or Blu-ray remote kind of thing. It's not maybe something everybody would push, but I still have a Blu-ray player in my house. I'm that kind of crazy person where I still like physical media; not in all cases, but in some cases. But my Blu-ray remote - I've got to memorize where the Pause button is, because it's usually dark when you're watching a movie, right? You don't have a good visual, so you don't have full light... And then the buttons aren't always illuminated buttons either, so you don't know where they're at, and so you have to sort of find it in the dark. If it was maybe a screen, it's always moving, or I've got to announce the screen, now it's bright, people's eyes are changing, their vision's changing... Whereas the typical remote is the remote. But I know where the Play and Pause button is. Sometimes I push the Menu button, and that's the sad part, because it ruins the movie.
+
+**Jerod Santo:** \[laughs\]
+
+**Adam Stacoviak:** But anyways, but I've got it memorized. And it's got a little -- kind of like a little knob, or a little feeling on it. Like, a little extra piece, a little dimple or something like that, that tells me "This is the Play button" or "This is the Pause button. It's okay to push." So I kind of like that physical touch feeling, where I don't have to have a visual of it. I can just feel it, and there you go. Or it could be Braille.
+
+**Jerod Santo:** Yeah, I think that's the nut of it, is like the physical feedback of it being right or wrong, without your eyes being involved. And the Apple TV remote, by the way, is terrible at this, because Apple decided -- I think Jony Ive decided unilaterally for all of us that we didn't really need any tactile feedback on our remotes. And the Apple TV remote - they've gotten better over the years. But if you go back to the original one, it's this Pez dispenser... Actually smaller than that. It's like a pack of gum, and there's really not much to feel on it. It's symmetrical, so it could be upside down or not, you don't know... And there's one physical button, I think the Menu button... Or maybe the Apple -- no, that's a couple buttons now. But they're few and far between. You can't tell them apart, because there's no distinction, like you were mentioning, Adam, like the Braille, or like the little protrusions you can put to differentiate the two buttons without looking at them...
+
+\[30:23\] Long story short, you almost have to always look at it, or just ruin the movie, or just like take 50/50 chances. Like, "I might be exiting this app, but here we go..." I'm trying to pause it so I go to the bathroom, but I'm going to just crank the volume up on accident. That's a terrible one. And because - perfect situation, you're in a dark room, and you just want like one more ounce of volume, but you end up exiting the app, because there's no feel to those buttons. And so I guess, to Rachel's point earlier, even buttons themselves are complicated. Like, you can do a button poorly, but a touch screen button is almost de facto poor in that context.
+
+**Rachel Plotnick:** Yeah. It's so situation-specific, I think... If we think about driving and having to keep our eyes on the road, or we think about being in a darkened room, those are times when we just need things at hand, and we don't want to have to be moving back and forth between these different systems. But in other cases, obviously - you're on your iPad, or you're looking right at the thing and you can manipulate it in a way with your finger that might feel really pleasurable and way better than having to push buttons. So I do think it's very context-specific.
+
+Gaming to me is one of the most interesting ones, that despite all the different systems that have come out, it seems like gamers still just really want to push buttons. They want physical controls. And you can see someone who plays with a video game controller how well they know those buttons; and they're not looking down at them, they're just expert mashers.
+
+**Adam Stacoviak:** Yes, for sure. I concur on that one. I feel like the modern game controller has more buttons than obviously like the NES from back in the day, or even the Atari. It began with a joystick, and one button. I couldn't imagine gaming with a joystick and one button right now. I'd just be like "What?! How do you do that?"
+
+**Jerod Santo:** Yeah.
+
+**Adam Stacoviak:** But sure.
+
+**Jerod Santo:** Well, that goes to the geography as well. I mean, our minds are very good at mapping geographies onto memory... And so we can remember with precision the location of buttons and to differentiate them. This is why a lot of people who are into like memorization and speed reading will talk about their mind palace, and like how they map actually memory to like made-up locations in their brain. I've never done any of that, I've just seen people talk about it. But it's because we just have these geographic memories. We tie locations and directions very well to memory, which is why you can very quickly navigate a new city or a new town, after a couple of days you can remember kind of landmarks and where things are... Unless you're just staring at your phone and letting it tell you which direction to go... Which - admittedly, that's what I do nowadays.
+
+**Adam Stacoviak:** Or you're ubering...
+
+**Jerod Santo:** Yeah, or you're ubering around.
+
+**Adam Stacoviak:** But if you're liming, you know...
+
+**Jerod Santo:** True.
+
+**Adam Stacoviak:** A little easier. Does that translate to you, Rachel, liming?
+
+**Rachel Plotnick:** Oh, yeah.
+
+**Adam Stacoviak:** Okay, good.
+
+**Jerod Santo:** I love those things. There's no better way to get around a downtown of a city that is a nice one than a nice electric scooter, you know? \[laughter\]
+
+**Adam Stacoviak:** We've scooted in some shady spots, I'm sure...
+
+**Jerod Santo:** Yeah, every once in a while. But they go 15 miles an hour, or 25... You can get out of there quick.
+
+**Adam Stacoviak:** You really do learn the city, though. You're right. This idea of the mapping the geography to your brain... We learned Raleigh pretty quickly even, Jerod. In a matter of a day or two we knew where we were going because we scooted everywhere.
+
+**Jerod Santo:** Yeah. And so I think we have a very good system for memorizing specifically the locations of buttons that we need. This is why when you're driving and you can't look away, and you need to push a button, maybe to turn your hazards on, for instance, most of the time when you turn your hazards on, you're doing it quickly, because there's a hazard. You're trying to tell people that something abnormal is happening. And if you have to dial down through a touchscreen of particular menus and buttons - you know, fake buttons - in order to find it, that's just not right. Now, you can probably, if you've hit your hazards enough time, even right now, picture in your head where that hazard button is on your dash, and you can hit it without looking, at least close to it, and then you can feel and verify "Yup, there it is", and hit that sucker. I don't think you can replace that with great haptics, or sounds, or any sort of digital feedback.
+
+**Rachel Plotnick:** \[34:24\] I do think something interesting about that too, what you're describing, is this kind of muscle memory. And all of that too just does take time with a system... Because as you were talking, I was thinking of all the places where I feel like an alien, or I feel like a totally incompetent user... You get into a rental car, you've never used any of the mirrors or the dash before... Or you're a Mac user and you start using a PC, or vice versa, or you've never used a particular game system... I mean, in all those situations you can just be totally clumsy and just futzing about, when on the other hand you become this kind of expert user when you've built that muscle memory and you know exactly where it is.
+
+**Jerod Santo:** So I think cars are an obvious one where we're going to see a reversal... I mean, we've already started to see a reversal and a return to knobs and switches... We're seeing it on phones as well. I mean, Apple -- okay, maybe they've run out of ideas. That's one of my thoughts on Apple, with the iPhone, is they're out of new stuff, so they're just like "Well, let's just throw a camera button in there now."
+
+But also, now there's a camera button, which is another one where if you want to quickly take a picture, and you're going to hold your phone sideways all of a sudden, having a separate button that controls this particular important sub-function is worth having a piece of hardware there... Versus overloading a button to do a bunch of things, which they also do. Or software, software, software. And so now they're putting another button on the iPhone. And so phones, cars... Is there anywhere else where they're re-buttonifying things?
+
+**Rachel Plotnick:** That's a good question. As you were talking, it was making me think that it's almost like we had to take the buttons away to figure out where we should put them back... I know there were some experiments with elevators, of taking buttons out of elevators, and you just have this kind of central nervous system of the elevator, where you just walk up to a kiosk and press your floor, and then it would queue up the elevator for all the different people... And so you get in and you just be in this steel box with absolutely nothing inside. And you can probably guess what happened... People really started to freak out about that.
+
+**Adam Stacoviak:** Yes...
+
+**Jerod Santo:** \[laughs\]
+
+**Adam Stacoviak:** "Let me out of here...!" Yeah, even if the button doesn't work, give me a button.
+
+**Rachel Plotnick:** Yeah. Well, that's the placebo button. We all want to feel like we're in control, and there's that that quality of agency that buttons have, like "I control this machine, not the other way around." So when you take away the buttons, people get really upset.
+
+**Jerod Santo:** Oh, yeah. What about -- you said placebo buttons... That's a real thing in elevators, right? Like, the Close Door button...
+
+**Adam Stacoviak:** It does not work.
+
+**Jerod Santo:** Did you did you do research on this? Did you confirm that it's merely there for us to think we're in control?
+
+**Rachel Plotnick:** The majority of the time the elevator is controlled by a computer that has a timer, and no matter what you do, it's going to go at that particular time. So most of the time it's not going to do very much of anything. Same thing with the crosswalk button.
+
+**Adam Stacoviak:** I was going to say...
+
+**Rachel Plotnick:** Yup, same thing with the crosswalk buttons. Yup, yup.
+
+**Jerod Santo:** Oh... I hit those profusely.
+
+**Adam Stacoviak:** I refuse... They don't work anymore. I know they don't work. \[laughter\]
+
+**Rachel Plotnick:** But people get angry when you take those away, too. They just want to press that button.
+
+**Jerod Santo:** They just do.
+
+**Adam Stacoviak:** Let me feel like I've let you know I'm here... I'm here now... Give me the timer. At least start the timer, right? That might do it. It may not be like "Hey, crosswalk now." It's more like "Someone's here. Let's begin the process to allow someone to walk." Yes... Or just jaywalk. Don't do that.
+
+**Jerod Santo:** \[37:43\] So this is related. It's not a button, but it's the area of lore. How about streetlights and cars? There have been stories of specific things you can do at particular streetlights... For instance - and I've been told these things. I've never confirmed any, but I've tried them all. Like flashing your brights... There's a there's a sensor on the top of the light, if you flash your brights at the light -- if you're the only car there, for instance, and you're at a red light at a four-way stop, flash your brights. It will then know that you're coming and switch it. That's one. The other one is there's sensors. There's weight sensors in the concrete at the stoplight. And you have to make sure you're on them. And if you are -- this is usually for stoplights that have not much traffic, and they always stay green one direction... But when a car is there, they just flip for that car. Maybe you're not an expert on this, Rachel, or maybe, Adam, you know... But are those just old wives tales that they might say? Or is that legit?
+
+**Rachel Plotnick:** I wish I knew... Maybe Adam knows. But I do think that gets at this issue of like we could make most things automatic now if we want to. We don't really need to push buttons to do anything. Everything can be sensors and automatic, timer this, timer that... We could live in a world where everything's just kind of auto-controlled for us... But yeah, I don't know, Adam, maybe you've got a better answer on that.
+
+**Jerod Santo:** Adam, do you know about that?
+
+**Adam Stacoviak:** I don't know about the flashing the brights, but I do know there's a sensor in some cases; not in all cases, but I can concur on this one, because we recently got a light leaving my neighborhood for which there was not a light before; there was a Stop sign. But there was also a lot of wrecks. And so there was an occasional wreck at least once a month. Somebody had a fender bender because "Hey, let me out." "No, you can't come." You know, all that road situation. But now we have a red light, and sometimes I'll be out late, going to the store, like 11 o'clock... Which is not super-late, but it's pretty late for a small town. Like, our small town shuts down at eight. There's nothing open after eight. So if you're out after eight, it might also be midnight. It's still late, kind of thing. And so I'm sitting at this red light for a good five minutes before I realized the thing had not turned green. And I was just doing my thing as a driver... And it's a new light for us. And so I'm like "Well, maybe there's a sensor! Let me back up, because I'm past the line, because I'm anxious." Because prior to this it was a stoplight; you would pull out far enough so you can see around the turn, and confirm and drive away. And so I'm still learning how to use this intersection.
+
+So I back up, and I wait. And the red light is still red. And I see the opposing sides, I can see that the lights are doing their thing. Still green. They go to yellow. They go to red. Here I'm still red. Now we're all red. Where is my green light?
+
+**Rachel Plotnick:** Brutal.
+
+**Adam Stacoviak:** So what do I do?
+
+**Jerod Santo:** You run it.
+
+**Adam Stacoviak:** I waited 10 minutes, because I was like "Okay, I'm not going to get pulled over..."
+
+**Jerod Santo:** \[laughs\]
+
+**Rachel Plotnick:** 10 minutes?!
+
+**Jerod Santo:** You're even more compliant than most people on Earth, I think.
+
+**Adam Stacoviak:** Well, I was like "Well, maybe there's a cop nearby here, and this is a trap."
+
+**Jerod Santo:** Have you ever sat there and you're just like "I should just run it" and then you realize no one's there, and you're like "I still haven't run it. I need to run it."
+
+**Adam Stacoviak:** Yeah... I wasted 10 minutes of my life. That's the story. I wasted 10 minutes of my life thinking there was a sensor, and there was no sensor. So I don't know, is the long story short. I don't know if there's sensors or not. I know I tried to see if there was one...
+
+**Jerod Santo:** \[laughs\] I thought you were gonna confirm.
+
+**Adam Stacoviak:** ...and I was waiting, and it did not happen, and I ended up running the red light.
+
+**Jerod Santo:** Good story.
+
+**Adam Stacoviak:** Don't listen to this, cops. It's not true. I'm just -- this is just fodder for a podcast.
+
+**Jerod Santo:** What's funny is these - what I call them as wives tales, because I don't know if they're true or not, but I've used them my entire life... You know, it's kind of like pushing the crosswalk. Like, "Well, it might help." However, now I'm getting 16-year-old children and I'm teaching them the drive, and I'm wondering "Do I pass on my lore? Do I tell them these tips? Or is it like completely insane that I do this, and I shouldn't actually tell them?" So thanks for not really helping, y'all... But at least now you know some of my struggles.
+
+**Rachel Plotnick:** Well, I think because we don't know how so many machines work, and they are so impenetrable to us, we have to come up with this folklore and all this local wisdom... Because we can't really see how the machine works, or what makes it do what it does, and so it leads people to do a lot of wild and crazy things to try to figure out how to game the system.
+
+**Jerod Santo:** Ain't that right. Just trying to like come up with a unified theory of how things work. You're like "Here's how I think this is working, but it's a black box, I have no idea how it's working." Oh man, software developers understand that...
+
+**Adam Stacoviak:** \[42:06\] Have you considered as part of your research in your book to create a decision tree of when to make it a button and when to make it "Okay, this could be either/or kind of thing"? Is there any guidance you have, or expertise based upon your research that says "I can go look at this and as I'm a designer or a decision-maker in the idea of a button or no button, or how to deal with the scenario" - do you have any wisdom to share, like a decision tree, or anything that says "Yes, in this case button" or "No, in this case digital is fine"?
+
+**Rachel Plotnick:** That's a great question. I think a lot of people are interested in that and wanting some kind of standardization or some kind of system for deciding when to put these things in which situations... And I'm not a UX expert and I'm not an HCI expert, so I don't do much in the way of kind of saying "This is right or wrong." I know there are regulators that are starting to do this kind of work. I just saw in the EU that they're actually requiring cars go back to having physical buttons instead of touchscreens for things like turn signals, and windshield wipers... So I think we're going to see more legislation like that, maybe more standardization of things... Especially in safety situations.
+
+I've also talked to people about things like defibrillators, and CT scanners, and X-rays - you know, medical situations when you have to push a button... And it seems like any time life or death is involved, or people's well-being is at stake, probably a touchscreen is not going to be the right way to go, because it's going to involve a lot more machinations to do that. And I do know that studies have shown that it just takes longer to push buttons on touchscreens. Even simple things take more time than when you're reaching for a physical button. So I wish I had a list of best practices... I do think a lot of it is situation-specific.
+
+**Adam Stacoviak:** Yeah, for sure. It's probably hard to create a decision tree like that for all buttons, in all scenarios, for all things, because you're right there -- back to the Cybertruck, for example... And maybe this is the case for most Tesla cars and vehicles, I'm not really sure. In my experience, I haven't really driven any Tesla. You can confirm this, Jerod. You drove a Tesla before.
+
+**Jerod Santo:** Yeah.
+
+**Adam Stacoviak:** The shifter is in the screen.
+
+**Rachel Plotnick:** Yeah...
+
+**Adam Stacoviak:** I'm like "Okay, I've gotta --" I think I had to slide my finger to park, or to drive... It was like some sort of abnormal, not the way of every driven car I've ever driven in my life... And so I'm like "What? How is that there?" And I'm sure once you've become accustomed to it it's okay, but imagine if you're James Bond. Just imagine this. And you have to get away from the bad guys, right? And you hop into a Cybertruck and you've never driven one before. You're going to get caught, because you don't know how to push it to drive, you know? \[laughs\]
+
+**Jerod Santo:** That would be a funny little sketch.
+
+**Rachel Plotnick:** That would be really funny.
+
+**Adam Stacoviak:** Or when you take your foot off the accelerator, it's going to start braking on you. Like, that's not what's supposed to happen. Don't start braking when I accidentally just remove my foot. You have to always have your foot on the accelerator. It's so strange.
+
+**Rachel Plotnick:** There's a decision tree for you. It's just "Could James Bond do it or not?" And then, you know...
+
+**Jerod Santo:** Right. If James Bond would die, then we need physical buttons here, man.
+
+**Rachel Plotnick:** Yeah, that's it. You just figured it out.
+
+**Jerod Santo:** If he can figure it out, you're good to go. Yeah, I think medical is an interesting one... I think you're on point with -- like, if it's dangerous, if there's large consequences, you need not just physical buttons, but going back to Adam's reference of like the nuclear buttons... Underneath a piece of glass, with timing, with like the key and the turn, maybe there's three buttons you have to push... Like, the more consequences are hidden underneath that button push, or the result of the button push, there should be more ceremony, more importance, more constraints around the button.
+
+**Rachel Plotnick:** Yeah, it's almost like it has to be made more complex in certain situations and simpler in other ones. Like the defibrillator case, a lot of people -- these are automatic defibrillators that are installed in places, AEDs like in shopping malls, or in a city, so that you could administer first aid if someone's having a heart attack right in front of you. And they want average citizens to do it, but they've found that a lot of people are just too anxious to push the button, because they're so worried "Hey, what's going to happen if I electrify this person? I don't want to push the button."
+
+\[46:16\] So they were trying to figure out "How do we make this simplistic enough that people won't be afraid to actually use it?" In other cases it's like "Alright, let's put 15 layers of security around this thing so that people will be dissuaded from using it.
+
+**Adam Stacoviak:** Yeah. Simplicity is good, and complexity is good. Simplicity in the fact that you've got one button... Back to the Kodak camera. "Hey, even you can push the button." I agree with that. It's kind of cool. Like, you've got this automatic camera, all you had to do was push the button and turn the thing. There was two steps and you could take photos. Or in the case of something complex or catastrophic, multiple steps before the button push, and don't make it so simple.
+
+**Jerod Santo:** But the world in which we don't need any buttons - probably not coming anytime soon. What I find so interesting, Rachel, is you go from this topic of buttons and tactile controls to what you're working on now, which is smudges, sweat, coffee spills...
+
+**Adam Stacoviak:** What...?!
+
+**Jerod Santo:** I've got to hear this story. So what you're working on now, "License to spill: Where dry devices meet liquid lives." Where did this come from? This seems like it's out of left field.
+
+**Rachel Plotnick:** You know, I like to keep people guessing a little bit, I guess...
+
+**Jerod Santo:** \[laughs\]
+
+**Rachel Plotnick:** ...but they're actually not as far afield as they see them. I've always been interested in interfaces, and just thinking about how people interface more, you know, metaphorically with their devices. And I guess maybe this project emerged out of the pandemic and also having children, but I got really fascinated by kind of messiness and care, and how we have to negotiate our devices in all these situations... Like, what happens when you spill coffee on your laptop keyboard? What happens if you drop your phone in the toilet? What about wanting to take your smartwatch in the shower, or if you start sweating with your Fitbit? ...things like that. And I realized there's just a lot of interesting things happening around our environments and our bodies and our devices, and maintenance and care is kind of a real problem. So I looked at issues of waterproofing and how that's kind of evolved over time, everything from wristwatches to current smartphones, and thinking about what we do to kind of keep our devices alive in all of these situations. It's actually pretty tricky.
+
+**Jerod Santo:** How do you go about looking into something like that?
+
+**Rachel Plotnick:** It was really not easy, actually. Just just like buttons, it was like "Alright, well, which buttons should I study? And which liquids should I study? Which devices?" But I kind of take this long historical approach. I actually start out with looking at babies who were teething on telephone cords in the 1900s. They kept getting saliva on the cord, and then people would lose their phone service, which was pretty fascinating.
+
+I look at things like vinyl records and how smudges on the disc would create problems with it being able to play music. And then I move all the way into things like the Sony Walkman and early PCs, and keyboards and all the problems around liquids and dust, and things like that. And then moving more into the present, with wearables and smartphones, and things like that.
+
+So I kind of took this long range approach of thinking about "Hey, this kind of messiness", what I call media hygiene, how we kind of protect our devices from ourselves - it has been a problem for a really long time, but we kind of negotiated it in different ways at different moments. Even the buttons - they get wet, they get dirty... You're going to have problems.
+
+**Adam Stacoviak:** Media hygiene. Is part of your study how many germs are on my iPhone kind of thing? Or is it in that realm?
+
+**Rachel Plotnick:** \[49:39\] I looked at that a little bit, yeah. I mean, there was a lot of talk, especially during the pandemic, about what's appropriate for cleaning your phone, what are you allowed to put on it, what are you not, what happens if you take it in the bathroom... And I think that there's -- you know, obviously there's contamination issues and health issues around that... But it's also like "Where do we want our media?" Is it socially acceptable to livestream from your bathtub? Do you put a TV in your shower? Is it good to disconnect in certain situations? Where do we actually want our devices to go? And some people actually say "Hey, it's better that this thing isn't waterproof, because that way I have to take a break." So there's an interesting kind of dialogue there.
+
+**Break**: \[50:16\]
+
+**Jerod Santo:** I'm reminded of the Truman Show, where everybody's watching Truman, of course, as the story climaxes, and will he or will he not escape his little enclave... New Haven? I don't know what it was called. And everyone's watching all around the world in their different contexts. And there's the one guy who's in his tub, and he's in his tub the whole movie, watching this --
+
+**Adam Stacoviak:** Splashing, too. Big splashes.
+
+**Jerod Santo:** Yeah. And by the end he's getting excited, he's splashing, and becoming going crazy... I'm thinking "You're about to electrocute yourself, dude." So that would be a time not to use, I think, electricity, would be when we're splashing around in the tub, but...
+
+**Rachel Plotnick:** Yeah, there were a lot of terrible stories about that, especially in the '20s and '30s, people dropping their radios in the bathtub, even children... You know, horrible deaths from electrocution. And obviously now we don't have to worry about that as much as we used to. But I think there's still kind of a stigma around liquids and electronics. Like, people immediately are like "Put that drink away. Don't take that there." We still kind of tiptoe around these issues. We haven't kind of overcome those thresholds, to some degree, the taboo of it.
+
+**Adam Stacoviak:** Who is this book for? Is it for everyday people like Jerod and I, that we're just going to start taking caution or knowing when or what to expect with different devices? Like, who's the consumer or the reader, generally?
+
+**Rachel Plotnick:** \[54:01\] I'm hoping it'll be interesting for academics and laypeople. I think it speaks to a story that all of us kind of navigate on a daily basis. We often don't realize just how much we're thinking about "Should I put this in a case or not? Do I take this to the beach with me or do I leave it behind? Oh crap, I poured this entire mug of coffee on my computer. What do I do now?" So I'm hoping that the stories will kind of resonate with people who live through that, but also from an academic perspective, trying to help people understand what media hygiene is, and how it's kind of been this pervasive problem for the maintenance and care of technology.
+
+**Adam Stacoviak:** How much did you open up the topic of right to repair, considering buttons? Like, going back to buttons, if it's behind a screen, it's less easily repairable. Whereas if it's a digital analog button, let's just say... I'm sure there's some digital behind the analogness of the button... But you can at least potentially swap out that sensor to correct that button issue. How much did you push the button, I suppose, on this topic?
+
+**Jerod Santo:** Haah...! \[laughter\]
+
+**Rachel Plotnick:** Nice!
+
+**Jerod Santo:** Somebody had to do it.
+
+**Rachel Plotnick:** No, that was definitely a part of my study. I think right to repair is really important to think about. And obviously, over time our devices have gotten harder and harder to repair. So if you do have an accident, the consequences for that are much greater than they used to be. And a lot of what I focus on is how there's this disparity between advertising, like "Oh, it's so sexy and pretty. You can take your watch underwater. You can go diving with it. Don't worry, a spill is just an oops, it's no big deal." But then watching how people have navigated horrible warranty plans, and liquid contact indicators... Oh, it got tripped, therefore now your whole warranty is invalidated, and we can't repair this thing. You have to get a new one.
+
+So I think there are a lot of kind of real world consequences for this, where we as consumers get penalized for our quote-unquote bad hygiene. We're treated like bad users, and these companies continue to kind of profit off of all of our mistakes, if you will.
+
+**Adam Stacoviak:** How long was this research? How deep did you have to go? How was it funded? What was the purpose of this research? Obviously, you wrote a book, but was that the whole point of the research? What's the backstory there?
+
+**Rachel Plotnick:** Yeah, it took a few years. I started probably about 2019 working on this. I didn't know exactly what it was going to turn into at that time. Initially, I kind of started with the tactile stuff. I was looking at smudges on our smartphones, and just kind of thinking about screens and tactility and handling, since that kind of goes along with buttons; just touching stuff, right? And it kind of morphed over time into thinking about liquids.
+
+So it's been about four years that I was working on the project. And I went into some archives, looked at Rolex's archives and kind of the history of waterproofing wristwatches in the military, and kind of went down a lot of paths as I was developing this... But the hope was always that it was going to turn into this kind of bigger book project.
+
+**Jerod Santo:** So stop me if you've heard this one. This is going to sound like a joke, but it's not a joke. When we talk about computer/human interactions and hygiene - how do you eat Cheetos while you're coding? This is not a joke. This is real. I know it sounds like a joke. I learned this from Johnny Boursiquot, who's a friend of ours here on the show. He either really does this, or he convinced all of us that he really does this. How do you eat Cheetos while you're coding? Anybody want to guess? Let me say it this way, "How does Johnny Boursiquot eat Cheetos while he's coding?"
+
+**Rachel Plotnick:** One-handed?
+
+**Jerod Santo:** No. Good guess. Adam, do you want to give it a shot?
+
+**Adam Stacoviak:** Dispensed. There's some sort of dispenser.
+
+**Jerod Santo:** Very close. Chopsticks.
+
+**Rachel Plotnick:** Chopsticks... I've heard that.
+
+**Jerod Santo:** Yeah. So that's real-world dealing with two desires. Your desire to work and your desire to eat Cheetos. I guess three desires - and not to have Cheetos all over your stinking keyboard... Because that's some hygiene problems right there.
+
+**Rachel Plotnick:** Yeah, there's a good section of the book that looks at torture testing and all of the kind of bizarre trials we put our devices to see what will they stand up to. And in fact, I look at programmers and think about all night coding, and kind of the desire to have food and drinks near the device, and how people have had to navigate that... Now I'm doing something on energy drinks and gamers... So I think the mixing of food and drinks and snacks with computers is super-interesting.
+
+**Jerod Santo:** Yeah, because they don't really mix, but we want them to so bad.
+
+**Rachel Plotnick:** Yeah. Yeah.
+
+**Adam Stacoviak:** \[58:16\] Gosh...
+
+**Jerod Santo:** \[laughs\]
+
+**Rachel Plotnick:** Whether that's to our benefit, I don't know.
+
+**Jerod Santo:** Yeah, I don't know about that either.
+
+**Adam Stacoviak:** There's a note here, at least in the description of the book, that says "Obsolescence culture." What does that reference to? Like planned obsolescence, or something else?
+
+**Rachel Plotnick:** Yeah, I think -- well, the way that things get made obsolete when they break, and that it's so easy to break things now. "Oh, I spilled this tiny bit of liquid on it, therefore now the whole thing is broken. Go out and buy a new one." It's pretty painful.
+
+**Adam Stacoviak:** I'll have to share my example then, since that's how it works... So we had a microwave. I say "had" because we had to swap it out. Thankfully, we could get the exact same one. And so this issue with the microwave was not the function of the microwave. It was the digital touchpad, which is kind of like analog buttons, but they're just really like these depressible buttons. It's sort of the non-screen buttons.
+
+**Rachel Plotnick:** Yeah, those are membranes; membrane keys.
+
+**Adam Stacoviak:** Membrane keys. There you go. Well, when you push certain buttons, they wouldn't do the function that the button did anymore. And then that was okay, because I was like "Well, I don't really push nine so frequently, so I'm okay with that. Fine." I can't do 19 minutes, which I would never do in a microwave unless it was something crazy, I suppose... Or nine minutes. But okay, I can live without the nine button. But then it became of the other buttons, like zero, or one.
+
+It's like "Well, I have to have one..." You know? And then the screen that shows you -- like the liquid display, the LCD, if that's what it was. I don't think it was an LCD, but some sort of liquid crystal display... Which is LCD. \[laughter\]
+
+**Jerod Santo:** I'm just watching you figure this out as you go...
+
+**Adam Stacoviak:** I was trying to be more specific and then I messed up again. So just follow me. Just do your best, okay?
+
+**Jerod Santo:** We're with you, man. We're with you.
+
+**Rachel Plotnick:** Stick with the buttons.
+
+**Adam Stacoviak:** The buttons. The screen stopped working. So we're like "Okay, this microwave is no longer usable." And we don't microwave so frequently that it was like that big of a deal, but we had to replace it, because there were certain things, like maybe popcorn, or just certain things you would use it for.
+
+**Jerod Santo:** Sure.
+
+**Adam Stacoviak:** And then it was like "Well, we have to replace it with one that fits in there", because it was a built-in microwave, and it had a surround around it. And so thankfully, we can go and replace it. So we literally took this microwave that functioned properly, that could not be commanded properly. And it became obsolete, because no one -- I can't give it to somebody. Maybe I can share it with the electronics person who can repair it, but then they have to have enough value in the process to repair it to get the value back out from their time.
+
+So here's me, spending 200 bucks on a new microwave or whatever the number was, for the same exact thing, just to get done. It's like "Ah, fixed." And it was just it was; like, the same microwave. And we got rid of it because of that.
+
+**Rachel Plotnick:** It's that single point of failure, you know? The fact that one button could ruin the whole device, I think, is so problematic. And you're so right. I mean, most people - you take it to any repair person and they say "Oh, we don't want to fix it. It would be cheaper to just get you a new one."
+
+**Adam Stacoviak:** Frustrating.
+
+**Rachel Plotnick:** Brutal.
+
+**Adam Stacoviak:** What's with this new research thing? Can you talk more about this energy drinks and computers, and - can you tease? Where is it at? Have you put anything online yet?
+
+**Rachel Plotnick:** Not yet. Actually, I'm doing a talk on Twitch on January 31st. I guess I'll plug that.
+
+**Jerod Santo:** Nice.
+
+**Rachel Plotnick:** For ROMchip, it's a game history journal. So we're doing a talk on Twitch, and... Yeah, I'm kind of thinking, again, sort of bringing the past and present into conversation. I'm looking at things like pilots and fatigue, and different uses of caffeine and chemicals to try to keep people awake to do their jobs... And I was just really interested in this question of how energy drinks have become so pervasive in gaming, the idea of the all night gamer, and what do we kind of put our bodies through to get them to consume more media? So it's kind of like the equivalent of binge watching, but for gaming, and that's been a part of gaming culture for a long time... So I try to think about building some kind of timeline or historical narrative around how we got to be this way, and why, and... Now energy drinks are super-interesting. They've got all these nootropics and vitamins and supplements in them... They're kind of being pitched as healthy, whereas they used to be pitched as junk food, and even that's been a really fascinating kind of shift.
+
+**Jerod Santo:** \[01:02:23.29\] Yeah, I wonder how far you went into the energy drink thing, because... I guess I've lived long enough to remember when they were junky, and now it's like there's this health aspect to it... And I don't know anything about nootropics or any of this stuff. I just look at it and think "This seems like -" and because they're so huge... I mean, there's a new energy drink every day... They're mostly differentiated by branding, and not much else, it doesn't seem... And kids love them. And I'm thinking like "This seems like kind of a predatory industry", where it's like they're -- these aren't good for you, and they're addictive, and stuff, but it's like "Hey, it's fun. It's healthy. It's this." Have you looked into the makeup of these energy drinks and what goes all into them, and stuff?
+
+**Rachel Plotnick:** Yeah, I'm still kind of working through that, but I think you're right. Most public health studies have shown that there are a lot of problems with these drinks. I think some of them have tried to cut out sugar and things like that, and replaced it with other things... But the health implications are certainly concerning, and worrisome in many ways. And what's fascinating is how these industries have embedded themselves so deeply in media, whether it's promotions, or going to conventions, or in-game kind of features and things... So it's almost like the drink has become media in and of itself, and it's another thing to be consumed. They pull influencers in to promote them... So it's a massive industry that I think hasn't been studied quite enough, so I'm excited to kind of explore a little more and see where that goes.
+
+**Jerod Santo:** Some of their marketing teams are amazing, too. Red Bull's marketing, for instance -- and Red Bull's been around for a long time, so they're not one of these new ones... In fact, when I was younger, Red Bulls were all the rage, but there was no sort of illusion that it was good for you. It was just going to keep you up all -- it's going to give you wings, and you're going to crash later, but you're like "That's alright, I've got to study. I'm going to hit the Red Bulls." But man, some of the stuff Red Bull does... Just associate themselves with extreme sports, and sponsoring extreme athletes, and marketing in a way that's really, really effective, and quality... So much so that sometimes I'm jealous of some of the stuff they come up with. Just ideas... It's really cool.
+
+So like that whole side of if your entire business is predicated on you being better at marketing and branding than everybody else, and that's really where like the battleground is for consumer sales... You can produce some really good marketing campaigns, right? Like, that just brings out some amazing stuff.
+
+**Rachel Plotnick:** Yeah, and I think it's very interesting too that in some ways these drinks are becoming more akin to Gatorade, or kind of sports drinks, or they're pitched that way, at least... And I think as we see the rise of e-sports and things like that, it becomes associated with the idea of endurance, and performance, and you know, get that kind of edge... So that's really implicit in the marketing, too; or explicit, I guess.
+
+**Adam Stacoviak:** Yeah. You're right, Jerod, Red Bull has done it well. There's a big fascination, too. And you've kind of arrived if you're sponsored by Red Bull. I pay attention to mountain biking, professional mountain biking. And there's several folks I pay attention to that are in the game, and one of the YouTubers - he's a YouTuber too, I suppose... But he shared his story of how 10 years ago it was a dream that he became a Red Bull rider, that he was sponsored by them. And he has been since, and how his life has changed. It's almost like arriving; it's like major leagues, but for this sort of like -- it's still a sport, but it's this uniquely styled sport, where it's extreme sports. Like skateboarding, or mountain biking, or maybe even snowboarding... I think there's lots of different stuff around that, but that's really wild how they've been able to use that marketing to entrench themselves.
+
+\[01:06:01.06\] And now they even have Red Bull TV, where they literally are a media company, more so than just simply a drink company. They've transcended their original -- it's almost like Apple, where they kind of began as Apple computers, and now they're just Apple, because it's beyond the computer. They've transcended. And I suppose they're also into music and films and stuff, too.
+
+**Jerod Santo:** Yeah. Well, they're just like culturally as relevant -- they want to be as culturally relevant as possible, because then you associate going to a concert with Red Bull, going to go snowboarding with Red Bull. All these things just get associated implicitly with Red Bull, and it's pretty smart.
+
+**Adam Stacoviak:** What I find funny though is that these folks I see, they drink Red Bull, and I'm like "Well --" And I don't know much about Red Bull, but I'm also in that camp where I don't know much about the health benefits or lack thereof. And I'm thinking "Really? Do you really drink Red Bull that much?"
+
+**Jerod Santo:** Well, that's what's interesting about the gamer side, because on the athletic side, especially with the history of these drinks being not good for you... I remember seeing -- I don't think it was Michael Jordan in particular, but like world-class athletes, and they're trying to sell you Coca-Cola and Sprite. It's like, I know you guys don't drink that when you're trying to train your body. And it's the same thing with some of these energy drinks. But if you're just trying to stay up all night and game, there's this fast twitch... There's an allure or an air of physicality to it, but really, you're just sitting there pushing buttons. So you don't have to be that in good a shape, so you can kind of buy it more... Destiny, or whoever it is, they drink this particular energy drink. You're like "Yeah, I buy that", because they've got to stay up all night. They don't have to like run 100 meters in eight seconds.
+
+**Rachel Plotnick:** But it is still extreme. It's hardcore to stay up all night and press those buttons.
+
+**Jerod Santo:** Exactly.
+
+**Adam Stacoviak:** So hardcore.
+
+**Jerod Santo:** \[laughs\] Oh, yeah... We're bringing it back to buttons. So Rachel, have you seen the H2 commercial? I told Adam about this, the new Hummer that has the crab walk...
+
+**Rachel Plotnick:** Oh, no. Tell me about that.
+
+**Jerod Santo:** Oh, man. Go out and check out this commercial. You could use it as an advertisement for your power button book, because the entire thing is like energetic music, and... This is the Hummer that goes sideways. All four tires turn, so you can do the crab walk... And the whole thing is "Push the button." And there's like a whole song.
+
+**Rachel Plotnick:** Oh, how have I not seen that? Oh, my gosh...
+
+**Jerod Santo:** Yeah, you've got to go watch it.
+
+**Rachel Plotnick:** That's in my near future.
+
+**Adam Stacoviak:** Yes.
+
+**Jerod Santo:** It's pretty catchy. Good advertising.
+
+**Rachel Plotnick:** See, selling the button still works.
+
+**Jerod Santo:** It certainly does. And then alongside your new book, what you could do is you could have like commemorative wipes, or something... \[laughter\] Some sort of cleaner that comes out with your book.
+
+**Rachel Plotnick:** There you go.
+
+**Jerod Santo:** I don't know, if you just want free mark -- we're talking marketing.
+
+**Rachel Plotnick:** I love it.
+
+**Jerod Santo:** You know, like one of these things they get from Apple, where they just clean your screen off... Or chopsticks.
+
+**Rachel Plotnick:** I want to do demonstrations where I just dunk the book in a big vat of water, and then just watch how people react. I think that's going to really play it up.
+
+**Jerod Santo:** That could go well.
+
+**Rachel Plotnick:** Yeah.
+
+**Adam Stacoviak:** And then sign it.
+
+**Rachel Plotnick:** Yeah, exactly.
+
+**Jerod Santo:** Yeah, I love it.
+
+**Adam Stacoviak:** "Here you go. Here's your wet book."
+
+**Rachel Plotnick:** Here's your soggy book. Have a great day.
+
+**Adam Stacoviak:** \[laughs\] "I can't read it..." "I know!"
+
+**Rachel Plotnick:** That's the point.
+
+**Adam Stacoviak:** That's the issue.
+
+**Jerod Santo:** You have to buy another one. You might have to buy another copy. Is there anything we didn't talk about on these two subjects now, regarding buttons and cleanliness?
+
+**Rachel Plotnick:** I feel like we covered a ton of stuff. No, I feel good.
+
+**Adam Stacoviak:** So you mentioned a Twitch coming up for your future research... But what about a website? Where can people go to find all the details of your books, and your research, and future research?
+
+**Rachel Plotnick:** Yeah, you can go to rachelplotnick.com, or you can find me at Indiana University. Either place, you'll get to something to do with buttons, liquids... All the things.
+
+**Adam Stacoviak:** All the things. And the Twitch is when again? January what?
+
+**Rachel Plotnick:** January 31st. It's at 2 p.m. Eastern. If you just search for "Rachel Plotnick and energy drinks", it should come up and you can find the registration... It's free, but we'd love to have some people join.
+
+**Adam Stacoviak:** Who should attend that? Who's that made for at this point?
+
+**Rachel Plotnick:** Anybody. Anybody can attend. Yeah, it's meant to be kind of academic light. So I'll be talking about academic things, but meant for a general audience. So if you're interested in energy drinks, come on by.
+
+**Adam Stacoviak:** Very cool.
+
+**Jerod Santo:** Thanks, Rachel.
+
+**Rachel Plotnick:** Thank you so much. This was great.

The world of embedded systems (Interview)_transcript.txt

@@ -0,0 +1,703 @@
+**Jerod Santo:** Alright, we are here with Elecia White from Embedded.fm. Elecia, welcome to the Changelog.
+
+**Elecia White:** Hi. It's nice to be here.
+
+**Jerod Santo:** Nice to have you. See, you are on the other side of the microphone, so to speak, as a longtime podcaster. I think you've done 491 episodes, if the most recent one on your homepage is, in fact, the highest number.
+
+**Elecia White:** I think it's 492 today.
+
+**Jerod Santo:** Oh. So there must be a new one out today. What's the most recent?
+
+**Elecia White:** Octopus Army.
+
+**Jerod Santo:** Oh, I see it right there. You know what? I scrolled right past it to "Oscillators Oscillating Other Oscillators", which is a great title.
+
+**Elecia White:** Like you, we like our titles to be amusing.
+
+**Jerod Santo:** I love that.
+
+**Adam Stacoviak:** Yes.
+
+**Jerod Santo:** And you're here today because of a Changelog++ listener; shout-out to Andrew O'Brien, who pointed us in your direction, a longtime listener of yours and ours, and thought "Hey, let's get them together and have them talk." So we're thankful to Andrew for pointing us your way, and we're thankful to you for coming on the show to talk all things Embedded.
+
+**Elecia White:** That is my brand.
+
+**Jerod Santo:** That's your brand.
+
+**Elecia White:** I guess so. I never really intended it, but that's where it is...
+
+**Jerod Santo:** \[laughs\] Well, you started a podcast called Embedded, so at a certain point you decided to embrace it, at least.
+
+**Elecia White:** Actually, the podcast was originally called "Making Embedded Systems", which is the title of my book. But then everybody just kept calling it Embedded, and we're like "Fine, okay..." And then NPR came and said "We want to have Embedded, too." And we're like "No, it's ours." But actually, we just said "Can we please interview an actual NPR person? Because that would be so exciting." So that's what we did.
+
+**Jerod Santo:** Cool.
+
+**Adam Stacoviak:** What was their approach to ask you about having Embedded? What does that mean? What's the backstory?
+
+**Elecia White:** Well, okay. Picture this. You are just sitting there and doing your work, coding away, and you get an email from NPR... And it's real NPR people, and they say "We'd like to talk to you", and you're like "Oh my God, it's my big moment. Finally, engineers are going to get to be on NPR, and we're going to get to talk about how exciting this all is." And they're just like "Yeah, so we're starting a show called Embedded, and we just wanted to know if you were going to have a problem with that." And we're like "We didn't trademark anything. You've got to be kidding me." Anyway. Yes, it was a huge disappointment, and yet, they did let us interview Kelly McEvers. And she had to keep leaving, because things were happening in other parts of the world that were actually important...
+
+**Jerod Santo:** Like, during your interview with her?
+
+**Elecia White:** Yes, yes. She was in the NPR studio, and she was like "Yeah, I have to go. People are doing this."
+
+**Jerod Santo:** Oh, wow.
+
+**Adam Stacoviak:** Huh.
+
+**Elecia White:** So now we're Embedded, or Embedded.fm.
+
+**Adam Stacoviak:** So they also have Embedded.
+
+**Elecia White:** Yes.
+
+**Adam Stacoviak:** There's two Embedded podcasts out there.
+
+**Elecia White:** There are. One is about how engineering is pretty cool, and how technology is interesting, and how you can make cool devices that interact with the world... And the other is about all the depressing stuff that happens in the world.
+
+**Adam Stacoviak:** \[08:10\] Well, that's a sad dichotomy.
+
+**Elecia White:** I hope ours is the former.
+
+**Jerod Santo:** \[laughs\]
+
+**Adam Stacoviak:** Yeah, I would stop at one.
+
+**Jerod Santo:** I was gonna say, I was interested in that first one.
+
+**Adam Stacoviak:** Yeah.
+
+**Jerod Santo:** It sounded cool to me.
+
+**Adam Stacoviak:** What's interesting, I think -- what I think Embedded is, is this idea of creating... I would say more so in the last, I don't know -- maybe since the Raspberry Pi came out, this idea of "I can do things myself. I can take this small thing, that is just a single-board computer, and I can learn to program, or I can add servos, I can 3D-print now", because those are more ubiquitous, "and I can start to make my own thing in the world." And the system I put on there is not necessarily connected to the internet or the rest of the world, but it might be through Wi-Fi, or whatever. Is that kind of what you mean by Embedded? Like, this kind of world where you have standalone devices that are not really tethered via wire to that network, so to speak. How do you define Embedded?
+
+**Elecia White:** I like to say that I write software for things that aren't computers.
+
+**Adam Stacoviak:** Okay.
+
+**Elecia White:** That pushes the definition onto "What is a computer?"
+
+**Adam Stacoviak:** You've got my interest. Go.
+
+**Elecia White:** If you're working on like servers, then developing an iPhone app is Embedded. But if you've been doing 8-bit microcontrollers, anything with an RTOS doesn't really feel embedded. So it is definitely a loose definition, but the whole software for things that aren't computers actually makes you think about your microwave, and your car, and your toys, if you have electronic toys for kids... There's a lot there. And it's huge. I mean, there's all the sensors that get interconnected for science, and for industry, and there's ICU devices. I mean, I did actually work on educational toys, and that was so fun. So it's a big field.
+
+**Adam Stacoviak:** That's a good explanation of it, but I was thinking a very simple one from my recent Christmas holiday... We got my oldest son some eDrums. So he has a drum kit; as you can imagine, an eight-year-old going on nine, banging on some drums is cool. But it's also, as a parent, not always the appropriate time to bang on drums. And we want to encourage the obvious curiosity that is drumming. However, we also think that eKits are cool, because they're less loud, and they're cool because they have this controller... It's not a computer, but it is a computing device, because it has MIDI controllers, and all these different sounds for drums... So you can choose the metal kit, or the eRock kit. It's got all these different drum kits in there, it's taking the sensing input from the microphone on the eDrum into the controller, into your headphones, or your amplifier, or to your tracking device, and it's turning that into MIDI control sound. I didn't think about that being -- I guess I did, but I didn't think about it until this moment... That's not a computer, but also kind of a computer, that somebody wrote software for...
+
+**Elecia White:** Exactly.
+
+**Adam Stacoviak:** ...and maybe I could even write my own software for it, if there was a way to. Is that part of what you do as embedded, is like you start to find ways to add to the non-computer computer?
+
+**Elecia White:** Well, I mean, I do it as a career. So most of the time people bring me ideas, and I help them build their systems. I may have taken apart a few toys...
+
+**Adam Stacoviak:** For sure.
+
+**Elecia White:** \[11:41\] BB-8 was one of my favorite ones when it first came out. Figuring out how to get its IMU data and maybe control it so that it did weird things... That was fun. So I don't do a lot of hacking and reverse-engineering, except for my own amusement. But the same tools are necessary. The way that I reverse-engineer a device is about basically debugging it without the code.
+
+**Jerod Santo:** So what are some non-computers that you've programmed?
+
+**Elecia White:** Well, I have a autonomous underwater vehicle that I worked on last year, and it wanders around its area. For example, one is in the Monterey Bay, in California. And it checks in with all of the various sensors that are on the seafloor. I also have worked on some of the sensors on the seafloor, measuring when we get big rainstorms, it changes how the mud moves... And that is really hard to monitor. So they dump a couple of smart rocks down there that have to live for years and years, so battery is a big issue... And it's not like I can go down there and push buttons and say "Update that firmware."
+
+**Jerod Santo:** Smart rocks? Is that what you call them? Smart rocks?
+
+**Elecia White:** Smart rocks, yes.
+
+**Jerod Santo:** Wow. Okay. Never heard of that.
+
+**Elecia White:** And then when something happens, when there's an event that causes a mudslide, or some weird currents happen, it monitors and it reports back to the little vehicle that wanders around the bay, and picks up all the data from all the sensors.
+
+So - I mean, those aren't entirely my projects. Those are working with the Monterey Bay Aquarium Research Institute, which is awesome... And they do lots of science and ocean-based technology. Let's see... What else?
+
+**Jerod Santo:** Can't top that one, can you?
+
+**Elecia White:** Well, I worked on a gunshot location system, where you sprinkle sensors and microphones around the city, and automatically call the police when there's a gunshot.
+
+**Jerod Santo:** Oh, that's cool.
+
+**Elecia White:** That was really cool technology, and had some amazing applications. I'm kind of glad I left... I mean, I am comfortable with the level of privacy associated with it, but I am less comfortable with how the company started being run.
+
+Children's toys... That was fun. I got to take some of my toys to a kindergarten, just volunteering. And one of them that I hated - it was a karaoke, and it was the most annoying toy. It would just sing constantly. And there was a voice changer that did not make anybody sound better. And I thought it was a dumb toy. I was way more interested in this other toy that helped kids learn to write.
+
+And I would take the toy to the kindergarten, and kids would play with them, and I'd watch, and it was fun. And then one of the boys, who was kind of behind, asked if he could take the karaoke home, and I was like "Good riddance." And a week later, he had caught up to his class. He had been really behind, but because singing was the way that worked for him... Yeah, it totally blew my mind that something so different than what I could have stood, worked.
+
+I've worked on DNA scanners, twice now, actually... Once pre 1995, '98... Well, anyway, once when DNA scanners were really cool, and very hard. And more recently, when -- like, you can just make them on your desk. That progress has been just phenomenal and amazing.
+
+What else have I worked on...? I worked on inertial measurement units, which used to be the way, when I worked on them, that I would make anybody I was sitting next to in a plane just stop talking to me. "Oh yeah, I work on inertial measurement units." I could have said "I sell home insurance. Are you interested?" No, inertial measurement units was so good. People were just asleep by the end of the words.
+
+\[16:06\] But I didn't tell them what I was doing on this airplane... I was going to a racetrack, so that we could test them on high-speed vehicles, or airplanes, or any of the other places where inertial sensing becomes really interesting.
+
+This was the day before Wiimotes. And so the whole idea of accelerometers and gyroscopes as being interfaces was very different then. Now you flip over your phone, it turns on. That's what you expect. But how does it even know? I think sensors are probably my favorite area of embedded systems right now.
+
+**Jerod Santo:** You've done a lot of cool stuff. How did you get into this world?
+
+**Elecia White:** By accident. I mean, isn't that how it usually happens?
+
+**Jerod Santo:** Usually.
+
+**Elecia White:** So in college, I took applied CS courses, because I couldn't see the point of theoretical ones... And theoretical engineering, because I didn't really like any of the actual engineering. And what that gave me was Fourier and signal processing, as well as the ability to develop products. I went to HP, and Netservers, big then cloud computers, and kept getting lower and lower into the hardware, monitoring temperature... The things that we monitor now in cloud servers. We were just starting to put in temperature sensors and alerts when things were happening, instead of just the bad old days of having computers that ran, or they didn't, or they caught on fire, and you found that out by going in.
+
+And as I kept going lower and lower in the stack, eventually I wandered over to HP Labs, and that was when we made DNA scanners. And I didn't know what I was getting into. My boss was just like -- or my perspective boss was like "You'll really enjoy this. I promise." And I was just like HP Labs. That is the best location ever."
+
+And the first time I moved a motor -- because I told it to, not just because I powered on it something. But the first time it was under my control, it was magical. It was like, software can touch the world. Software can change physical things. It's not just zeros and ones. It's "Move this here. Read what's actually happening." That physicality of embedded systems is what got me into it. And that's what's kept me there. I really like -- not blinking lights, but a lot of people get into it for blinking lights. A lot of the hobby things are blinking lights. And I appreciate it. I hate blinking lights, but I appreciate their enthusiasm. But for me, it's touching buttons and having things happen.
+
+So that's how I got into embedded, is I found out you can actually write software that changes things, and not just other software.
+
+**Adam Stacoviak:** That moment of discovery is the fun part, right? When you get that "Oh man, I had no idea that's what that really meant." When you kind of get that taste. And it sparks this perpetual curiosity, I would say... Because that's -- I mean, mine is not as cool as yours. I was writing CSS, and my mom was like -- and I was doing a blog, because I was far away from home... And she was catching up with me and my life via a blog, back - gosh, 2003, I want to say; 2004. And she was like "You are good at this. You should keep doing it." And because my mom said that, I'm like "Okay, this is kind of cool. I should probably keep doing this." And that's kind of like where my curiosity in software development began, which was just like learning CSS. But that initial moment is so strong.
+
+**Elecia White:** Teaching people - it's always amazing to get to see that moment, or to see part of it. I like mentoring folks, in hopes that I can be there when it hits.
+
+**Jerod Santo:** \[20:15\] On platforms like the web, it's so easy. You can get to that moment in minutes, because you can load up their favorite web page, and show them how to open up the dev tools, and change the background color, or something; or delete a line, and the title disappears. And you see that a-ha moment very quickly. And I know in embedded systems a lot of times now it's the blinking lights, right? It's the Arduino or whatever it is... And that moment of turning a light on, that's so cool.
+
+When you mentor, when you teach, what do you do? How do you get people to that moment? What do you show them? And how long does that take? Because it has historically been considerably more steps in embedded than on platforms like the web.
+
+**Elecia White:** Yes. I mean, teaching now, there are some simulators. There's one called Wokwi, and it will simulate an Arduino. And you can program it under C, or Arduino, or MicroPython; lots of interfaces. And it takes out that pain of wiring, getting the wiring wrong, blowing it up... All the things that can happen with embedded systems. It also takes away the pain of cross compiling. But in the end, it's on the web. And I agree, if you want to see that coding can do more than just -- it's not just typing. It's doing things. It's doing cool things. Changing web pages and interacting with simulators is still really the way to go. But yeah, it is sadly a little bit of a step function with some of the embedded systems, trying to get from "I opened a box" to "I'm doing something."
+
+If anybody wants to try it today, we've said Arduino. Arduino is really great. It was the system developed for helping artists do installations. And so the whole system is very directed towards art, and making things accessible that way. They don't expect you to have 20 years of C experience. The other one is one of the MicroPython boards. That one - Python's a pretty easy language as far as languages go. It's not C. And it isn't compiled, so you can try stuff on the command line. And you get a MicroPython board, you plug it in and it acts like a thumb drive. You just take your Python file over, put it on the thumb drive, and then there are all these things, like motor.run, and suddenly it runs. And at that point, you do have to have things wired up, but there are a lot of kits for the MicroPython, and it's a good way to get started.
+
+**Adam Stacoviak:** Wow.
+
+**Elecia White:** And with MicroPython or CircuitPython, which are very similar things, you can get started right away, making bigger and bigger things. Arduino - I love Arduino. There's a lot of Arduinos, there's a lot of good things you can do with Arduino, but it's an old system, and it's really expensive for what it is. So I would recommend MicroPython for the future.
+
+**Adam Stacoviak:** I pulled up the Wokwi, and I'm a couple layers deep, while you're talking...
+
+**Elecia White:** \[laughs\]
+
+**Adam Stacoviak:** And I'm on the Wokwi for VS Code, and I'm like "Wow, this is so cool." I mean, you can do it in the web too, but a lot of people are really happy in VS Code. I'm not unhappy in VS Code. It's not my favorite, but it's certainly a quite usable code editor... And I think it's a great platform for a lot of people.
+
+\[24:09\] So the fact that you can have another tab next to your code in VS Code, and it's Wokwi, and you're kind of like programming in software what could be future hardware, and learning... That's so cool. I've personally been intimidated by, I suppose, hardware coding like this, or things that do stuff, that isn't software... Because I don't know how -- and I suppose everything's findoutable, right? The word 'findoutable'... But I just haven't found out or had the time to find out. So the burden and the hurdles to start in the way that I want to start seem so high. This Wokwi thing seems so cool. I feel like it might be a bit more accessible now.
+
+**Elecia White:** It's a great system. Uri Shaked is the developer, and he's very responsive, and they've got a Discord... It's really fun. And we have an episode about it, if anybody wants to look up that.
+
+The other -- and I feel you for "How do I get started without breaking everything?" Adafruit and Sparkfun are two companies that are for hobbyist electronics, and they both have boards you plug together. You're not going to solder these. I mean, you can, but the goal for them is you just click these things together and you don't have to worry about that "Am I providing enough power? Am I going to cause a small fire if I leave this on overnight?" So Adafruit and Sparkfun are the places you go to find these things out. They both have a really great tutorial system as well. So if you're thinking about getting into it, it's a good place to go.
+
+On the other hand, Wokwi lets you do things like have a robot army that you control from an Arduino, because you can put in 500 motors, and Wokwi doesn't care. If you do that in real life, you're talking about a lot of power, and figuring these things out. But if you do it on Wokwi, it's all just really fun.
+
+**Adam Stacoviak:** I like the idea that, one, it is software, so you can scale, just like you said. Which is a great thing. But at the same time, it's like "Well, this isn't real. It's software." But the moment I feel comfortable literally buying or spending my cash on some hardware and some wires, and a kit or whatever it might be, I can take the thing that I've been playing with, iterating on, and get, like every other software developer's doing, and I'm learning those software development skills... But now, once I've kind of gotten to the point where I've got my really good version that I want to really make in the real world, I can translate that to - it seems like at least, translate that to the real world. Now, I've written the software for it. Now all I have to do is just pair it up with real hardware, versus this software version.
+
+**Elecia White:** Yes. Yes, indeed.
+
+**Adam Stacoviak:** That's so cool. That's dope. I had no idea it existed.
+
+**Jerod Santo:** How big is the chasm between "I'm able to use a simulator in Python", or maybe it's even in C, or "I'm even able to get something onto some device and have it do stuff", and your world, of "I'm writing professional embedded software for autonomous vehicles that go underwater, and sense things"? How big is that chasm?
+
+**Elecia White:** Everybody come to embedded says "It's not that bad. You just read my book. It's all fine."
+
+**Jerod Santo:** \[laughs\] Right.
+
+**Elecia White:** \[27:50\] The real me has to admit that over-the-air updates are never going to be easy. Security is never going to be easy. Low battery... All these things where you have -- like in toys, where you have to shave off pennies, so you end up with a processor that should not be able to do this much, but you are packing in every piece of RAM, every cycle is spoken for... And that optimization... People bring me their Arduino projects and they say "It works. All the features work, most of the time." And I look at it and see all of the things that isn't going to work. Usually, there's like "I just want it to connect to the internet, so I can get the data back." It's like, okay, so you have to have a database for the data you want, the cloud data; the cloud data of interest. But you may also need to monitor the devices, to make sure that they check in. What if this is a medical device, and you need the user to be using it? So if they aren't using it, you need to know. And then what if it's had some faults? What is the failsafe? Can you actually hurt someone? What firmware version is it on? And these are all cloud problems, but these are cloud problems that have to be solved with the device. And as you scale from one or two or three devices in your house that you're playing with, that you're building, or even 10 devices in your tiny startup office, into thousands or millions... I worked at Fitbit, and this was a huge problem as they scaled - I mean, orders of magnitude. That database became bigger, and more people had to access it, and yet you don't want someone to know where President Obama is right now because he tends to wear a Fitbit watch. That's the sort of data you don't want to get out. And yet you still need to know if every device is up to date, and if their batteries are degrading more quickly, because on the new firmware...
+
+Yeah, there's a lot. There really is a lot. It's not insurmountable, and there are companies all the time that are trying to make it simpler. Do they succeed? Some of them... Some of them succeed pretty well. But it's never... Sometimes they talk about it being Lego blocks. Like, I used to have to develop a lot more stuff, but now I get my drivers from whatever repository I'm working... I get them from the chip vendor, or from the RTOS vendor. I don't have to write drivers anymore. And other pieces of code, like IMUs - most of that code comes from somewhere else. I used to have to write it. I don't anymore. It's really nice. But there's still the shimming of it together to try to get it all to work together, given the time constraints, or whatever other limitation there is to the embedded system. I guess that's another way that I have defined embedded systems, is that they are purpose-built for what they do, and so they don't have a lot of extra resources. Whether it's needing to live underwater for five years on one battery, or whether it's needing to live in your pocket overnight, on one battery charge. They don't have extra stuff. And that causes a lot of problems when you need to do extra stuff.
+
+**Jerod Santo:** \[31:49\] Well, there's a sideways application of Parkinson's law here... Which, you know, Parkinson's law is the one that the scope of a project will expand to match the time allowed for the project. Something like this. That's a bad paraphrase. And I was just teaching my kids about budgeting this morning, and I was talking about how if you're not diligent, that your lifestyle will expand to match the size of your income, right? Sometimes beyond. It's just kind of a human nature thing... And I think our software expands to match the size of the hardware that it's running on. To use every last bit. And it has allowed for many of us, web developers like myself included, to be lazy, and do things the easy way, the fast way, the cheap way... But in that embedded world, where every resource counts and it's heavily constrained, you must have to know all of the optimizations, all the little tricks, the way to manage your memory so as not to bloat, or leak, or... Do you like -- I mean you must like that... I mean it's literally a constraint.
+
+**Elecia White:** A puzzle.
+
+**Jerod Santo:** The challenge of -- yeah, it's a puzzle. And to many of us that would be a huge bummer, you know? Because it's hard. And there's things to do. But you must like puzzle solving a lot.
+
+**Elecia White:** I do, I do. I really like puzzle solving. And I'm not super-afraid of math. So if I need to add up where all of my memory is going, or be able to look at a memory map file and see hex addresses, those don't really scare me. I know that they're just addresses. They just look funny.
+
+**Jerod Santo:** Mm-hm.
+
+**Elecia White:** But I really enjoy the puzzle part. And like I said, I really enjoy the motor moving.
+
+**Jerod Santo:** \[laughs\] Yeah.
+
+**Elecia White:** And when the robot -- for a personal project, I wanted to do some higher level software, so I got a Jetson, which is NVIDIA's AI embedded system. It's a Linux system. It's a computer. I used it as a computer for like two weeks. And I wanted to get a really janky arm, a robotic arm, that could go up, and down, and grab things. And I wanted to make a typing robot. I wanted to have a camera that would see the keyboard, and then the robot would hold a little, I don't know, pin, tooth-like thing. And it would punch out "Hello", or whatever I told it to punch out. And it kind of worked. And I used some machine learning, and I did a lot of computer vision, which is an area I'm always interested in. But then, one day I'm telling it "Okay, type Hello" and it types "Help." And I'm like done. I'm like "This is the end of this personal project. I cannot make it any better. This is the best." And I forgot why I started this story, but I hope it was amusing.
+
+**Adam Stacoviak:** Well, I'm wondering maybe if it wasn't a Fiona moment. My friends who are big fans of a very awesome TV show called Silicon Valley would know what I'm talking about... Because in this episode - spoiler, for those who haven't watched it yet - she asks for help. She is a robot. She needed help. I'm wondering maybe if your robot needed help.
+
+**Jerod Santo:** \[laughs\] It might have.
+
+**Elecia White:** It certainly did. \[laughter\] It was in the way of "How much software can I throw at a problem to make the hardware that is very bad better?" It turns out you can throw a lot of software at some bad hardware, and it still is bad hardware. But it can be really fun.
+
+**Break**: \[35:48\]
+
+**Adam Stacoviak:** I'm curious about the possibilities of embedded systems. Jerod, maybe your question regarding the chasm is "If I begin, how far is that dip, so to speak?" How far is it from beginner to potentially professional? How big is this embedded world? Is it an underserved world? Is there a lot of opportunity? Is there one place to go to learn? Is it all in university, or college? How do you get to be a pro? And if you do become a pro, or someone who's professional, earning their money and career via this route, is it big? Is it expansive? Is it growing?
+
+**Elecia White:** You might've noticed that your home has more blinking lights than it used to.
+
+**Adam Stacoviak:** Yeah. I'm sad about that.
+
+**Jerod Santo:** \[laughs\]
+
+**Elecia White:** Electrical tape.
+
+**Jerod Santo:** There you go.
+
+**Elecia White:** It is growing. Some things are becoming easier, some things are becoming more straightforward... But I see more job openings than people looking for jobs. So I would say that's pretty good. How do you get there? There are a lot of paths, because like I said, what is an embedded system depends on your perspective. There are quite a few embedded Linux projects that still are embedded, because they need to be very conscious of their constraints. And yet to most software engineers, it wouldn't look that different. It might be boring, because it's trying to figure out how to provision 100 small sensors with IP addresses, but the actual algorithms and such wouldn't be strange. It would be the interactions that were a little different, having to deal with actual boxes out there. But then when you start getting into higher -- I guess it's lower tolerance; things where you have to last for five years, or have to squeeze out every microamp possible, or even things where a firmware update is very hard, and if you do it wrong, things go very bad for you. Like a firmware update at the bottom of the ocean. Those -- you don't start with those. I think that's what I need to say; there are better places to start. And there are a lot of things that software engineers know, that embedded software engineers are still picking up. We are not really that good at CI/CD. If somebody could just tell us how to use Jenkins instead of making us figure it out each time, that would be great. I mean, I use Jenkins...
+
+**Jerod Santo:** Sure.
+
+**Elecia White:** ...but it's not -- some of the software engineers I work with, Jenkins is just like obvious. Every part of it is obvious. And to me it's like, I can do what I need to do, and when you ask me to set up a new project, I just go home and I, I don't know, turn off my computer for the rest of the day.
+
+Design patterns was one of those things that I really wanted to bring to embedded systems. And I feel like I have kind of talked about it for the last 20 years, about how important it is to not keep solving the problems in new ways, but to start solving the same problems in the same ways each time, so that we don't have to rewrite all the code.
+
+\[41:59\] And things like object-oriented programming... It's not new. And a lot of embedded people really understand it. And yet we mostly program in C still. I know the gasps, the horrifying screams... Yes, we still mostly program in C. But we don't usually program in an assembly anymore. We're making progress.
+
+**Jerod Santo:** There you go.
+
+**Adam Stacoviak:** There you go, yeah.
+
+**Jerod Santo:** So what about open source? Has that permeated your crew?
+
+**Elecia White:** Yes. There's open source -- I mean, the vendors of our chips have figured out that they don't really want to keep giving us their crummy software in ways that only they can manage it... So like ST and Microchip and Nordic all have big repositories, and they have examples of working with different sensors, they have their drivers there... And then we have things like Zephyr, which is an operating system that is kind of like you took Linux and you made it much smaller, and then instead of anything you could run at runtime, you hard-coded it at compile time. Which has some interesting things... The device tree is one of those things that most people in their first three months of Zephyr experience will start cursing. And once you have -- yeah, so Zephyr is huge, and FreeRTOS is now open source. Wokwi is open source; most of it, I think. There's a little bit that isn't.
+
+The sensor vendors, the IMU vendors will give us code. And sometimes that's partially open source, and sometimes it is not. But the easy stuff now, the common filters that used tot be so complicated, now that's code they just give you. They tell you how to do.
+
+**Jerod Santo:** What's that?
+
+**Elecia White:** A Kalman filter is a way of -- I could talk about this for like three hours, so I'm trying not to... \[laughter\] Okay, accelerometers and gyroscopes. Accelerometers measure gravity. They measure acceleration too, that's what they say, but really, they just measure gravity. Gyroscopes measure spinniness; how fast you're turning. And it doesn't really matter whether you're spinning around a certain thing or whether you're spinning around a part of a thing. So you can measure how your arm moves with a gyroscope. It's not like it has to go around and around. It's just how much is it turning.
+
+When you are in a car and you're going around a cloverleaf, you're spinning and you feel acceleration. You feel acceleration, the outward push of you as you're spinning around. An inertial measurement unit can use both of these sensors, and know which one is wrong. And it's kind of like your ears and your eyes. If you are spinning around in a circle and your eyes are closed, you're going to get dizzy much faster than if your eyes are open. The Kalman filter is a way to combine sensors with different error features, different error characteristics, into a way to get a much better response, because you can characterize the error when nothing's happening. Or you can characterize the error in certain conditions where you know one is good and one is bad, like on a cloverleaf.
+
+**Jerod Santo:** Gotcha. And so the vendors of the device would provide a set of filters for you to use?
+
+**Elecia White:** \[45:58\] They would provide the algorithm, yeah. And they would provide the base code for how good each of their sensors are.
+
+**Jerod Santo:** Okay. Because they've done those tests and they understand the sensors and their error rates, and these other things.
+
+**Elecia White:** Exactly.
+
+**Jerod Santo:** Better than you do.
+
+**Elecia White:** Better than I want to.
+
+**Jerod Santo:** \[laughs\] Yeah. I mean, after a while you may understand them better than they do. But those used to be closed source, or they used to be unavailable to you, and now they're just provided.
+
+**Elecia White:** In 1990, it was a patent I worked on.
+
+**Jerod Santo:** So you helped invent these things.
+
+**Elecia White:** Well, no, I helped implement them.
+
+**Jerod Santo:** Okay.
+
+**Elecia White:** It came out of a -- and to be fair, a lot of people started doing IMUs and Kalman filters in the early to mid '90s. And it exploded as to the applications... But a bunch of PhDs were done at that time. And so they became very, very popular because they solved the inertial measurement unit problem pretty well. And now they solve a lot of other problems that involve a bunch of sensors.
+
+Usually, when we talk about self-driving cars, the GPS has one set of errors, mostly; it's really slow. The LiDAR or camera can have another set of errors, and it provides a different set of information, and the different types of Kalmans can put all of these together. And the idea is just you put the errors here and you put the data here, and you try to make the best out of the bad situation you have.
+
+**Jerod Santo:** Gotcha. So you mentioned self-driving cars and you mentioned LiDAR... So I have to ask your take as an embedded person on Tesla's proposition that they can get to autonomous, or whatever level you need to get to, with cameras only. And then Waymos using LiDARs, other people using LiDARs. How important is the LiDAR, do you think? Pure opinion. Versus having a bunch of cameras and letting that be your only input, or your only data streams.
+
+**Elecia White:** I can't picture doing it only with cameras.
+
+**Jerod Santo:** That's what a lot of people say.
+
+**Elecia White:** I mean, the argument is humans do it only with cameras.
+
+**Jerod Santo:** Right. Like, all we have is eyes, right?
+
+**Elecia White:** But we have a few other things.
+
+**Jerod Santo:** Such as...?
+
+**Elecia White:** Our eyes and our inertial measurement units are very well connected.
+
+**Jerod Santo:** Okay.
+
+**Elecia White:** So there is no disconnect that if you jerk the wheel, you know what your vision's gonna do.
+
+**Jerod Santo:** Okay, I'm tracking you.
+
+**Elecia White:** And as a human, I just -- cities and self-driving cars... I worked on autonomous vehicles, and not just underwater, but self-driving, large off-road vehicles. And we did it almost entirely with cameras. But part of our goal for the future was dealing with Stop signs and pedestrians. And I spent a lot of time with the machine learning systems and the reinforcement learning systems... And it's a hard problem. And I don't know how they're going to solve city driving in a way that is truly safe and doesn't have systemic failures.
+
+That's one of the problems with medical devices. When we talk about doing embedded systems in medical devices, it's usually there's some element of "Oh my goodness, there's so much documentation. There's so much documentation the FDA needs." And that's primarily to make us understand the ramifications of the points we're making. And the reason for that is because -- oh, I've started this in the wrong place. Okay, moving backwards. If you work on airplane stuff, there are different levels of concern. If you take down a small airplane, a Cessna, it may crash into something important, like a school. And so you have a certain amount of damage you can do with a small plane.
+
+**Jerod Santo:** \[50:23\] Sure.
+
+**Elecia White:** You have a much bigger amount of damage you can do with a larger plane. Medical devices are a little confusing, because -- and safety critical devices, medical or vehicles, have a problem that you can have thousands of them in the field when they start to fail. And so your problem is not that you have hurt one person. Your problem is that your software has decided that the 29th of February is not a real day, and has gone nuts, and is now on its murdering spree. And that's what you need to avoid.
+
+And so the thing with cars is that we keep trying to make them better in small ways, but I feel like we've kind of reached a local minima, that we're not gonna get better with just cameras. I'm not even sure we're gonna get better with just a LiDAR. But I don't know how to build in that feeling of where the car is, the proprioception that you get as an experienced driver. I'm afraid that cars will drive like 16-year-olds forever, and I don't know that that's what we want.
+
+**Jerod Santo:** Yeah.
+
+**Adam Stacoviak:** As an outsider who - I'm not sure I would ever even venture into the world of what it would take to be on a team that creates autonomous cars... I'm a fan from afar, let's just say, of the idea.
+
+**Elecia White:** Me too.
+
+**Adam Stacoviak:** I like the idea, but I think it's -- this is lame, and this is from the outside. It seems perpetually flawed, because it's not a closed system. There's just so many flaws in the system that you can almost never predict. You can predict a lot of the various ones, but you can't predict all of them, ever. And because it's not a closed system that you control, then it seems like you're perpetually gonna be flawed. No matter what, there's always gonna be something that changes the game, potentially catastrophically, that you can't circumvent.
+
+**Elecia White:** As a human, you can't do either.
+
+**Adam Stacoviak:** Right.
+
+**Jerod Santo:** Yeah, I think your point about the horizontal scale, Elecia, if that's just the way I'll say it, is that we have crazily bad 16-year-old drivers, but it's not like at a specific moment in time this percentage of all cars on the road are gonna turn into that.
+
+The error that happens spread across a fleet of cars all running the same software, for instance, that makes the same mistake all at once is catastrophic. Of course, the famous or infamous -- I think this was the first Tesla that killed somebody... That was in FSD mode. It was because it came over a hill and the sun was coming up, and the sun shone into all the cameras at once. I'm just giving my depiction of what happened... It's roughly that. And it couldn't handle it. Well, of course, a human couldn't handle that either, because your eyes are completely blinded. And so likely, maybe the same result in that one instance. But if something like that happens across 50,000 cars at the same time, that is a catastrophe. Of course, it's a tragedy when one person dies. But when this fleet suddenly goes bad... I mean, the stakes are way higher, right?
+
+**Adam Stacoviak:** Yeah. I mean, I've been that human, Jerod, that's coming up over hill with the sun.
+
+**Jerod Santo:** And you're 16?
+
+**Adam Stacoviak:** \[53:57\] And I can emp -- well, not even at any age, really. I mean, period. As even a good driver, I would say, a marginally good driver. I think I've had -- the most recent accident I had, which was like the one I didn't cause... The deer hit me. And this was like a year-ish ago.
+
+**Jerod Santo:** Well, some things can't be avoided.
+
+**Adam Stacoviak:** I mean, this deer wasn't like random at nighttime, when you would think a deer would be out. It was nine in the morning, in town. Like, where other cars were at. It wasn't a time when deers should be running necessarily.
+
+**Jerod Santo:** He was looking for your car.
+
+**Adam Stacoviak:** Ran right into me. And then the other time was - geez, probably 20 years ago. And it was just wet. And I just didn't stop in time. Like, I stopped, but my car decided to skid, because that's what happens.
+
+**Jerod Santo:** It slid a little bit.
+
+**Adam Stacoviak:** Yeah. And sure, I could have been paying a little better attention... But I'm a pretty good driver, is the point of the last 45 seconds.
+
+**Jerod Santo:** That's the point you're trying to get across?
+
+**Adam Stacoviak:** I'm a pretty decent driver.
+
+**Jerod Santo:** Right.
+
+**Adam Stacoviak:** But I've been in a scenario where the sun has completely overtaken my vision, completely... I do recall, seconds beforehand, what was in front of me... And so I went now from literal memory, with confidence of my next driving maneuvers straight, and around the bend a little bit... Because I was going from my memory before my sight was taken away. In the case of the cameras, the thing was freaking out... I don't know what happened there. Maybe it was like a literal panic mode. But in my own body, I had a panic version. But I just didn't crash. There was no Adam BSD, or BSOD. There was just continued driving. I was safe.
+
+**Jerod Santo:** One of the reasons why I've been tentatively positive about this progress is because of how many accidents there are, and because how distracted and bad we are at driving, even though we are amazing, Elecia, to your point about the actual physicality of observing the world that we're in, and that connection between our hands and our eyes being just grounded in, I don't know, biology. We're pretty amazing computers. But I don't think these cars have to get so good to be better than us, because we have so many failure rates already. The failure rate's very high. That being said, it does seem fraught. It's always been 18 months around the corner for the last 10 years. There's progress being made. I mean, Waymo is opening in new cities; it seems like maybe a city-by-city rollout. And I believe they have LiDAR involved as well. It's working. They also -- I think they have a call center, I believe... And I could be spreading fake news, but I think they have actual humans, not in the loop, but available in order to take over a car if need be. That seems like a good fail safe. There's progress being made. But it does seem -- I don't want to call it an impossibility, but at this point I just don't feel like we're close.
+
+**Adam Stacoviak:** Can we go back to my idea of this closed system? Like, is that even a possibility? To your point, Jerod, can we create a closed system?
+
+**Elecia White:** Freeways and trucks.
+
+**Jerod Santo:** Yeah.
+
+**Elecia White:** Self-driving on freeways and trucks, I think, is a better possibility. Freeways have a much smaller number of things that can go wrong. And it is more of -- it's not a closed system, but it is a differently constrained system.
+
+**Jerod Santo:** Right. And a smaller problem set.
+
+**Elecia White:** I mean, if they had their own roads, it would be much simpler. Then it would be truly closed. And in the meantime, I think we can get closer on freeways more than anything. I mean, there have been a lot of -- most of the new cars now will have lane keeping, and keeping up with the person in front of you... All of the things that the initial full self-driving Tesla did, we can do now in my new Bolt. It is progressing, whether or not we call it self-drive, or if we just call it new safety features.
+
+**Jerod Santo:** \[58:15\] Yeah.
+
+**Elecia White:** I think it's getting there. And I like the idea of a closed system, but at some point you do go back to being a train.
+
+**Jerod Santo:** And so you're laying track.
+
+**Elecia White:** Or laying pavement that can only be used by this form of self-driving.
+
+**Adam Stacoviak:** Right. Some sort of guardrail, parameters, boundaries, something that says "This is where you should be. And if you're not here, then you can no longer operate", or something changes to no longer be the danger possibility, so to speak.
+
+**Elecia White:** We'll put sensors in the roads. We'll have it monitor the car in front of you. And humidity conditions. And then the sensors in the roads will talk to the cars. "Wait, wait, let me scribble this down..." \[laughter\]
+
+**Jerod Santo:** I need to invest in a smart rock company, because we're gonna be selling a lot of smart rocks.
+
+**Adam Stacoviak:** Yeah. Well, as we're knee-deep in this conversation, I was just thinking, one of the most probably globally known, most people have bought one at least at one point in their life, or want one, an embeddable system - is like a GoPro, for example. That dude who founded that company was just surfing, and was like "Man, I would love to film this." And he went away and did his thing, and now GoPro exists. And I'm just thinking like "Wow, this must be a proving ground for not just innovation, but potential future founders." This seems like a good place to go and not so much get rich, but like change the world in dramatic ways. I think GoPro has changed the world in dramatic ways. It's given you, let's just say, a camera view, obviously, of something dangerous. But then you have that GoPro effect, like "Oh, that doesn't look that steep", until you get there and you're like "That's pretty steep." This embeddable system world seems like if you want to go and invent something, this is a place to go and invent something. Because it's real. It's physical.
+
+**Elecia White:** Yes. There are lots of times when it's like I have a problem, and I know how to fix it because I know how to wire things together. But I also have that same feeling in software, that sometimes I can invent something just pure software. I do origami, and I write Python scripts. I write very long and complicated NumPy scripts for doing origami. And I love that.
+
+**Adam Stacoviak:** That's crazy.
+
+**Jerod Santo:** What do the scripts do for you?
+
+**Elecia White:** They create crease patterns, which I can then use a laser cutter or Cricut or other device to pre-score paper. Then I can fold along the wavy lines, which will give me origami that is not flat, that is patterned, and weird and wavy, in organic methods.
+
+**Adam Stacoviak:** Does this exist in the world? Can you do that? Can you go buy that product right now?
+
+**Elecia White:** No.
+
+**Adam Stacoviak:** Well, then you just invented something.
+
+**Elecia White:** I mean, there's curved origami out there, yes. Yes, there are other people who do curved origami.
+
+**Jerod Santo:** This is just a world that I have not been introduced to, so I don't know it exists.
+
+**Adam Stacoviak:** Yeah.
+
+**Jerod Santo:** But yours is custom curved origami, that you want to make. And so your scripts give you the crease lines.
+
+**Elecia White:** Yes. The scripts give me the crease lines, and they let me combine straights and creases in order to get tucks and weaves. And - I will go off the rails here... \[laughs\]
+
+**Adam Stacoviak:** Go off the rails! You're allowed.
+
+**Jerod Santo:** We're with you. We are with you. I'm listening.
+
+**Elecia White:** \[01:01:58.23\] Adam was saying how you can invent things.
+
+**Adam Stacoviak:** Yeah.
+
+**Elecia White:** And I think you can invent things in software, too. And it's some of the same impetus, that you invent them because you want them. I had a stuffed animal that would Wi-Fi through the cloud to another stuffed animal, and when you padded it, the other stuffed animal would light up. And you could have this sort of communication to someone else.
+
+**Adam Stacoviak:** What did the other animal do?
+
+**Elecia White:** I would pat mine, and theirs would light up. They would pat theirs, and mine would light up.
+
+**Adam Stacoviak:** Ah, I see. So there's like a --
+
+**Elecia White:** It was like a goodnight sort of thing.
+
+**Adam Stacoviak:** Yeah. Maybe for like brothers. They're in different rooms, like "Goodnight, brother" kind of thing. Not a walkie-talkie, it's more like "I'm petting mine, you're petting yours."
+
+**Jerod Santo:** Right. You just know that somebody else is there, doing something. It's a connection.
+
+**Adam Stacoviak:** Because it's a two-person system. Yeah, it's a peer-to-peer, with two peers.
+
+**Elecia White:** Well, this actually went through the cloud, so it was more like you're in college and you're telling your parents you're going to bed.
+
+**Adam Stacoviak:** Yeah. That's cool.
+
+**Elecia White:** But yeah. All kinds of just... Yeah. Two people connecting quietly.
+
+**Adam Stacoviak:** I think I'm liking this idea more now that we're talking about. I mean, I've always liked this idea, but... This idea of software paired with hardware that creates these newer experiences that isn't software-only. Like an iPhone or a computer screen or a browser software, which is like - maybe that's the kind of software I'm thinking of. But then I'm also thinking not just like GoPro, but I use -- so if you didn't know this, Elecia, I like to barbecue. And so I own a Traeger. A Traeger is a smoker, but it's basically a computer. It's an embedded system. Somebody wrote software to have sensors that says "How many pellets should I push through to maintain this temperature? What temperature is it in the ambient? What temperature is it at the burner?" And this thing that I just push the button, turn the dial, set it to 275 and wait an hour or two - that's a computer, an embedded system that somebody invented. Because they were like "Well, doing it with the old wood method was too cumbersome." It wasn't ubiquitous enough, because too many people didn't want to tend to fire. And so now Traegers exist as a smoker to give way more people access to push a button, oven-like control of a barbecue system.
+
+**Elecia White:** Exactly.
+
+**Adam Stacoviak:** That's crazy. I never really thought about this Traeger as this embedded system until this conversation. Obviously it is, but I'd never thought about the fact that I could have made that.
+
+**Elecia White:** You could have made it.
+
+**Adam Stacoviak:** If the things were more accessible. And I think a lot of this conversation is like how accessible is this technology to people, to lay folks like Jerod and I, or - we're seasoned in software development, but not in embedded systems necessarily.
+
+**Elecia White:** But with an Arduino or one of these little micro Python, circuit Python boards, think about what you would need to do. You would need to -- okay, so mentally build the system architecture... You need to push pallets, you need to be able to tell time, and you need to monitor temperature. If you went online and looked up "How do I get a thermistor? How do I tell temperature?" Really easy. Like, that would take you three minutes. Wiring it is going to be easy, powering, it is easy... Doing the timing - a little bit of research on that will probably turn up that, you know, you don't care whether you're doing two hours in one minute, or one hour, or 58 minutes. So you don't have a big constraint between really good time, versus kind of crummy time, either way, as long as you're around two hours.
+
+\[01:05:52.19\] And now you have to figure out how to push the pellets in in a way that's probably metal, because you don't want it to burn in plastic. You're not going to 3D-print this. And now you need a motor subsystem. And that's where things get a little complicated, but motors are very common, and you can buy the kit that just says "I want you to move here, and I want you to move there." And so you push the pellets in... Maybe you have another door closer thing. Now you're getting into the mechanical. And I do get intimidated by mechanical, but you have a unit that works. You can start to look at it and see "How does this work? What are the mechanical pieces that I need to be able to push, and close doors, and scoop pellets?" And it becomes -- you want to build this, so you build each piece, and then they all work together, and then it's done. And then you want to do firmware update, and you call me. \[laughter\]
+
+**Adam Stacoviak:** Yes.
+
+**Jerod Santo:** Love it. It seems like the -- you said you get intimidated by mechanical, and it seems like the pusher would be the hardest part of that. Maybe you could utilize gravity to just have an opener and shutter, versus a pusher, and then you're dropping the pellets in. I'm not sure how Traegers work, but...
+
+**Adam Stacoviak:** Well, I know exactly how it works. I'll tell you. So there's a hopper. And the hopper, you fill it up. It uses gravity to --
+
+**Jerod Santo:** So it just drops.
+
+**Adam Stacoviak:** Yeah. As the pellets fall down, it falls into this space that there's an auger, and the auger just turns. And so it slowly just moves the pellets along the system. And so you fine-tune the temperature within the firebox based on the speed of the auger. So if you slow down the auger, no pellets feed it. If you speed it up, it speeds it faster. And you probably have some - maybe an algorithm or machine learning of like how the speed of the auger versus temperature control... And you're obviously maintaining temperature, but it's just literally a turning mechanism. It's not even smart, aside from knowing speed and temperature, and how speed affects temperature.
+
+And the hopper is just the pellets are falling based upon gravity. So gravity did the work there. Thank you. No pusher necessary. And there is a sensor inside the hopper that will tell me if it's low, but I'm smart enough to know how much I put in there. I know when it's low. And my cooks are never that long where I need to literally monitor it, ever. I've never had to know if it's low. I think it's just a gimmick, really.
+
+And in terms of time, I'm the timer. I mean, Traeger does have an iPhone app and an Android app that controls it, and you can set time, and stuff like that, and you can tell it to turn off... But that's a rich person's feature. Not everybody needs that. You really just need the ability to maintain a temperature for a sustained amount of time. Any barbecue person will tell you "All I'm teaching you is time and temperature. I'm not teaching you anything special about cooking food. Just time, how much time do you need, and at what temperature." So that's the recipe for cooking, is like you know time and temperature. And so the Traeger just basically does those two things. It's time and temperature.
+
+**Elecia White:** And that mechanical is so much smarter than what I was coming up with.
+
+**Break**: \[01:09:05.26\]
+
+**Elecia White:** So Adam, what do you want to build next?
+
+**Adam Stacoviak:** Geez, let's see... Birthday cards.
+
+**Elecia White:** Birthday cards...!
+
+**Jerod Santo:** Say more.
+
+**Adam Stacoviak:** Yes... My family is a fan of singing birthday cards. And my son loves to take them apart; later I'll tell you a story. So my son is curious like this. He loves to take things apart and figure out how they work. And so he took apart a birthday card we got him. And obviously, you open it up and it starts to play the song, or whatever it's supposed to do. Well, he decided to take it out of the back of the package, discover the sensor, and discover this thing and this mechanism, and he decided to pull the wires out and put them in different places. He made this birthday card do things it wasn't intended to do for us. It was just meant to be open, do the thing, and it's a closed system in that regard. But he found out you can pull the sensor out, or whatever the mechanism is, and rewire it... And he made it do something different. He's like "Dad, look at this." I'm like "What is that?" He's like "It was my birthday card." That's kind of cool.
+
+**Elecia White:** Okay. Chibitronics. C-H-I-B-I-tronics.
+
+**Jerod Santo:** Okay.
+
+**Elecia White:** This is away from the software, although they have little microcontrollers that it works with. So we can get to the to the music and whatnot. But what they've done is they sell these kits that have a copper tape, and you copper-tape around and you can make little buttons by having pieces of paper touch each other. You can make little sensors by having patterns of copper tape do different things. They sell stickers that are LEDs. So you put your copper tape up to one side, your copper tape to the other, and then you put it to power and you have a little battery. Usually the clips - what are they called? What would you call a clip like this...? Binder clip.
+
+**Adam Stacoviak:** Oh, yeah.
+
+**Jerod Santo:** Binder clip. Yup.
+
+**Elecia White:** You put a binder clip with the battery, and you make your own card like that. And it lights up, or you can have it do music... They have all of these different neat sensors. This is more on the electronic side. This is how people understand how do you go from battery through a circuit. And you don't need to know Ohm's law, and all of that. Kirchhoff's laws... That's not that important for this. It's really about you follow the copper tape to the different pieces, and you make neat lit things. And you can add a controller, which can then even add a little tiny motor, which can either vibrate, or it can move... Like, you do a paper engineering cut out of a butterfly, and it will gently wag the wings... Yes. Chibitronics. They're super-fun.
+
+**Adam Stacoviak:** Chibitronics.
+
+**Elecia White:** They were directed towards women at one point, getting girls interested in storytelling through electronics... But I don't think they are as gender-specific anymore, because I've been seeing lots of their new sensors and new systems being a little louder.
+
+**Adam Stacoviak:** Okay. Since we're here then, I have to ask you... I'll just say one brand, and I want a reaction from you, Elecia. Crunch Labs.
+
+**Elecia White:** Nothing. I'm terrible.
+
+**Adam Stacoviak:** Oh, my gosh...
+
+**Jerod Santo:** Nothing... \[laughs\]
+
+**Adam Stacoviak:** Well then...
+
+**Jerod Santo:** Ah, that's great.
+
+**Adam Stacoviak:** So I don't know his name. I'm going to google real fast... Crunch Labs. The fellow that runs -- this is a very massive YouTube channel, but it's not a YouTuber, like a... I don't even know how to describe it...
+
+**Elecia White:** Oh, I've seen this guy.
+
+**Adam Stacoviak:** Oh, yeah. This is the coolest stuff ever. This guy used to work at NASA, and he is all in STEM, and all kids-focused, but he has his subscription box that you can get to like build different stuff. And as you're describing this - how do I pronounce this? Chibitronics... I'm thinking like "Wow, here's more kits that entertain and educate these --"
+
+**Jerod Santo:** Oh, is this Mark Rober.
+
+**Adam Stacoviak:** Yeah, Mark Rober. Thank you, Jerod.
+
+**Jerod Santo:** Yeah, everybody loves this guy.
+
+**Adam Stacoviak:** Yeah, I love Mark Rober. He's the coolest. I mean, he is -- I don't even know how he does what he does. He's amazing. He's living his purpose, for sure. But my son loves Crunch Labs, is so entertained about this stuff, and has become smarter because of Mark Rober. I mean, educational, curiosity, STEM, embedded systems, I'm sure, electronics... Rube Goldberg things etc.
+
+**Jerod Santo:** Sure.
+
+**Adam Stacoviak:** Just all this cool stuff. So I thought you might know Crunch Labs.
+
+**Elecia White:** I don't, but I'm gonna sign myself up. I love the little robots. And they had a crossbow... How do I get the crossbow? \[laughter\]
+
+**Adam Stacoviak:** \[01:16:03.18\] "How do I get the crossbow?" That's great. And these kits - they come on this, frequently. So I was thinking -- what made me think about Crunch Labs was the monetary accessibility of the Chibitronics... And you know, building these things that no one else seems to be doing, to -- I mean, at some point you've got to want more people in your field, right? You're probably lonely, to some degree, in your field. You want more folks who have the same drive and curiosity as you do... And the only way you get there is by inviting them in.
+
+**Elecia White:** Yes. And my field was so lucky to have Arduino, because it really drew a lot of folks to be interested, both in electronics and embedded systems, but also in software. And there are the Chibitronics, and some of these other robotics kits... First Robotics is just wonderful. That's the competition for middle and high school students. And what the kids learn is just -- I just want to hire all of the First Robotics engineers. Because once they get into the robotics and they start understanding how the mechanical and electrical and software all work together to make the robot move, they get a systems understanding that's really hard to teach.
+
+**Jerod Santo:** We've gotten so many amazing links from this conversation. Are there others, if we're like "Elecia, what's another cool thing?" You just showed us Chibitronics first... This firstinspires.org is the First Robotics people. Adam pointed us to Crunch Labs... Of course, we have MicroPython, Adafruit, the Wokwi... What else have you got?
+
+**Elecia White:** Hackster.io, Instructables.com, and Hackaday.io.
+
+**Jerod Santo:** Oh, Hackaday I'm well aware of. What was the first one you said?
+
+**Elecia White:** Hackster.io...
+
+**Jerod Santo:** Hackster.
+
+**Elecia White:** ...and Instructables.com. And I'm grouping these all together. Even though they have very different flavors, they are all places you go and tell people about your projects.
+
+**Jerod Santo:** Nice.
+
+**Elecia White:** As an engineer, it's like, okay, I'm learning to document. This is why I tell people to use these sites, because I want them to learn how to document and reproduce their projects. But as somebody who's doing it, as somebody who's excited about it, it's getting to show off what you did.
+
+We don't always get to go to these hackathons, or to these hack meetups where you get to bring your project... But this lets you connect to people who are maybe in the same boat. Maybe there's a competition for a Nordic dev kit, and so everybody's looking at the same set of sensors... And you get to see what other people's ideas are. But you also -- you know, they all have community aspects, where you can talk about what you're doing and talk about what you want to be doing... As well as demonstrations where other people have said "This is what I did, and these are the steps so you can reproduce it", in case they have a new project.
+
+**Jerod Santo:** Love it.
+
+**Adam Stacoviak:** Where's this going? This kind of thing. Showing off is kind of cool, right? I'm looking at this one here on Instructables, in the Circus section... Because it lets you see things like crafting, and cooking, and living... So there's more than just simply this embedded world there. It's a sleep cycle lamp, and I'm seeing this lamp with a common shade, probably a similar -- or sorry, a very small board beneath it that's powering this thing, and probably basing it on time it's doing different sleep cycles, colors, tones etc. But I'm thinking like "Man, where could this world go, where you have infinitely accessible Lego-like things?" Not so much Lego itself, but how you can take this part, pair with that part, and make a product, or create something new, or invent something new. Where can this go? Where's the future going for this?
+
+**Elecia White:** \[01:20:20.10\] Like you have this idea and you want to market it?
+
+**Adam Stacoviak:** No, no, no. More like you. Where do you think -- I mean, given your knowledge of embedded systems and where you've seen it go, what's next for, I suppose, this world of... From the simple hobby project to underwater things with sensors, looking at dirt moving around... You know what I mean? It's such a chasm, but like, where is this going? It's a big question, I'm sure.
+
+**Elecia White:** I mean, it's going everywhere. I fought against the term internet of things for so long, but that's what it is. Like, everything wants to be on the internet, wants to help you be better, or do something different. And I don't love all that. I mean, I don't need everything I have to be on the internet. But having two stuffed animals essentially on the internet, talking to each other over distance - it turns out to be kind of cool. The applications turn out to be really interesting. The ability of my doctor to see if I'm using some medical device they've prescribed... Okay, that gives them more tools. If it's not okay with me, then I make it not okay. But it gives other people the opportunity to engage with us if we want to, or to monitor elderly folks. Or to teach kids to read, or... Adam, I have to say, it's going everywhere.
+
+**Adam Stacoviak:** Yeah.
+
+**Elecia White:** And I'm not sure that's the best thing, but it's also going into all kinds of science. One of the projects that I worked on a year ago, two years ago, that's going out this year, is they're just gonna chuck these boxes into various rivers around the country, and be able to monitor the environmental DNA. So they would know if an otter moved into this river before anybody could see if they were actually there yet. And they can also monitor things going wrong... Like, human DNA being in the water too much is bad. The Internet of Things... That's where we're headed. And it may not be great, but some things will be great.
+
+**Adam Stacoviak:** So embrace it, right? You've lamented your fight against IOT, but... Just embrace it, basically?
+
+**Elecia White:** Be aware of your security. Be smart. You don't let strangers into your house just willy-nilly.
+
+**Adam Stacoviak:** Yeah.
+
+**Elecia White:** Maybe look at the names of the companies you're letting into your house, and make sure that they have nice friends, and good references, and all of that... Because you don't want to create bot networks or allow people to spy on you. So when you do engage with the Internet of Things, do so wisely. And it's not just people in their houses, it's companies, and UAVs. You need to know where you get your UAV from, because some of them report back to their home countries. And that's kind of scary, depending on what you're doing. If you're using it to take pictures of the beach, who cares? If you're using it to map your new mining prospect, I don't know if you care.
+
+**Adam Stacoviak:** \[01:24:12.17\] Yeah.
+
+**Elecia White:** But you should think about it. Security is unfortunately the downside of IoT. It's also the S in IoT, which is to say "You really need to think about it for yourself."
+
+**Jerod Santo:** And privacy too, which those two are, you know, kind of eternally linked, aren't they now? Because being able to throw a smart rock into a river and monitor things is awesome. But being able to throw it into my pool and monitor my swimming, maybe not so much. Just for instance, you know?
+
+**Elecia White:** Yes, there are a lot of downsides to this.
+
+**Jerod Santo:** The internet of things... And also, for IoT makers out there - just leave us physical buttons. Let us opt out every once in a while and just use the dishwasher without having to launch an app, for instance...
+
+**Elecia White:** Oh, yes...
+
+**Jerod Santo:** \[laughs\]
+
+**Adam Stacoviak:** For instance, yes. I take it your dishwasher requires you to use an application or something, Jerod?
+
+**Jerod Santo:** No, that was just a for instance. I'm sure somebody's does.
+
+**Adam Stacoviak:** Yeah. Well, I have LG washer and dryers, and I can tell you when my washer is done and when my dryer is done. And I can tell it to start drying something right now, if it's on. So that one's safety -- security and safety. Maybe the similar S'es... It's that I can't control it from afar or remotely if it's not on currently. I can't power it on and power it off. I can only instruct it if it's been previously human-touched, like recently, kind of thing.
+
+**Elecia White:** And would you have paid for those features?
+
+**Adam Stacoviak:** Apparently, I did.
+
+**Elecia White:** I know. We did too, but I would have never meant to pay for those features.
+
+**Jerod Santo:** No. I never want almost any -- like, sometimes you're like "Oh, this actually is kind of nice." Once you start using the feature. You're like, "Yeah, I kind of appreciate this." But it was never --
+
+**Elecia White:** When the dryer's off, it's kind of nice.
+
+**Jerod Santo:** Yeah, it never has sold me on any piece of appliance.
+
+**Adam Stacoviak:** So the one thing I would say about that that I've appreciated - because I try to be not the person who only complains, I try to be the person who appreciates pretty much any scenario. I do my best; I'm always the greatest at it... It's that it's a different interface to a simple interface that doesn't make any sense. And so the iOS application actually makes a bit more sense to use its advanced features, like the dry clean version. There's a dry clean setting on the dryer, for example. It hooks to the water, and it can humidify the air kind of thing. It can do some cool stuff. That feature is not easily accessible on the device itself. I mean, you can; it's just harder to like map your brain to like "Okay, push button, go." Whereas the iOS app is a bit more easier to use. I think of it like a different interface to this thing.
+
+The notification is -- I turn that off in most cases. I actually don't have it on, unless I'm washing clothes personally. And when I say that, I'm like -- I don't usually wash our clothes. So I'm that person who gets to just sit back and enjoy life, for the most part.
+
+**Jerod Santo:** You'd better not be complaining then...
+
+**Adam Stacoviak:** Until I have no more underwear, and then I'm like "Oh my gosh, I have to wash my own clothes now."
+
+**Jerod Santo:** Then you're checking the status at all times.
+
+**Adam Stacoviak:** That's right. Notifications on, thank you.
+
+**Elecia White:** I mean, the app does give you more information than could possibly fit on the front of your washing machine.
+
+**Adam Stacoviak:** It does. Yeah. So that's how I look at it. It's like, I wouldn't have paid for that feature. I think I did because I bought it, obviously... I didn't buy it because of that feature, and I'd say that it's a nice to have, not a need to have... But it is nice to actually have, because I've used some of the advanced features via the iOS app, and I've -- it's just a different way to interface with the device, that's actually better and easier.
+
+**Jerod Santo:** \[01:27:53.28\] But you know, those bits of information that they're showing to you can also just seep out of your local area network, and into the hands of somebody.
+
+**Adam Stacoviak:** Adam is washing his clothes again. Get him.
+
+**Jerod Santo:** You know? Someday you're gonna get an email from a hacker and he's gonna be like "Adam, you're out of underwear."
+
+**Adam Stacoviak:** Yeah. "Hey, listen, you probably should wash your clothes... It's been about two months since your dryer's dried anything. Clearly you're out of underwear at this point. I've checked your Walmart checklist for your underwear, or wherever you buy your --" I don't buy my underwear from Walmart, but I just said that.
+
+**Jerod Santo:** TMI, dude.
+
+**Adam Stacoviak:** Just so you know, okay?
+
+**Jerod Santo:** Alright, Elecia... I assume the website is embedded.fm. Where else can folks connect with you?
+
+**Elecia White:** You can get Embedded from any of your regular podcast sites. And if you are faced with one that is Embed in all black and looks scary...
+
+**Jerod Santo:** Don't do it.
+
+**Elecia White:** We're the one with the robot.
+
+**Jerod Santo:** Look for the one with the robot.
+
+**Elecia White:** It's like a little robot head; robot head, slash radio, old-timey radio... So that's us.
+
+**Jerod Santo:** Nice.
+
+**Elecia White:** You can get my book. It's called Making Embedded Systems, and it's in second edition. It's from O'Reilly. It's on normal Amazon and bookshop sorts of places. If you're interested in making embedded systems, it's a -- the goal is for hardware engineers to understand more about software, and for software engineers to understand more about hardware, and how to design and make embedded systems.
+
+If you want to contact me for consulting, I think the answer to that is please don't. \[laughter\] Please don't...
+
+**Adam Stacoviak:** Please don't? Is that right?
+
+**Elecia White:** We're booked... We're booked for like a year. It's great.
+
+**Jerod Santo:** Congrats.
+
+**Elecia White:** And yet I always love to hear about neat things people are making. So if you want to tell me about your embedded system, hit the Contact link on embedded.fm, or email me at show \[at\] embedded.fm, and I'm just fascinated by the technologies coming out.
+
+**Adam Stacoviak:** Very cool. Any conferences or like IRL places? I know you mentioned several digital ones, which we've definitely collected and we'll put in the show notes... But is there anywhere that's like common? Like, is there an Embedded Conf or something like that, where it's, you know, international or accessible to folks?
+
+**Jerod Santo:** Like the one you can't miss kind of event?
+
+**Adam Stacoviak:** Yeah, like you must go there. If you're really getting into it, you must go to this conference kind of thing.
+
+**Elecia White:** This is another one of those areas where there is a good hacker subset, and there's a professional subset, and they don't quite overlap. I will be keynoting the Embedded Online Conference in May of 2025. Please don't ask me what my talk is about, I don't know yet. But I will really soon.
+
+**Adam Stacoviak:** Embedded things.
+
+**Elecia White:** Probably sensors. And for the hobbyist, or - see, it isn't just the hobbyists. I've gone to Supercon from Hackaday; I would happily go to Teardown. There's Open Hardware Summit... There are a ton of really fun places where people will show you how to get into embedded systems, or hacking devices, or making cool things.
+
+There aren't as many professional conferences, unless they're about certain topics, like Zephyr, or security. Hardware.io is one of the international conferences that's pretty popular, and more security-based. I can't think of any other ones, but I know I will when we get off, so I'll send you an email and you can provide links.
+
+**Adam Stacoviak:** For sure. I know IRL is important to folks... We can obviously shell out lots of digital locations to hang, but IRL - can't beat that, right? Real humans, real people, real fun.
+
+**Elecia White:** \[01:32:05.00\] Really?
+
+**Adam Stacoviak:** I think so.
+
+**Jerod Santo:** \[laughs\] Really?
+
+**Adam Stacoviak:** Maybe not. Maybe not.
+
+**Jerod Santo:** To each their own, as they say.
+
+**Elecia White:** I'm kind of an indoor cat.
+
+**Adam Stacoviak:** We'll have to get some sensors for that, and maybe a motor to judge it.
+
+**Jerod Santo:** There you go. Or one of those things you can touch that we know Elecia is still there.
+
+**Adam Stacoviak:** That's right.
+
+**Jerod Santo:** And she can just send a sensor to the conf.
+
+**Adam Stacoviak:** "Still here."
+
+**Jerod Santo:** Like light up a light. You know, blinking lights. We hear you like those.
+
+**Adam Stacoviak:** Yeah, for sure.
+
+**Elecia White:** As long as they're blinking for other people...
+
+**Adam Stacoviak:** Well, Elecia, thank you for your curiosity, too. I know that sometimes you're going alone, right? Sometimes you're just in the dark. But now you're not, obviously... You're book-solid with your business, you've written books, you've come on this podcast, and you're doing cool stuff, so thank you for sharing all that with us. It's appreciated.
+
+**Elecia White:** Thank you for having me. Your podcast is a lot of fun.
+
+**Adam Stacoviak:** Well, thank you.

The world of open source metadata (Interview)_transcript.txt

@@ -0,0 +1,416 @@
+**Jerod Santo:** Today, we're joined by, for us, an old friend, but a long time no talk... Andrew Nesbitt is here with us. And you know, Andrew, I came across Ecosystems, which is ecosyste.ms - nice domain hack; hard to say out loud, but it looks cool in the URL bar...
+
+**Adam Stacoviak:** It does look cool.
+
+**Jerod Santo:** I came across this and I thought "This is a very cool project. It seems somewhat familiar. I can't quite put my finger on what it could possibly be... And then I saw it was from you, and I'm like "Oh, it makes total sense." This is right up your alley. We've had you on the show many times back in the day, talking Octobox, talking Libraries.io, talking Ruby ecosystem and dependency management... And it looks like you're still out there, kind of beating around that same bush. So first of all, welcome back to the show...
+
+**Andrew Nesbitt:** Thanks for having me. Yeah, it's great to be back.
+
+**Jerod Santo:** Ecosyste.ms. I mean, okay, we have a lot of context that maybe our listeners don't share, but take us back to what you're interested in, which - it seems like you've been interested in similar things for a long time. And you built libraries.io around this, and Ecosyste.ms is a very similar thing... I'm wondering if it's the same old thing, or if it's a new-new thing... So tell us about your past and like collecting and organizing dependencies, and the information about them, and open source projects, sustainability, and then how that brought you to Ecosyste.ms.
+
+**Andrew Nesbitt:** Yeah, okay. So I have been swirling around the world of open source metadata for - it must be coming up to 10 years now. Starting with 24 Pull Requests.
+
+**Jerod Santo:** That's right, 24 Pull Requests, yeah.
+
+**Andrew Nesbitt:** And that didn't kind of start from metadata, but the idea of that project was to encourage people to contribute to open source as part of kind of the run up to Christmas... And after kind of like first getting that off the ground, we quickly ran into "Oh, how do we suggest -- where should people go and contribute to?" And a lot of people would try and send a pull request to a project that had no activity, and like the maintainer was gone, or just were struggling to be able to even like work out how to send a pull request to some projects, because they were really not very friendly or easy to contribute to.
+
+And that kind of led me down this path of "Okay, well, how do you define what a good project is?" And then "Can we scale that up, rather than manually having to have people kind of like submit their things and keep those things up to date every year?", because that project would just kind of come and go every December, and shut down afterwards... So the maintenance there couldn't be entirely human, because there was thousands of people contributing to that project, and sending pull requests... And it was a lot of data to try and work with.
+
+So I started to build out some basic metrics there to try and go "Does this project look like it has activity that's happening on it? Does it look like it's ever received third party contributions?", and things like that. And that led me to kind of -- I got a job at GitHub from there, and then GitHub promptly fell apart internally... Tom Preston-Werner left, it was a horrible time... And so then I left there and started Libraries.io as essentially like a - okay, well, looking at package manager metadata is a different way of kind of getting some measure of what's an interesting open source project. Like, rather than just using stars, which - stars is a terrible metric, and has very little kind of bearing on a lot of projects, especially as you go down from the massive frameworks, those huge keystone projects. Once you get down to smaller libraries, and also especially the kind of low-level critical projects that are doing a lot of the real work... They don't get a lot of attention, and a star is basically a measure of attention, how many people are landing on that GitHub repo page.
+
+So package manager metadata was like "Oh, this is really juicy", because it kind of gives me a hook into saying "These libraries are being used by other people." But download stats - again, available for most package managers, but not all - is often kind of wildly all over the place for certain projects, especially if they're used a lot in CI; you'll just see really inflated download stats. And you also don't necessarily see those for dev dependencies, the things that people, especially maintainers are installing on their laptops to be able to work on those projects. But they're not necessarily a runtime dependency of all the applications; there are definitely gems that Ruby and Rails devs use locally, but aren't shipped with the Rails app, so you would never see those numbers.
+
+\[08:18\] And the insight that I kind of accidentally tripped over was if we go mining the dependency information out of open source repositories at a large scale, you actually start to get a really good picture of how people really use open source, and how they don't use open source. Like, if a project breaks, you probably don't go and un-star that project, let's be honest. Not many people are un-starring things. They don't remember. And also, you don't un-download a thing. The download counts remains after you downloaded it and was like "Oh, this doesn't actually work" or "This is not what I wanted", or it has become unmaintained. Whereas actual "I depend on this thing", if I remove that thing as a dependency, then numbers go down, and you get a really interesting, strong signal that something is maybe not quite right with that project.
+
+So that kind of led me onto a path of "I should just try and index the dependencies of every open source project ever." And libraries.io started out as a search engine, designed to be like "I can help you try and find the best package."
+
+And that was primarily like "This package is well used, so therefore that implies it has good documentation, that it actually works, and other people are using it as kind of a proxy." And it grew and grew, and became a massive and expensive and difficult project to maintain as a side project, whilst I was doing contracting. And me and Ben, who were working on it, we're like "Well, what are we gonna do? How can we turn this into a sustainable project that can fund itself?" And at the time, GitHub had just implemented its own dependency graph as well, along with purchasing Dependabot... And that basically -- they started giving that away for free. That pulled the rug out of any plans we had to monetize libraries.io directly, as well as a project I was building called Dependency CI... Which never really got off the ground, but was back in the day was like "Oh, this is really cool", because it could literally like block your pull request to say "You're trying to add a dependency here that is not good, because it doesn't have a license", or it's got security issues, or other things. And so we ended up selling to Tidelift, to try and find some way of recouping the costs of building out that project... But just before we did, we also made all of the code open source, and all of the data open source. So it was kind of like an airdrop into the community to be like "This is always gonna be here if you wanna use it for purposes."
+
+Didn't really work out at Tidelift... There's a big cultural difference in the founders at Tidelift compared to me and Ben. Me and Ben are very -- we really like building and solving problems in the open, and shipping stuff really quickly, and kind of iterating on those things... And Tidelift's culture was - because they just sold to another company...
+
+**Jerod Santo:** Yeah, who bought Tidelift?
+
+**Andrew Nesbitt:** Sonar? I can't remember the name. It's a security company.
+
+**Jerod Santo:** Okay.
+
+**Andrew Nesbitt:** And as a shareholder of Tidelift, I can tell you, I didn't get anything from the sale.
+
+**Jerod Santo:** Bummer, dude...
+
+**Andrew Nesbitt:** \[12:02\] But Libraries.io was there, and was open source, and after -- I took a break for a little while during the pandemic, which - you know, everyone had a kind of a crazy time... I went to do some contracting with Protocol Labs, basically kicking the tires on IPFS and Filecoin, and trying to use it as a real user... It was an interesting time to actually try --
+
+**Adam Stacoviak:** \[unintelligible 00:12:31.22\] was pretty cool.
+
+**Andrew Nesbitt:** Yeah. \[laughs\] And then at the same time was talking to Schmidt Futures, which is now Schmidt Science... But one of the kind of sub-foundations of the Schmidt Foundation, who were basically saying "We have researchers that were using the data from Libraries.io for research", but now Libraries.io -- when I left Tidelift, they started to remove features of Libraries.io, especially the API access and the data... And Schmidt Futures basically kind of came along and said "Could you stand up another copy of it?" And I was like "We could do that... But what if we rebuilt it from the ground up as infrastructure for research purposes?", rather than taking the same code, which is like one big search engine, one honking great Rails app, and actually make it into kind of a slightly more -- like, take all the lessons learned, but instead of building it as a search engine, instead build it as a base layer of open source metadata, which then can be used to build a Libraries.io on top of it. And that also means we can take some of those lessons that were like "Oh, actually, it turns out contributing to a project that has one absolutely enormous database schema is really difficult." Like, trying to stand that up yourself is really hard as a contributor. So people will just bounce straight off the project, because they're like "Well, there's no way I can possibly comprehend how big the stuff that's going on here..." And then also, the performance implications of deploying a change, that might be like "Oh, you're about to touch a table with like a billion rows in it." That's gonna be difficult for you to test without me giving you production access... And I really don't wanna do that to random third party open source contributors.
+
+So Ecosyste.ms is essentially a do-over of Libraries.io. It's many different Rails apps that are focused on collecting different kinds of open source metadata, and then combining them together in different ways. So there's a packages service, there's a repo service that collects the dependency information from repositories, there's an advisory service, and a commit service, and an issue service... Basically, all the different things that you might be interested in. And each one of them can then be independently worked on and scaled up as different amounts of data \[unintelligible 00:15:10.20\] and kind of collected in different places.
+
+And that has been going on for nearly - I wanna say three years now... Really kind of like going from -- it was a nice kind of year where I just worked on it myself, didn't really tell anyone about it, just kind of like plugged away... And there are core pieces -- because Libraries.io is open source, I was able to reuse the dependency parser and a load of the mappings to the package managers... Actually take that code and kind of reuse that in a way that also allows you to have multiple different package manager registries, where Libraries.io would only support one... Which was really nice when RubyGems had all of its drama recently, and the gem.coop popped up... I was able to go "Oh, I can quickly start indexing gem.coop." It just fits straight into that new schema. And then since kind of like the past year, it's just absolutely exploded in usage. The amount of traffic today alone was 50 million requests to the API.
+
+**Adam Stacoviak:** \[16:23\] Wow.
+
+**Jerod Santo:** Wow,
+
+**Andrew Nesbitt:** And it has become quite a piece of critical infrastructure to a number of different kind of areas of open source in terms of SBOM enrichment, and also trying to find those critical pieces of open source that need security work or need sustainability efforts to be kind of coordinated around them.
+
+**Jerod Santo:** Well, I'm happy to hear that you got to reuse some of your code from Libraries.io, because what I thought was gonna happen when you said "I airdropped it", I thought you were gonna just catch your own airdrop a few years later and be like "And because I open sourced it, I just relaunched it under a new..." But obviously, the big rewrite is a very tantalizing thing, especially when you've been living with all your mistakes for this time. It's like "Let's start over..." But you got to reuse some of your code, which is really awesome. So nice job open sourcing that when you still had an opportunity to do so.
+
+**Adam Stacoviak:** Yeah, absolutely. You mentioned this is used in research... I guess, research terminology, so to speak. What exactly does that look like? Who are those folks? What kind of research are they doing? Are they developers? Are they developer-adjacent?
+
+**Andrew Nesbitt:** I think mostly developer-adjacent, or in the research space I guess you'd call them research engineers... Where lots of computer science researchers are like "We want to study what these behaviors are like across different package managers", or comparing what are developers doing in this space versus that space, especially around the dependency stuff, to be able to go "Oh, the average number of dependencies in a JavaScript app compared to a Ruby app", for example, which I think is about 10X... And then looking at kind of "Can you go down those dependency chains and find where the security problems are, or the license problems are?" And also leading into kind of "How can we encourage best practices in this space?" Or work out "How many projects have taken on these various kinds of", specially -- just recently I had a call with someone who's looking at all the attestations around trusted publishing. Like, can we see the share of usage of packages that have the trusted publishing setup, and are publishing attestations into a SIG store, compared to the overall space? And also then breaking that down across different ecosystems as well.
+
+**Break**: \[19:00\]
+
+**Adam Stacoviak:** This might be silly, but let me ask you this... I've been researching some CLIs and I've been researching how CLIs install themselves. Sometimes they'll leverage the actual package manager of the distro, like a Linux distro or something like that... But most, by and large, just give you a URL to curl and pass to Bash, essentially... Which can be problematic if you don't trust the script.
+
+If I wanted to somehow research CLIs and how they install themselves, and the various ways they install themselves, is that something that this service could do? Is that the level of research it could do?
+
+**Andrew Nesbitt:** Yeah. I mean, for one thing, you would be able to quickly find everything that had kind of tagged itself up as a CLI program. I've also been indexing every public image on Docker Hub, and basically running an SBOM scanner against each one of those. There would be some juicy insights there, to be able to go "How many--"
+
+**Adam Stacoviak:** Juicy? I like juicy.
+
+**Andrew Nesbitt:** "...of these things were installed via a distro package manager, versus "we just have a URL for this"? Which would be recorded in the SBOM, basically to say "Oh, we've found this known bit of open source, and it appears to say that it sits in this file system here", which implies it was installed by apps, or it's in a random space, like it was probably curled down, along with the Docker file that was used to build that image. And there's a good kind of million open source Docker images on Docker Hub, or at least individual versions of things. And you also get the interesting aspect there that you can kind of multiply that by the number of downloads that some of these Docker images have... And some of those numbers are crazy. Millions and millions of downloads of a particular image. And of course, those numbers inside that one container are never reflected in the package managers upstream. So just because it was downloaded in Docker doesn't mean that that actually shows up as being a million downloads in RubyGems, or on npm. So you start to see some really interesting things, and you start to see those download numbers, or the proxy for a download number of distro packages as well... Which is a really hard number to get a hold of, because every distro package manager is very heavily mirrored, and basically just a file system somewhere exposed over HTTP or Rsync. So no one has good download stats for those things. The only place you really find that is the Debian popularity contest, which is opt in, not opt out.
+
+So you'd be able to go "Oh, okay. Well, I can see -- here are the CLI programs that are being manually downloaded inside of Docker images as part of this install process." It's not gonna give you everything, but it certainly gives you a good proxy for "Okay, well, I can see where --" Like, relative usage of these things starts to show up, which is where I've found the most useful ways of kind of sorting different piles of packages or whole registries, is to go "Okay, well, if I sort this registry by the number of dependent repositories, or the number of dependent packages, which things show up at the top?" And then also, which of those things make up 80% of all of this stuff?
+
+\[24:28\] And you actually end up looking, like -- I like the 80/20 rule, but it doesn't actually turn out to be like 20% of packages make up 80% of usage. It's like 0.01% of packages make up 80% of usage. These tiny amounts. There might be 2000 Node modules total that make up 80% of all of the usage of npm in terms of downloads, and in terms of discrete dependent repositories...
+
+When you then start to really focus that lens, you see a long tail of stuff that never gets used, and it's also like all kinds of spam and malware and stuff that floats around. But there's 10,000-15,000 packages which are the packages that make up most open source usage across all these ecosystems.
+
+**Jerod Santo:** It's kind of amazing how massive that asymmetry is when you pin that down to the individuals...
+
+**Andrew Nesbitt:** Yeah, and that's on average one maintainer per package at that critical level as well... So that's like 15,000 people maintaining all of open source usage.
+
+**Jerod Santo:** That makes the XKCD comic even more poignant. Now you're the one person in Nebraska - replace Nebraska with wherever they are in the world; probably in different towns...
+
+**Andrew Nesbitt:** And how many of them have you had on the Changelog?
+
+**Jerod Santo:** That's a good question. Probably a good percentage of those. Oh, man... So there's 15,000 people basically running the world for free. \[laughs\]
+
+**Andrew Nesbitt:** Well, I have done a little bit of indexing of how many of those have GitHub Sponsors, or are their projects on Open Collective, or they have some other kind of funding link... And in terms of those top critical packages, it comes out to kind of like - depending on the ecosystem, it's somewhere between 25% and 50% have some way of "Here's an automated way you can give me a donation to the project."
+
+There's a good chunk of those as well that are massive, corporate-funded projects. Like, all of the AWS RubyGems that make up the AWS CLI are in the top of RubyGems, because they're just massively used. They don't need any funding, because Amazon has full-time staff. But there's a good --
+
+**Jerod Santo:** They might need some funding. I hear they're laying people off again... \[laughs\]
+
+**Andrew Nesbitt:** Hopefully, they didn't lay off all the Ruby people maintaining the CLI there.
+
+**Jerod Santo:** Yeah, that would be awful. So do you track -- so you're tracking those who are able to receive funding in some sort of automated fashion. Do you track funding itself? Like, who's getting how much money, and how?
+
+**Andrew Nesbitt:** Yes. Well, where possible. So I'm tracking -- I call it a funding link, and some package managers have funding links support, where you can say "Oh, you can donate to me over here." Repositories have the funding YAML file, and I go looking for that wherever possible. And you actually see that even on GitLab and Codeberg. I don't know how well those platforms display it in the UI, but it definitely -- because obviously, GitHub Sponsors is not... I don't think there's a GitLab Sponsors or a Codeberg Sponsors. Those files do show up all over the place.
+
+\[27:54\] And then also being able to go "This repository is owned by a user on GitHub who is part of GitHub Sponsors" is another way of kind of detecting that. Even if they haven't added their funding YAML file, we can kind of make a hop to say "Oh, here's one of the maintainers to be able to support that." I then collect the data from GitHub Sponsors of every -- because GitHub Sponsors users are public. You don't get any financial numbers, but you do get "Here's the number of active sponsors of things, and here's the total, like all-time..." It's quite hard to get time series data out of that API, so instead I basically just kind of snapshot it on a regular basis, to go "Oh, here's what's the current state of the world in terms of GitHub Sponsor funding."
+
+It's a bit weird, though. A lot of people who have realized that GitHub Sponsors is actually quite a good way to sell digital goods. If you go looking at the top users of GitHub Sponsors who have the most people funding them, they sell things like avatars, and Discord memberships, and eBooks, and things like that. They're not necessarily kind of selling "Oh, I can maintain this project better for you." That's not -- like, Open Collective is so much bigger in terms of actually like supporting the projects as a collective, because they're just set up in a totally different way to GitHub Sponsors.
+
+**Jerod Santo:** Yeah. That's fascinating. So they're kind of doing sponsorware, insofar as it's not a donation, or "You're supporting my work on this project." It's like "Actually, there's a quid pro quo here. We're going to trade a good or a service for that sponsorship money." Really, it's a purchase of some sort of thing.
+
+**Andrew Nesbitt:** Yeah, yeah. If you go looking, it's easy to see GitHub doesn't make it particularly -- like, they don't have a leaderboard... Which is a good thing, to not -- like, putting a leaderboard on things can often produce some very strange behaviors...
+
+There's also an interesting breakdown of like number of users who sponsor other maintainers, versus companies. Obviously, companies are going to sponsor a lot more in total amount per company, but the distribution is quite surprising in -- you're looking at easily 10 times as many individuals are sponsoring other people on GitHub Sponsors compared to the number of organizations. Like, it's quite small, really.
+
+**Jerod Santo:** Really?
+
+**Andrew Nesbitt:** And most of that activity is public. So it's not like there are -- you can be anonymous as a GitHub Sponsor, but you can't really hide the fact that there is a sponsorship happening there. There's also on Open Collective some massive donations that go to certain projects through company sponsorships, because they're acting as a fiscal host, rather than just being a platform to collect tips, which is basically how GitHub Sponsors works.
+
+**Jerod Santo:** Right. It reminds me of way back in the day, Chad Whitaker's Gittip, which was later called Gratipay.
+
+**Andrew Nesbitt:** Oh, yes.
+
+**Jerod Santo:** Remember that?
+
+**Adam Stacoviak:** Oh, yeah.
+
+**Jerod Santo:** And it felt all warm and fuzzy, because people were getting money for their open source, but when you go looking at it very closely, most of that was like the same 50 bucks getting passed around between friends... Not a slush fund, but like a -- they just felt good. So I would make 20 bucks a month and I'm using open source, so I would give it to somebody else. And there was really no new -- not enough, new money coming in. It was really just money that already existed amongst all of us maintainers kind of patting each other on the back... Which was unfortunate, but just the way it started.
+
+**Andrew Nesbitt:** I definitely do that. I sponsor 35 different people on GitHub Sponsors with just a few dollars a month, to just be like "I appreciate your work." I don't have a huge amount to support you with, but just as a way of saying "I noticed you and appreciate that you continue to maintain these things that I use."
+
+**Jerod Santo:** \[31:59\] Well, I hoped GitHub Sponsors was big enough and mainstream enough to kind of change the shape of that. And maybe it's done it some, but it sounds like there's still more indies passing person-to-person kind of sponsorship than there is corporate-to-person.
+
+**Andrew Nesbitt:** Yeah, I think the change of interest rate across the world had a massive impact. The nice thing about Open Collective is they are -- especially Open Source Collective is very public. You can see the amounts of donations going in and going out... And there was a big drop around the time that -- like, post COVID hit and changed all of the finances of these things, and it was like "Oh. Okay, well, open source is no longer one of the--" It's an easy line item to drop, because "Oh, everything is free, and it just continues to work..." For now, until a security problem comes along and then everyone starts scrambling again.
+
+**Jerod Santo:** So you've got 12 million packages being tracked, 287 million repositories, 24.5 billion dependencies, 1.9 million maintainers... I'm reading these stats off of your website. There's a timeline of like public events on GitHub, there's issues, there's commits... I mean, there's just tons of different data points that you're tracking. How do you store all this stuff? Where do you store it? How big is it all? Because I'm just thinking this is a data management nightmare.
+
+**Andrew Nesbitt:** So that 24 billion dependencies is a bit of a headache.
+
+**Jerod Santo:** \[laughs\] I bet. I mean, that's crazy.
+
+**Andrew Nesbitt:** Almost all of this is stored in Postgres.
+
+**Jerod Santo:** Okay...
+
+**Andrew Nesbitt:** Individual Postgres instances on dedicated machines in France and Amsterdam, mostly because they're very affordable. Online.net is a very reliable host, similar to Hetzner or some of these other kind of bare metal machines.
+
+So I do the maintenance of the machine myself, and obviously, scaling up is a little more tricky, because there's not just a nice Heroku slider anymore... I use Dokku as essentially like the open source Heroku, which is really nice.
+
+Just git push, it builds your Docker image, and then it handles putting NGINX, kind of proxying all of those things. Very nice for like an individual machine. It doesn't really give you any kind of multi-machine things, but I try to avoid too much complexity when there's only a very small number of people working on doing the infrastructure. And it's mostly me, rather than -- I calculated, like a back of the napkin thing the other day, I think it would cost me 15 times as much to host on AWS as it does to host it on dedicated machines right now... But these Postgres -- each service basically has its own database. So rather than it being one that is enormous, it's split out... Which at least makes it kind of like I can work on individual ones and be like "Oh, this one is reaching capacity, so it's time to scale it up", or "I should make another box of web machines or Sidekiq workers separately. I don't need to kind of do everything in one big lockstep", which keeps it fairly easy to do.
+
+And then the whole website is basically read only. Like, you can't ask -- you can't put data into it as a user. You read from it. And all the data comes in in the background through loading data from package managers, and repositories, and... There's about 2000 different Git hosts in there that I'm constantly crawling at different rates to go like "Oh, there's new activity over here." So I can cache things very aggressively at the kind of HTTP layer. I think the cache hit rate at the moment is about 60% in Cloudflare. At some point I've got it all the way up to like 95%, but then you get some AI bots come along and they do some weird stuff, and it's very hard to cache such a long tail of billions and billions of URLs that might exist on the platform. And Cloudflare on the free plan is not gonna cover an unlimited amount of cache. You just kind of keep rolling over the cache, over and over again.
+
+**Adam Stacoviak:** \[36:31\] Is this a solar project again, or is this you and Ben back together...? Who's the band?
+
+**Andrew Nesbitt:** So Ben is working on it part-time. He is also one of the directors at Open Source Collective, which is -- you know, that's a lot of work in itself. And then we have a few people who are doing some part-time work. Martin has done all the design work... Which looks so much better than my efforts of the original -- you can see, there's a couple of older hidden web pages there that are very poorly designed, which is just me making some plain bootstrap pages... And we just had James come on to help with making the project better-documented and easier to onboard as a contributor... Because I was running so fast on standing everything up and scaling it up and collecting all that data that I didn't really leave a lot of documentation along the way... Which is terrible, but - hopefully, these are pretty basic Rails apps. There's not a lot of interesting stuff. Like, intentionally trying to make it the most boring tech possible, so that I can focus on the interesting stuff, which is like the parsing or the mapping of the metadata... Which is like each app has that core little nubbin of "Oh, here's where the real logic sits." And that's a nice, well-tested bit of functionality, with a load of Rails scaffolding around it to be like "Okay, write this into Postgres and then serve it up in kind of the quickest way possible."
+
+**Jerod Santo:** How many apps is it now?
+
+**Andrew Nesbitt:** Oh, good question. It must be coming up to 20... But some of them are quite small. There's a load of services that are kind of like stateless. Like, I will just give you a SHA-256 of a tarball that you get from RubyGems or similar, and a lot of those I basically have on the chopping block to try and turn into something a little bit more like -- imagine a GitHub Actions, but for analyzing packages... So rather than it happening every time that you commit or every time you open a pull request, instead it'd be like you can define "I wanna run this kind of analysis on this package when a new version comes out." That might be like copyright and license extraction, or it might be "Do me a capabilities analysis of this go package using the Caps Lock library..." Which will basically go like "Oh, this library just gained network access and it can read environment variables, and it became a crypto miner." It would be a great way of like being able to highlight some of those changes.
+
+So I wanna pull it down, and make it a little bit kind of like fewer services, but one of those services will be basically the "Which open source analysis do you wanna run against this package?" And then "Here's a massive fire hose of every activity that is happening", and you can hook those analyses in, to say "Okay, I wanna run Zizmor every time I see a GitHub Action change", because Zizmor does the security scan on the YAML config to go like "Oh, you've just introduced a foot gun of GitHub Actions here." And then try and publish all of those analyses back out as a public good, just basically fling that into S3 or something as a way that allows researchers, again, to go and do broad analysis over the whole ecosystem, or multiple ecosystems, without having to spend all their time collecting all of that base data, and normalizing it, and then setting up infrastructure to run all of that across all of those packages. I see that time and time again, where the paper is -- like, 50% of the work is "Oh, well, we had to collect all of this data, and we had to make sure that it all fit into the right box."
+
+\[40:28\] And then we could actually start doing the interesting research. So what I hope is we get to a place where it's like "Oh, you don't need to do that. You can just use this open dataset", and that gives you a good starting point to then start to really dig into like "What's going on in these ecosystems?" That's the dream anyway.
+
+**Jerod Santo:** Well, you're certainly working your way towards that. So does Schmidt Sciences - do they foot the bill for all of this work?
+
+**Andrew Nesbitt:** So they gave a grant initially, to get started. Luckily, they gave it in dollars, and the exchange rate was very positive for a while, so we actually managed to stretch from a one-year grant into a two-year grant... And then Open Collective has been supporting the project as well as a fiscal host, but also as like a customer. So I built a number of tools for them, to help them kind of investigate ways of trying to expand the ability to kind of let companies fund open source, and then also to try and measure the return on investment of giving two projects and try and be able to see "Oh, if I donate money here, or resources, does that turn into actions and changes on the repositories?" And that kept me busy for a good nine months, I think, of building out tools for them whilst they financially supported the project.
+
+And we also have a number of customers who pay for a different license for the data. So the data is CC BY-SA, which is like a copyleft license. You can use it for whatever you like, as long as you also persist the license and you credit where it came from. But if you don't wanna do that, then you can pay to essentially have a CC0 license. It's not actually CC-0, because there's some things there to say "Oh, don't just completely undercut us and sell that on again." But we have a number of customers there... And that basically pays for all the hosting costs.
+
+So it's self-sufficient, it runs itself, as long as -- but you don't get any extra feature development on top of that. So that's where I'm trying to work on right now, is to get that level of sustainability higher. And we've just received a grant from Alpha Omega, to basically make that happen.
+
+Alpha Omega is part of OpenSSF, and their goal is turn money into security. And they have become a big user of Ecosyste.ms for doing analysis of like who are the critical projects in a particular space, who are the ones that are gonna be most likely impacted if there's a big security vulnerability? Who are the ones who have never had a security vulnerability and maybe don't know what to do if they get one? ...things like that. So they have basically given us a grant to try and help make Ecosyste.ms long-term sustainable. So that's things like making the project easier for people to onboard onto, and also to be able to kind of charge large companies in different ways. That might be like "Oh, you want an even higher rate limit than the very friendly rate limits that are already on there? Do you wanna go even harder? Well, then you can pay for a super-rate limit, or similar."
+
+\[44:07\] And then also this kind of pipeline of analysis will be another way that -- it'll basically be like "Oh, you wanna run your LLM queries across all these package source code? Well, then you can funnel it through here. We'll just like tee that up and trigger it every time that we see a new release of a package, or similar." It will be another way that I think would be -- essentially, just like "Oh, you've just paying for our CPU to do this analysis", and then the analysis that comes out the other side, if it is idempotent, I guess... LLM queries are not idempotent. You're gonna get a different thing every time you do an analysis. But for a lot of those things it will just come out as a public good, and companies will have paid to have it generated, but then it's shared for everyone to use... Which I think is a nice thing.
+
+I mean, what I'd really like to be able to do then is to actually do revenue share with the people who are maintaining those individual command line tools that do the analysis. Imagine being able to go like "Oh, we can help with supporting Zizmor, and Bullet", or all of these different things that are command line tools that analyze source code. And rather than you build a whole enterprise company around your command line tool, you can just focus on making that tool really good, and then we can run it at scale for customers, and then just funnel the money back to the maintainers, after whatever infrastructure costs there were to run it, so that you can actually focus on building the open source tools, rather than building the scaffolding around it.
+
+**Jerod Santo:** That would be super-cool. So it sounds like there's a collection of potential income sources, some that are currently working, other ones that you're working on... The relicensing of the data for a fee seems like a good one. Is that potentially -- could you see a world where there's enough people that want to do that, that that could be enough, or no?
+
+**Andrew Nesbitt:** Yeah, I think so. Especially this kind of dependent data, the 25 billion row table is really juicy in terms of the insights that you can get from that. The general package data though is often -- like, you can get Claude to generate you an npm scraper very easily. If you ask it to do it in Ruby, you get code that looks a lot like Libraries.io \[unintelligible 00:46:33.17\] \[laughter\]
+
+**Adam Stacoviak:** That's awesome.
+
+**Jerod Santo:** Do you get a nickel when that happens, or what happens? \[laughs\]
+
+**Andrew Nesbitt:** No, unfortunately not.
+
+**Adam Stacoviak:** \[unintelligible 00:46:39.08\]
+
+**Jerod Santo:** Yeah. Well, you know, imitation is the sincerest form of flattery... So just remember that.
+
+**Andrew Nesbitt:** Yes. It's tricky to get that kind of balance of like -- we want to give away as much as possible, especially as all of this data comes from open source. Like, it should be open, because it is data about open source. But then how do you continue to pay for that? ...whilst companies also can kind of go like "Oh, I could just go fetch it from the source myself." And trying to get as many different ecosystems support in is a good way of kind of going -- like, you really don't want to try and index the R package manager. Like, you're not going to have a good time.
+
+So we try and take care of all of the horrible bits... And then also being able to fetch the Linux distro package managers, which is something that I'm trying to add more distro support in... Because each one of those has its own kind of like horrible rabbit holes of weird and wonderful metadata. And trying to work out "How does this fit into the schema?" A lot of it is kind of trying to tie it around the package URL format. Perl - but not Perl word language... Although you can have a \[unintelligible 00:47:55.09\] That has kind of come out from efforts in the SBOM world, and originally, one of the inspirations was Libraries.io being able to map these things into different ecosystems and kind of say "You have an ecosystem, you have a name of a package, and you have a version. Can we talk about this in a fairly standardized way, as a way of transporting these package bits of metadata between different platforms that are doing analysis of different kinds?" And SBOM is the natural conclusion of that.
+
+\[48:41\] Of course, you have two different SBOM standards. There can't just be one standard for things... But being able to look things up by \[unintelligible 00:48:50.15\] is something the ecosystem does really well, because you can basically then take an SBOM and just work through it, every single package that's in there, and say "Can you tell me about this package? Can you tell me what security advisories are affecting the version that I've got in my SBOM?" And that is the biggest use right now, is there are lots and lots of people with GitHub Actions that are just enriching their SBOMs with this kind of information.
+
+It's funny how much more traffic we get on a weekday than on the weekend... And I think it's just because of the GitHub Action kind of like "Oh, this is happening every time someone commits", so you see a smash of traffic of them enriching their SBOMs and checking out every package that is in there... And then the weekend comes along, everyone stops working, and the traffic shape completely changes. And also the cache hit rate completely goes through the floor, because suddenly it's like "Oh, there's all kinds of other weird and wonderful things happening at the weekend", especially lots more like researchers and hobbyists using it.
+
+**Jerod Santo:** So you've mentioned a few of the weird, gnarly things like multiple SBOM specs etc. You have 35 ecosystems on here. npm, Golang, Docker, to name a few. Crates... NuGet, so you're in that world... Across 75 registries - so I'm assuming some ecosystems have multiple registries...
+
+**Andrew Nesbitt:** Yeah, Maven especially. There's lots of registries in the Maven world.
+
+**Jerod Santo:** And then -- oh, even Bower.io. I remember Bower. I don't know if people are still using that... Anyways.
+
+**Adam Stacoviak:** Forever ago, man.
+
+**Andrew Nesbitt:** No one adopts anything, they don't accept any new packages, but you'll still find people that use them and download stuff through them, yeah.
+
+**Jerod Santo:** So what I'm wondering is, where are the black sheep? Where's the gnarliest, weirdest -- like, let's not... I don't wanna create any enemies for you, Andrew, but which of these ecosystems are, in your own heart of hearts, notoriously hard to work with?
+
+**Andrew Nesbitt:** Well, the hardest bits are often the change over time, especially when you go back to the really old stuff. The classic one is that you'd think "Oh, npm - their names are case-insensitive." But if you go and try and index every name in npm, you will find about a thousand that are case-sensitive and have clashes with a different, cased version of the name. And those still exist on the registry. They haven't been removed. And so if you try and make an index against that, you're gonna have a bad time, because as soon as you actually go to run that, you're like "Oh, that's not like that anymore."
+
+So there's things like that, that when you go back into the time -- going back further and further is like "Oh, there's weird things here", especially when the package manager registry has a document database, rather than something that is always enforcing its schema in every record. And npm used to be CouchDB, which is "Oh, they've changed some schemas of the package metadata, so in new packages it looks different than old ones." Of course, now it's actually Postgres underneath, and it pretends to be CouchDB, which is interesting, and I imagine a headache in terms of actually maintaining that... But they still have some really old and weird -- you just run into like "Ah, this bit of metadata isn't right for these few packages", because it was frozen in time. There's JSON in Postgres now, somewhere... Similarly with Maven, they've got lots of different kinds of POM XMLs...
+
+\[52:40\] And there's so many features in the way that Maven can have these nested and parent POMs that is -- I don't really have a background in Java, so I've never used Maven as a user, but the amount of different ways that you can describe the data that is stored in a POM XML, and then published out to Maven Central... Of course, once it's on Maven Central and it's like frozen in time almost, they don't then go and update -- like, if RubyGems adds a new attribute to their registry, that becomes available in the metadata for every single endpoint, because you know, it's just a Rails app that's generating JSON. But for the things that store the files as a historical "We just dumped this file somewhere", then you're like "Okay, my code needs to be able to know every different possible version of this, how this worked, and then also be able to recover from it."
+
+The worst one is the R package manager. It's not huge, but it is used a lot in the research space... And they don't have an API. You have to scrape HTML from the thing. They also remove packages quite regularly, which is very strange. So R has this really weird -- I think it's because it's come from a scientific kind of like non-developer background. It also has one indexed arrays, which not many programming languages have that, right? But their package manager won't let you pin to an older version of something. It won't say "I want version one", even though version 2.0 is out. And the knock-on effect of that is that -- so as a user, if I'm gonna say "Install my R packages", I always get the latest version of everything. That means that if something's broken because something else got a new version, rather than the new version causing the breaking change be told off, it's actually the package that didn't upgrade to fix the problem with this other package that just updated.
+
+\[54:57\] So if you're not proactive in fixing breakages with your package being used with other packages, your package gets removed. It gets kicked out of that registry. Which is pretty wild, because people, especially in science, trying to make their science reproducible, are like "Oh, my package got yanked. How am I supposed to reproduce this science? It's no longer here." So they have some very strange behaviors where they'll actually make snapshots of the registry, and then -- so you can say "I wanna install my R package from this registry on this day", so you actually have like a weird historical aspect of the thing... Which is not like a lot of other package managers. And it's very hard to change, because there's just not a lot of -- we don't have a lot of funding in open source, but in terms of research software engineers, there's no incentive there to maintain and develop software, unless it has a paper attached to it. If you can get citations, great. You can continue to make a case to keep working on those things. But once it's done, it's done, kind of thing. You're like "Oh, you already published that paper. I don't need to continue maintaining the software." That's something that I have an interest in trying to solve, but it's a very hard problem to kind of break into. But what I'd like to be able to do is go "Can we connect the world of papers and citations back to the software that's being used", to especially - like, there's a lot of Python code that might not look like it's massively used, but then when you kind of go "Oh, but it's mentioned in all these papers", especially the kind of AI papers as well, which are just like exploding at the moment... If you can then say "We can send some of this transitive citation credit down the dependency graph to the transitive dependencies of the things mentioned in a paper." Like, I bet there are maintainers who have no idea that their low-level Python or Julia code is being like referenced in these massive papers. That's the discovery aspect there... But also, for the people that do know, to be able to go back to their institution and say "Look, my software is supporting all of this research that you're publishing. You should also support me, because that will make your research better", would be a really cool thing to make happen.
+
+**Jerod Santo:** Right. Until they say "Well, we already published those papers, so... Who cares?"
+
+**Andrew Nesbitt:** \[laughs\]
+
+**Jerod Santo:** That attitude makes it tough, for sure.
+
+**Andrew Nesbitt:** Yeah. There's a lot of still that kind of "Oh, open source is just there. I can just use it. I don't need to contribute back in any way, because someone else will do it." It's still a totally unsolved social problem, I think, in the wider open source space.
+
+**Jerod Santo:** Well, if somebody wants to write a paper on the reproducibility problem in scientific papers due to mismanaged packages in the R language, I think that would be a hit. I think it'd be a hit.
+
+**Adam Stacoviak:** Oh, my gosh... I'm still dumbfounded that they would not let you pin to an older version.
+
+**Andrew Nesbitt:** I know.
+
+**Adam Stacoviak:** I feel like that's gonna break so many research projects that go stale, essentially.
+
+**Andrew Nesbitt:** Well, there's the Software Heritage Project, which is a massive index of the hashes of every file ever published to any open source thing. It basically was produced to try and help solve that problem. Like, you had to make a full index of every file in every Git repository to be able to try and get around the fact that you can't pin to older versions in R's package manager.
+
+I mean, there are still other package managers that don't have lock files in them, which... If you think like years ago - yes, it wasn't such a problem. But nowadays lock files are so critical to the way people build and maintain and share their software, to be able to go "Oh, it works on my machine. It should work on yours, because you're literally installing the same set of dependencies." And Docker works for that high level, but as soon as you wanna change one thing, you'll obviously blast away the whole Docker image and have to start over. Whereas a lock file works really nicely at the language level to be able to kind of solve that problem. If your package manager doesn't have one, you should definitely try and get that added in somehow.
+
+**Break**: \[59:34\]
+
+**Adam Stacoviak:** Behind the scenes I've had some AI literally obliterating your API... With the polite mode on, of course. I've passed my name in so you can track all the things I'm trying to do here, but... It has finally found a way to craft a script that will pull back essentially some version of curl-fssl blah, which is the URL where the thing lives, and then piping that to sh. And so I've got a nice, dramatic list of projects to research that use that command, and what that install.sh script looks like, and what are some of the details in there. So it didn't take long, but my gosh, if I did not have AI to do this for me, I would have pulled my hair out so badly. Probably not your API by any means, but just more like - you can get the data, it seems, but you've got to comb through it, you've got to be persistent, and very...
+
+**Andrew Nesbitt:** Well, there's a lot of -- the schema is not simple.
+
+**Adam Stacoviak:** No.
+
+**Andrew Nesbitt:** Unfortunately. And it's hard to find a way to describe that in a way that doesn't just -- people will just switch off and kind of glaze over as you start going into the levels. Something that I've also tried to do over the past couple of years as the AI bots have kind of gone mad is actually let them scrape the website. Rather than block them, I've said "You can go mad", in the same way as I used to let Googlebot go mad on Libraries.io... Because two years in -- like, we've had a full training cycle of the frontier models. They actually know what Ecosyste.ms is, and they know the structures of the APIs, and they can actually just suggest those things... Which is a good and a bad thing, but I think in terms of being able to get into the training data in terms of like "My API is here, and my service exists" is helpful to people who are using AI coding agents to do some of these things.
+
+I have dabbled in the MCP world with this stuff, and it would be very easy for anyone to build an MCP adapter on top of this, but the security implications really hurt my brain. So I have kind of like held off going hard into it, because every string that is returned by the MCP is essentially like a prompt injection. So you imagine your version number that is pulled from an npm package and then fed through an MCP server into your context... They have the ability to make a version number, especially if it's like semver with your pre-release string on the end of the version number, you could make prompt injection \[unintelligible 01:05:00.19\] where I just start putting like "Ignore all previous instructions -1.1" in the strings of the thing that come from the package manager is suddenly a security vector. Or even just the description of the package, or the name of the package... There's a lot of trust that happens on that kind of go through when it comes out as an MCP server on the other side. If you're just saying "Blindly install whatever the MCP server told me", then there's a lot of trust that you're putting into many layers of indirection that could happen. And we've definitely seen loads of threat actors have realized how -- I'm gonna use the word 'juicy' so many times... But in terms of being able to go like "There's no restrictions. I can publish things to a package manager, and that might be the \[unintelligible 01:05:58.05\] level of indirection before I actually get to my target." That is very hard, to see all of the moving pieces until they actually kind of all come together... But most of these package managers have zero restrictions in what you can do.
+
+\[01:06:15.27\] Even GitHub only just recently started kind of saying there are certain restrictions in how you automated the npm publishing can be. Because people were literally like "Every commit, I'll just publish a new version. Why not? There's no restrictions. A hundred versions a day." Which is like, "Why are you doing this?" Well, because we could. And the cost to the registries is mad as well. You see that PyPI are just showing their numbers continue to grow, and they're like "Well, how the hell are we gonna continue to fund this? Because it doesn't look like it's gonna stop anytime soon." It feels like there's a lot of challenges that are kind of coming down the pipe for these shared open bits of infrastructure to keep them as open as they currently are.
+
+**Adam Stacoviak:** What is your take then with this rate limits and polling when it comes to this polite nature you have here? How do you leverage that? Because I can pass in my email, but then you say "Well, I can reach out to you later", you're watching my rate limits, of course... Can you just shut me off because of me passing that email to you? Or how do you curb the enthusiasm, so to speak?
+
+**Andrew Nesbitt:** So right now we have this -- we have the anonymous rate limit, which I think is 5,000 requests an hour per IP address, basically... And then the polite pool, which is a term we borrowed from a service called OpenAlex - which is basically like Ecosyste.ms, but for research papers. They have this -- if you pass in your email address as part of the user agent, then you just get an uprated rate limit, so that if we see that you're smashing the API, we can contact you and say "Oh, what are you doing? Can we help you do this in a different way?"
+
+So far, I haven't actually been tracking that particularly closely. I've literally just like "Great." Cloudflare is still catching most of that stuff before -- if you hit anything that's cached, it doesn't even touch your rate limit. So it's only the uncached things that actually affects that rate limit. But even then, it's like, 10,000 requests an hour - if you're really, really hitting it, you're gonna run into that. And then a 429 request is very cheap to serve up. So I can serve up a lot of rate limit used requests before things start to fall over.
+
+And then looking at the patterns and going "How are people using this? And is there a way I can do a higher-level API that avoids having to have someone do that \[unintelligible 01:08:59.00\]? Or is there ways of being able to export big chunks of data", rather than doing individual... Lots of little queries is another thing that we're exploring. It may be like a big click house, with a read-only - like, you can write your SQL-ish query against a column store worth of data, similar to BigQuery, but without the "Whoopsie, I spent $3,000 on my one query through BigQuery, because it pulled in terabytes of data." But that is a bit of an ongoing side project. It's not actually live yet for anyone else to use. But hopefully, for researchers especially, you'll be able to just be like "Oh, I can just do big, sweeping queries in kind of an offline way", rather than it having to hit the live Postgres databases... Because that's like the source of truth of these things. And often, researchers aren't like "I need the most up-to-date, within the hour changes." They're like "Ah, actually, I'm fine with this if it's like a day or a week old." It's really not too much difference compared to, you know, "I'm looking for the security advisory stuff that is as fresh as possible", which is often where you're scanning your SBOM and trying to find "Where are there new vulnerabilities that are affecting me?"
+
+**Adam Stacoviak:** \[01:10:24.19\] Yeah. How do you prioritize your time, I suppose? It's a lot to cover, it seems. It seems there's a lot that -- you know, even discoverability. Like, if I am naturally interested, "How can I pull this data out? It seems like I would have to spend a lot of time to figure that out..." That's okay, but... Who is your user? How do you prioritize your time? Who are you building the platform really for? I know who's using it, but how do you prioritize your time to how it's being used?
+
+**Andrew Nesbitt:** Well, to be honest, the number one user is me right now.
+
+**Adam Stacoviak:** Okay, good.
+
+**Andrew Nesbitt:** That's who I prioritize for, because I have a good picture of how you'd want to be able to pull this data out. So the APIs are -- each one has its own open API YAML spec, which kind of tells you "Here are all the different endpoints that you'd want to use." And then I'm building applications on top of this data as well, and going "Oh, this is not here." Or "I want to be able to do it like this." So often, a lot of those APIs have shown up because I couldn't get them to work right.
+
+Josh Bressers has also had a good amount of input in just like absolutely thrashing various aspects of it to look up lots of data around CVEs, and the kind of rate of versions being published. There's also kind of loads of tools that have been built on top of it. Snyk has a tool called Parlay, which does SBOM enrichment... So I can then go and -- these things are open source, I can go and look at them and see "Oh, how are they currently using the existing API? Is there a better way that I can do, or do I just need to beef up the caching in some of these kinds of places?" It's very much like the prioritization a little bit is like just running around, putting out fires, but then occasionally, it's like "Right, I'm gonna turn everything off and I'm gonna go and tackle one of these slightly chunkier problems", of essentially like solving a bigger challenge than just "Oh, there needs to be a new API." Often that's like "Oh, there needs to be another service for another kind of data", or "There needs to be another way of querying this thing, because lots of people have been asking for this."
+
+The biggest thing is just -- coming and asking for things on the issue tracker is a great way to kind of kick off that conversation and say "Oh, I've been trying to do this... I'm trying to solve this problem, but I can't work out how to go through -- I've hit a wall here, or there's just too many individual bits of data over there... Can there be an aggregation of this thing somehow?" And sometimes that's easy, and sometimes it's like "Oh, actually, if we make this index, it's gonna be like the index itself is like 500 gigabytes in size." That's hard to fit into RAM, so maybe we think of another way to solve that problem, rather than just adding an index for every single different way you might wanna query Postgres.
+
+**Adam Stacoviak:** I've found the introducing Parlay post, they even mentioned that "We're enriching Parlay", it's enriching these SBOMs using Ecosyste.ms. So are they one of your paying customers then, considering this tool is probably part of their...?
+
+**Andrew Nesbitt:** \[01:13:57.00\] No, they are using -- so Parlay is an open source tool that other people can use.
+
+**Adam Stacoviak:** Okay, gotcha.
+
+**Andrew Nesbitt:** And it's primarily companies, because open source developers don't actually care about SBOMs, because they're like "Here's the code."
+
+**Adam Stacoviak:** I had to search what SBOM enrichments was. I guess I should have guessed that by -- take a little bit of data and make it better, I don't know... \[laughs\]
+
+**Andrew Nesbitt:** Well, most SBOM extractions don't -- like, when you produce an SBOM from, say, a repository or from a Docker container, it will go "Here are the packages and the version numbers", but it's not gonna tell you "And here is all of the information about that package", because they just don't have that on disk, available, most of the time. Some package managers, especially the distro package managers do actually have that information right there... But these SBOM generation tools don't go and hit the npm API directly to fetch all of those things. So if you wanna be able to get a high level overview of all of the license breakdown of all the different packages in your SBOM, then you need to enrich it by basically going through each one and fetching some extra information and filling in the license field. Maybe they like maintainers -- there's a load of different things in there and it depends on which SBOM standard you're looking at as well, because they're different... But also, just being able to look up all the security CVE stuff... It's nice if you're only working in one particular ecosystem, because you can use npm audit or bundle audit. But as soon as you get into the multi ecosystem things, which every Docker container is, right...? It's gonna be like "Oh, I've got my Django app with a JavaScript frontend, and also all of the backend low-level distro package stuff... Like, there's a big collection of random bits of software in there, and I really don't wanna have to use 10 different tools to enrich it. I just want one thing that will just sweep across and support everything."
+
+**Adam Stacoviak:** You mentioned a couple of times building things on top of. Since this is sort of a redo for you, it's kind of like a take two, do it better... Is this the substrate for many things? And what are some of those things that you mentioned? You mentioned some things being built on top of, but what are those things? What's the world you envision?
+
+**Andrew Nesbitt:** There's a few that are listed on the Ecosyste.ms homepage. So we have the things that I've built for Open Collective, which are the Funds app and the Dashboards app. Those two things are like definitely -- they don't have their own data, they're essentially aggregations of various bits from Ecosyste.ms to solve particular challenges. One thing I've not built is a search engine. I've kind of been like "I'd like to see if someone else would build that." I already did that in libraries.io... But that would be a natural one to add in there.
+
+What I'd really like to build is things that help maintainers understand who is using their software. And this is going back to that 24 billion rows of dependency data, to be able to say like --
+
+**Adam Stacoviak:** How much are bots, how much is Docker pulls, how much is just like CI builds...? I guess those are all still users, right? I mean, if I'm that person releasing a hundred times, I'm still pulling the packages, right?
+
+**Jerod Santo:** Everytime they commit...
+
+**Adam Stacoviak:** Yeah. It's like "Boom. New version. Because I can", you know?
+
+**Andrew Nesbitt:** Yeah, yeah. And also, to be able to go - like, if we can flip that graph upside down and show you "Here are the key people downstream dependencies of your library", then rather than you find out that you broke them because they come into your issue tracker after you just published that release, and say like you broke stuff, like maybe building a CI that is an inverse, that goes "Okay, well, you committed something. Let me go and test this against your most popular downstream users, to make sure that you didn't break those things." And there's some difficult bits there in making sure those downstream CIs are reliable, and they're not just gonna be like "Oh, actually, our tests pass all the time, regardless." Or they fail all the time, so you can't trust if you actually broke anything or not.
+
+\[01:18:23.06\] But to be able to do that would give maintainers insights that would be like -- they can actually be proactive about some of these things, and maybe even be able to coordinate and go like "Oh, I'm able to reach out to these projects and say "I'm gonna break this thing, or I'm gonna change this thing to make it better. Can I help you upgrade in the process, rather than just firing out into the world and then not being able to know what the impact was until after the fact?"
+
+I've also been indexing Dependabot data as a way of being able to show -- I've no idea why GitHub hasn't done this, but as a maintainer of a thing, if I publish a new version, I wanna know how many Dependabot PRs actually were successfully merged, or were closed, as like "No, I don't want this, because it broke my CI." Or just completely left. Like, give me more context, so that I can understand what's happening with the people that are using my stuff... At least in the open. Because there's so many open source users now that it's a good proxy through to closed source.
+
+Tools like that, that enable maintainers to do more with the same amount of time that they're putting into the project by being more data-driven, or being able to just have more visibility... Because I think a lot of them are working in the dark a lot of the time, partly because you put the blinkers on and you just focus on getting what you need out of your project... But also because they just have no good idea of where their key consumers of those things are, and the knock-on effects of being able to go "Oh, I'll make a breaking change." That breaks this other library, and that ends up having a significant impact... As well as - you know, if you have a security advisory, to be like "Hey, significant end users of my thing, there's going to be a security update, FYI. Get ready to bump, rather than be like "Oh, we're stuck on this version and now we're going to have to like scramble to try and get it updated." To be able to get a little bit more coordination and collaboration by being data-driven I think would be amazing.
+
+So that's kind of -- that's my slightly bigger picture of what I would like to build on top of it, is to really empower maintainers to have an impact, to make their process better, but also then make their open source software better... Because everyone uses open source software, and so then you make all software better by just improving the base layers of the most critical packages.
+
+**Adam Stacoviak:** For sure.
+
+**Andrew Nesbitt:** It's a pretty big goal, but I think there's enough untapped data there that I think can be really powerfully leveraged to make a good go at improving some of these things.
+
+**Adam Stacoviak:** Could you maybe discuss how that interface manifests? Like, what would you show the maintainer? Where would you begin when it comes to exposing the data? Like, how do I get to know my users, the people using my thing?
+
+**Andrew Nesbitt:** Yeah, so I can imagine you would see "Okay, well, for this particular package" - and maybe I've got lots of packages, but I just drill down to one of them... Then I can see "Here are my top dependents", and top being - there's lots of different ways you can define what a top would be... But we can just use the Ecosyste.ms usage metrics is one thing... "Here are the key projects that are using your stuff", and then which versions they're currently pinned on as well. So they might be just like "Oh, I always pick up the latest version. I've got to Dependabot doing the updates." But maybe there's someone who's really heavily using your stuff, but they're actually pinned to an old, major version. And that's like an insight into "Okay, well, why were they stuck? Maybe I can go and help them upgrade, or I can learn that actually I made the most horrific breaking change ever, and they really, really don't want to upgrade, because it completely causes them too many headaches to do that." And maybe I can consider that in how I then continue to maintain that project going forward.
+
+\[01:22:40.04\] You could also then use that interface to say "Okay, well, can you show me everyone that's on this specific version?" Or "Is 50% of my users stuck on an old version?" Or "Are they stuck on an insecure version?" as well, to be able to go "Well, we had this CVE three months ago, and most people..." Especially thinking about this from like individual packages that depend on me, to be able to see the knock-on effect of like all the users of those packages are my transitive users.
+
+There's a lot of data there, but being able to highlight where those key points are of leverage that are like "These things could be improved here", that would be one way of that kind of being manifested.
+
+The other way you could do it is rather than it be a UI, it's more like a notification system of being like "Oh, did you --" You've got the proactive kind of things of like your dependents have updated to your latest version, or your dependents are having problems with - they tried to upgrade to this thing, and "Here's the context of this Dependabot pull request and the discussion that they had, and they haven't yet merged it." To be able to show you that "Oh, wow. Okay, that's interesting... It's having a problem for them that we didn't even imagine, because we're not using the same database as they are for our testing purposes." Something like that.
+
+And maybe there's an AI element there, once you get to very large amounts of users, that you're like "Actually, this is too many downstream users to reach out to. Maybe I can empower Copilot or Claude through some kind of prompt that is like "I've described the changes in my changelog in a way that helps them upgrade from one version to the next." But there's a lot of people that are very reluctant to take on some of those things, because they can be wildly unreliable sometimes, when you try to do things the same way over and over and over.
+
+**Adam Stacoviak:** It's kind of like telemetry via exhaust, too. You're not like literally tracking your users, you're tracking them through natural usage pattern of the ecosystem of open source... So you're not like ask them to opt into too much telemetry either.
+
+**Andrew Nesbitt:** Yeah. I really try not to be too invasive. I try not to track too much data about the individuals, and instead keep it at the project level... Because for one thing, the projects are all like licensed in a way, that says "Yeah, you can share this, and you can like understand this." The licenses let you do that. Whereas tracking individual people is a much more messy thing to do, because people come and go, and they change their names, and they change their email addresses, and it can be hard to try and pin them down.
+
+But also, most open source projects, they're all volunteers. Trying to pin requirements on an individual is asking a lot of someone who is probably not being -- like, they're just giving away their code. So instead it's like "Oh, well, we'll look at these as if you want to do something to help, then here's data that you can do it", rather than being like "We're going to force to upgrade", like "You must do this."You wouldn't want to use Ecosyste.ms to power a massive wave of automated pull requests, for example, for one thing...
+
+**Jerod Santo:** Right.
+
+**Andrew Nesbitt:** \[01:26:22.00\] GitHub would just shut you down straight away. They're allowed to run Copilot or Dependabot at a large scale, but you wouldn't want -- it would be horrible for maintainers, to just have... You hear Daniel from Curl constantly saying like how many different AI bots, especially if it's incentivized in any kind of way... Then you're going to make a mess. But Ecosyste.ms tries to kind of just watch what's the vibe of these ecosystems going on at the moment, and then you can use that to try and have impact on top of that.
+
+**Jerod Santo:** Have you found any information black holes in your desires for features, or tracking things? I mean, exact amounts of funding is an example, I guess... But anywhere else where you're like "Man, I could build this, but I went looking for the data and there's no data"?
+
+**Andrew Nesbitt:** Ooh... So yeah, the funding one is a big one. The other thing that I'd really like is kind of more data around the non-code contributions... But that's really hard to get, right?
+
+**Jerod Santo:** Yeah...
+
+**Andrew Nesbitt:** Your Discords, and your Slacks are not open enough to be able to really index without -- you know, you need an API key, or you need a ghost user sat in a Discord, collecting everything...
+
+**Jerod Santo:** Right. Now you're getting creepy; you're getting real creepy.
+
+**Andrew Nesbitt:** Yeah, it is way too much. There are tools --
+
+**Adam Stacoviak:** "There's Ecosyste.ms, again, tracking us... Get out of here, Ecosyste.ms...!" \[laughter\]
+
+**Andrew Nesbitt:** I'd start joining all the community Zoom calls with an AI chat log kind of thing... But no, there are tools like that. Bitergia has one that you can configure it to track your own community, and you can feed in mailing lists, and you can feed in your Slack, or your Discord, or similar. But you're kind of doing that per community, or even just a per repository level. Trying to do that at a mass scale is stepping into worlds that I'm not really comfortable in terms of the amount of tracking of stuff. It also is just really, really messy. Open source metadata is messy, but it is like tangibly -- okay, yeah, I can see how I can connect the dots here... Whereas once you get into --
+
+**Jerod Santo:** Right. \[unintelligible 01:28:43.26\] structured.
+
+**Andrew Nesbitt:** ...like unstructured text of discussions of things, you're quickly into like "Right. Well, we're just going to try and have LLMs process everything here", and it's a horrible mess, and it's incredibly expensive. We use no LLM stuff in Ecosyste.ms, because we just don't have any budget for that kind of stuff. The amount of processing to analyze 12 million packages...
+
+**Adam Stacoviak:** Well, you do now... Our friends at AMP have free -- just advertising as I use it... And it's like free docs, essentially. I was just telling Jerod about this on the pod we're releasing on Friday... If you're not using AMPCode for free, at least two hours a day or so...
+
+**Jerod Santo:** Ad-supported.
+
+**Adam Stacoviak:** ...then you're missing out on a little bit of LLM work that you can get for free.
+
+**Jerod Santo:** Ad-supported, that's the way I'm saying it.
+
+**Adam Stacoviak:** Well, yes, sorry. It is ad-supported. So you're getting advertised to, but you know... I think that if you're not using that and you have a use for a couple hours a day at no cost... One of the 17 advertisers they have in the network are supporting your open source, essentially. It's kind of cool.
+
+**Jerod Santo:** What else, Andrew? Anything else we didn't ask you about Ecosyste.ms-wise, or...?
+
+**Andrew Nesbitt:** I mean, we covered a lot...
+
+**Jerod Santo:** Yeah.
+
+**Andrew Nesbitt:** \[01:30:02.17\] I'm trying to think if there's anything... I think I covered most of my thinking of the future things, and that's mostly everything that I'm working on at the moment, is Ecosyste.ms. I haven't got any other side things...
+
+**Jerod Santo:** Octobox is dead, or...?
+
+**Andrew Nesbitt:** Octobox is ticking along... GitHub copied most of the features of Octobox, and then we lost most of the customers...
+
+**Jerod Santo:** \[laughs\]
+
+**Andrew Nesbitt:** I still use it every day, but there's not a lot left there... So it still works, but it doesn't have any AI features, so it's not particularly interesting in terms of that aspect. Yeah, I think that nicely covers most of what I've been working on.
+
+**Jerod Santo:** Awesome. Well, it's really cool stuff. I've always been impressed by your abilities and willingness to just collect all the things, and then organize them, and give them back out for free for people to use for various reasons. It's probably exciting when you see somebody using it in a new way, that maybe you hadn't dreamed of, or wouldn't even care to, but you're like "Oh, that's cool." It shows that you're providing real value to folks, and...
+
+**Andrew Nesbitt:** Yeah, especially with the researchers. People will come to me and say "I'm working on this paper, that's like investigating ways that we can get LLMs to suggest better projects for you to use, or packages", or "We're trying to reduce LLMs coming up with old versions of things. Is there good ways of training it on reducing that --" What do they call it? It's like a data lag, basically; the training lag...
+**Adam Stacoviak:** Drift.
+
+**Andrew Nesbitt:** The drift, that's it. That's an interesting challenge, without resorting, again, to kind of RAG or MCP - are there ways of doing short fine-tunes after the fact, of like "Here are the latest versions of things?" and people are doing some interesting research in that space using big chunks of Ecosyste.ms data.
+
+The other thing I just started noodling on is an open source taxonomy... So to try and define a taxonomy that describes the different facets of what makes an open source project -- you know, what does it do? Who is it for? What technologies does it use? There's about six different facets in about 130 different terms that I put together as like a v1 kind of thing, of going like "If you were to put these packages into a box, or six boxes, which ones would it do?" Rather than just going like "Here's some free text keywords, here's a load of the kind of chunks of things", including the role of the user as well, rather than just thinking about "Oh, it's a frontend React app." It's like, but it's for a -- is it for an end user or is it for a sysadmin, or for a developer? To be able to -- and then what domain is it in as well? It's like really early, but I'm hoping it is another way that can produce some alignment in this open source discovery world... Because I worked at GitHub for a while, on open source discovery, and wasn't really able to make a good dent in it there... But I think there's still a lot of low-hanging fruit in terms of just helping people find the right kind of tools to use, because not many other people have really -- also, there's just not a lot of money in that space. It's a loss leader for most companies; searching for open source is not gonna turn you into... You can't even run a lot of ads against that kind of stuff, because open source developers are the number one user of Adblock. So those ads will disappear pretty quickly. But I'm hoping that that taxonomy will be like "Here's a nice blueprint of ways that you can define your project, and put it into ways that then allow you to kind of go "Okay, well, I've got five dimensions here, but I wanna rotate around one of them. I want a web framework for researchers, but I wanna rotate about the technology. What are my options there?" Or "I'm definitely in this technology space, and looking at this kind of position in the stack, but what options do I have here for different users?"
+
+\[01:34:31.12\] And to be able to kind of like twist the picture a little bit, but in a fairly defined space, rather than in just arbitrary free text... Because again, you just end up in this soup of words, which is like "Yeah, we kind of just get very fluffy." And often, projects just don't have very well-defined ways of finding things. Like, they don't add a description to their GitHub repo, or any keywords or topics, so you just kind of like never find it, unless it's in a generic search engine... Which is then really hard in terms of like "Oh, well, what are my options in this space?" And I made this as just like "Surely someone has made one of these already..." And I found a taxonomy of software in the research space, but I did not find a taxonomy of open source software. So I was like "Okay, I can make a stab at one of these. I've never made a taxonomy before...", but I put it together as a "This should be interesting..." And it's been useful so far, and it started some interesting conversations, but I really need some people with more experience in actually defining taxonomies than I have to give more input and also expand it and cover the problems, because I'm pretty sure there's gonna be loads of problems in it... Because I basically just put it together in a couple of days as like "Okay, I think this should work", but mostly untested.
+
+**Jerod Santo:** Where does that live?
+
+**Andrew Nesbitt:** That is on the Ecosyste.ms GitHub. There's also a really quick webpage I made of taxonomy.ecosyste.ms. It's literally just a few days ago, so it's not anywhere on the website, but it is on the GitHub org as oss-taxonomy. I'll get a link in the show notes.
+
+**Jerod Santo:** Awesome. Yeah, send us that... And anything else you wanna make sure we get into our show notes, so that y'all can just click through and find that and help Andrew figure out this taxonomy, so that we can all start to kind of formulate around it. Categorization is always useful, especially for otherwise gray areas such as these; especially if you're self-defining, it helps you to even flesh out your idea or your project better.
+
+**Adam Stacoviak:** I think this is fertile ground right there, honestly, because you've got so many... I would describe it as like ecosystem explorers. Previous to LLMs being a ubiquitous thing and agents helping you, you may have just stayed in the zone that you're comfortable in, because you're the mere human that cannot think 10x faster. And then you get into this LLM world and you're like "Man, I can actually explore new languages, because it knows it. I know this ,language and I can at least translate my knowledge..." And so now you find yourself exploring Go, or Rust, when you would have normally just stayed in the Ruby world, because maybe that's where you're comfortable. And so when you go into those worlds, you're like "Well, how do people test here? How do people deal with HTTP? How do you deal with security things?" And so you find yourself exploring new worlds; while you know the Ruby world well, you don't know the same kind of projects that would help you in a different lens. So I can think that's going to be useful, honestly.
+
+**Andrew Nesbitt:** Yeah, definitely. There's also kind of the ability to see where are the gaps in a particular space. Where have there not been many people working, or there's only just this one old library. Is there an opportunity to kind of jump in and improve that? Or, as you say, you come into a new ecosystem and you're like "What is the Sidekiq of X?"
+
+**Adam Stacoviak:** Exactly.
+
+**Andrew Nesbitt:** \[01:38:17.03\] And often it's like "Oh, well, actually in Erlang world we don't need Sidekiq, because we have OTP. It's kind of all built in." But to be able to learn what is the alternative to this thing is gonna be an interesting way of challenging that. And maybe also kind of breaking down some of these massive projects into sub pieces as well, to be able to go "Okay, well, you've got something huge, but actually there's lots of individual components here that can be used without you having to take on oh, like, "I've got a massive Apache Airflow install now that does everything. Actually, I really only wanna do a piece of this." But how the hell do you go finding that if like their discovery is just folders full of strangely named projects? That's not particularly helpful necessarily in terms of discovery.
+
+**Adam Stacoviak:** Well, let's close with this. What do you want from the world? You seem to be a pretty quiet guy... There's definitely a blog there, so you're active... I don't know how frequently you podcast. We haven't talked personally in years, at least me personally; maybe you've talked to him at least once, Jerod, without me, in the meantime... But what do you want from the world for this project? What kind of response do you want from coming on the show, or producing all this work?
+
+**Andrew Nesbitt:** Well, I have had my head down for -- basically, since leaving Tidelift, and then COVID happening, I basically just like got my head down and just started plugging away. I also started doing track days in a Subaru BRZ, which is an excellent way to get away from the computer. If you've got interesting cars, track days is brilliant fun. But Ecosyste.ms has kind of been building up and building up, and it's now reached the point where I'm like "I need more people helping, kind of not just contributing to the code, but helping it work out where it should go next." Because I can definitely come up with lots of things I would like to see happen, but I need more input from more people on like "How would you like to have a impact on the open source world through data?" So that's input in like feature requests, or thinking about that from a slightly higher level of like collaborations, ways that Ecosyste.ms can support different efforts, be it like security, searching for projects that are like "Oh, there are ways we can improve this part of an ecosystem."
+
+Collaboration is really what I would like to see more of, and I am starting to do more podcasts and various kinds of -- I started a working group with the Chaos Metrics people around package manager metadata, as trying to share the kind of learnings that I've done in developing Ecosyste.ms, and being able to kind of like map metadata across different ecosystems into standardized ways... But if they're interested in ways of understanding and using data in open source to have impacts, then Ecosyste.ms is literally rearing up right now through the Alpha Omega grant that we just received to be able to bring more people into this space and help them have real impact on/knock on effects of improving open source.
+
+**Adam Stacoviak:** Wow, very cool. I'm glad COVID is over, obviously... I'm glad that you're poking your head out of the hole, little rabbit, and showing the world what you've got. It's kind of cool, I like it.
+
+**Jerod Santo:** Good stuff, Andrew. Thanks for coming on the show again.
+
+**Andrew Nesbitt:** Yeah, thanks so much for having me.

Turso is rewriting SQLite in Rust (Interview)_transcript.txt

@@ -0,0 +1,326 @@
+**Jerod Santo:** So what would happen if you took SQLite, the most widely-deployed database in the entire world, and you rewrote it in Rust, with some new twists and new ideas? We are all going to find out because Glauber Costa and his team at Turso are working on it. It's called Limbo. Glauber, welcome to the Changelog.
+
+**Glauber Costa:** Thank you. It's great to be here.
+
+**Jerod Santo:** It's great to have Adam here, despite flu season hitting hard in the Stacoviak house...
+
+**Adam Stacoviak:** Oh, man...
+
+**Glauber Costa:** It sounds great, though. Adam, you sound great.
+
+**Adam Stacoviak:** Thank you!
+
+**Glauber Costa:** You'd mentioned you're on the tail end of the flu, but you don't sound like it.
+
+**Adam Stacoviak:** I've been trying so hard to sound better... So thank you.
+
+**Jerod Santo:** Where should we start? I think the story goes back quite a ways. I remember LibSQL, which was a couple of years ago now, your guys' effort to create a different open source SQLite. I thought SQLite was open source and public domain. Can you tell that story?
+
+**Glauber Costa:** Oh, happy to do that. And the story - in a sense, it's still the same story, so I think you hit the nail on the head there. It's just really how it started. I don't feel this is a completely different thing, what we're doing now. The story there is that we were using SQLite in our company. Pekka is my co-founder. We founded a company... If you happen to stumble upon it, I'll just mention it here by a name; it's called ChiselStrike. And it was in the API space, we were doing a bunch of things, and we were using SQLite very heavily in that company. And the reasons we were using SQLite were reasons I think that a lot of your audience will understand and relate to. Look, you would just do npm install, the name of the package, and then everything's there. Your database is there, everything is there, everything works, you don't have to install anything... So SQLite for us always had this appeal of a database that is just always working, and then you can build stuff around it, and it's just always there. So there was always something that was very appealing to us. However, we also stumbled upon this thing a little bit about SQLite not being open source. So let me address that.
+
+SQLite is a public domain software, which is technically a difference that doesn't matter. I think only a lawyer would be concerned with what is the difference between public domain and open source. For all intents and purposes, it's the same difference between like the BSD license and the MIT license. It's minor. Nobody cares. I don't care, and I think nobody else does either. But SQLite is, according to their own website, by the way - this is something that they put in big, bold letters in their website - open source. Again, I'm not claiming that they're wrong to say that. It's all good. Open source, but not open contribution.
+
+So when people think of an open source project, there are two things that come to your mind. The first one is the code is open. And SQLite is that. But you also think about -- instinctively, "Well, the code is going to be on GitHub somehow, and then if I find an issue, I can go contribute to this project... And I have some source of ownership to the code. It's not just that I can see the code, I can also modify the code." So that's usually classically one of the tenets of open source. Now, SQLite is not like that. The maintainers of SQLite, the creators of SQLite do not take contributions. This is not something that we're saying, this is something that they're saying.
+
+\[07:57\] And yes, I mean, I think it happened in the past that people managed to contribute to SQLite in very special occasions, but it's not a project that is designed to take contributions from other people. So that is how SQLite is.
+
+We started running into a couple of, hurdles with SQLite. Part of that is that we wanted to deploy our SQLite databases to the edge. We wanted to deploy it in a way that was replicated, we wanted to do things like read replicas, and we wanted to put it to run on call for workers, and all things like that. And we considered this a good thing for our project. And technically, we knew we could do it. And for context here, Pekka and I, we've been working together.
+
+Pekka is my co-founder. We've been working together for 15 years now. We met each other when we were both working on the Linux Kernel. Pekka was one of the maintainers. Each one of us worked in Linux for almost 10 years, the first five without knowing each other; I think we knew off each other, but we never met. And in the last five years, very closely to each other, where we developed a good friendship. After that, we actually joined a startup that also became a database company. We spent eight years doing that. So we felt technically like SQLite, all things considered, is pretty simple, given the scope of things that we used to do. So technically, we can make the changes that we believe we can. And so let's just come up with our own version.
+
+And we saw at the time a lot of people dancing around the subject. That's what we saw. So there were some projects like LightFS that were super-interesting projects. Dqlite... We saw lots of people trying to get SQLite to work as a distributed system. The problem that we saw with all of those projects is that they were all, again, dancing around the subject that SQLite cannot be changed, so we're going to create layers and layers and layers around SQLite to help with that. And for us, it was -- technically, this is actually a pretty simple problem. You just have to change SQLite here, here, and here, and you're good to go.
+
+And I think the last piece of the puzzle, or maybe the two last pieces of the puzzle - number one is that we were very concerned with who's going to want to run our SQLite plus patches, because SQLite is a very trusted project... So you make random modifications to SQLite, and everybody starts distrusting what you're doing. So we knew that - look, the solution to this problem is that we should just create an open contribution version of SQLite, because now it's not like SQLite plus Fink, it's its own project, and then people come to trust this one project. And with that, the diagnosis that we had at the time was like "Hey, SQLite is great code-wise. The thing that is holding it back in our reading is exactly the fact that it cannot take advantage of the dynamism, of the creativity of a truly open source community. So let's create that."
+
+And the last piece of the puzzle is that at the time, we were actually discussing "How do we do this?" And one of the options was maybe we should just rewrite SQLite. And we had a lot of experience in our previous company, Scylla... Scylla was a re-implementation of Apache Cassandra and C++, fully compatible. And maybe we over-indexed in the fears that we had going that route. We said, "Hey, the problem with Scylla is that it took a long time for us to put in the market. SQLite is much smaller, so maybe we can do it." But maybe because we had just come from an experience where we wrote a database, we figured that forking was a better strategy.
+
+So at the time, our decision, after much deliberation was "We're not going to rewrite SQLite. We are going to fork it instead." The main advantage is that you have something tomorrow, because you start from a codebase that is already working. And then we're going to start making changes to that.
+
+\[11:59\] So that was LibSQL. LibSQL was essentially at the time -- and the story of Limbo, if I had to tell the sanitized version in which the fork never existed, it would be the same up until this point, because all of those things, all of that, it's still valid. It's just that at the time, a year and a half ago, we decided that forking was the best alternative. And now - and I'm happy to go into the why, but now we decided to try "Okay, what if we had done that, one of those options that we considered? What if we just rewrote it?" And then when we put it out there, the results were -- I've never seen anything like that before. So that's the story.
+
+**Adam Stacoviak:** I think the public domain-ness of SQLite is unique. I think the not open to contribution is also quite unique. Very valid in the "open source world." And even, they say on this public domain page they have, where they say "Warranty of title." So if you are a company that needed something where you needed to be indemnified against the copyright infringement or anything happening inside this codebase, it gets to be a little bit hairy trying to use SQLite in unique ways beyond. So it kind of makes sense, some of the challenges you had. And then maybe the fork made the most sense, but ultimately, the rewrite seems to make the most sense, because what you want is the attributes of what SQLite offers the world, but not the... And I don't want to say it like a negative thing, because it's great, but it's dominated very well in the marketplace. I mean, it's used on Mars, I believe, right, Jerod? This is now on Mars...
+
+**Glauber Costa:** It has trillions of databases deployed, right.
+
+**Adam Stacoviak:** Yeah. I mean it's not... And Dr. Richard Hipp is one of our friends, we love him. We think he's an amazing person. But I think you're right though that the public domain-ness and the not open to contribution makes it challenging, because that -- and Dr. Richard Hipp even said on this podcast in the past that the test suite is proprietary in the fact that it's not in the open; you can't see it. So making these large-scale changes -- so to build on top of what is SQLite, if that's your desire, becomes just problematic.
+
+**Glauber Costa:** This was one of the big contributors for us to revisit that decision, was exactly the test suite. And again, it's not that we didn't know about it, it's just that we thought things would go one way. But look, I always love to - if I may, to use opportunities like this to clarify something... Because when you're telling a story, when you're saying something, it's easy for people to misunderstand you, so I always want to make this 100% clear. There's nothing wrong with SQLite and the way they manage their community. Sometimes people assume that because we went this direction, we believe that what they're doing is wrong. And I want to clarify that that's not the case. I think the beauty of having different people with different points of view, coming from different backgrounds and doing things differently is exactly that you can experiment with a lot of models. And those models have advantages and disadvantages, and it's fine... So that's not the way we -- we believe, we truly believe, and the goal we have with Limbo now is really to replace SQLite. So we believe that we can something much better if we're able to tap into the creativity and the dynamism of a modern open source community. We believe that. Which does not mean that what they're doing is wrong. It's just the choice that they made, and we're making a different choice. It's okay, right?
+
+**Jerod Santo:** Well, I'm glad you make that explicit, because as an observer, I remember your initial announcement, and maybe it's because of maybe the social baggage of what a fork implies...
+
+**Glauber Costa:** Exactly, yeah.
+
+**Jerod Santo:** ...but my initial impression was - and I don't remember the words that you use anymore; it was two years ago. But I was like - you and your team seemed a little bit upset, or... You know, "We're kind of disappointed with this, and so we're going to fork and do it a different way." And it's hard to do that without ruffling feathers, or without getting -- even if you don't mean to, people think that you're angry, or mad, or whatever, you know?
+
+**Glauber Costa:** \[16:05\] Exactly. And we did spend a lot of time, by the way, crafting that message... But regardless of how much -- there's always some room for misunderstanding, which is why I always appreciate... Now, I also don't want to run from it, and I want to make something clear. We disagree, technically, with that decision, right? It doesn't mean we think it's wrong. There's nothing -- we don't think badly of them. They clearly created something very successful. But the way Pekka and I always run open source communities and the way we always participated in open source communities... For us, this component of being able to tap into the creativity of the community was always something very important.
+
+And you have to put it in context as well. Pekka and I grew up in the Linux Kernel community. And a lot of people say this about SQLite; and with that, I disagree. SQLite, being what it is, has to be kept as this very small thing, because that's the whole beauty of SQLite. And again, I disagree with that, because if you look at Linux, Linux started as this project that would only run on x86 computers etc. And today, it is an operating system that runs everywhere. It runs very well on the server... Pekka and I worked together in memory management for very big box machines, like how to make sure that when you have two terabytes of memory in 2010 - which was a lot of memory back then; it's still a lot of memory today... But those algorithms to manage memory don't take three minutes and hang up the system. It's a problem that doesn't happen in embedded devices, and Linux still runs on embedded devices. So through this big community, you actually can create something that runs well everywhere, and don't necessarily have to make those compromises.
+
+So we don't agree that SQLite has to be protected and has to be maintained by just a single individual to be the success it is... But look, who knows? Maybe I'm wrong, maybe I'm right, but we act based on what we believe. We don't believe there is anything wrong with what they're doing. We just think that it could benefit from a different direction.
+
+**Adam Stacoviak:** But you do believe that you can replace. You said "We believe we can replace SQLite."
+
+**Glauber Costa:** Yes. 100%. And we will. And I want to clarify that now, because with LibSQL it became very clear to us in maybe a month that that was not going to happen. So with LibSQL -- I joked about it today, we had this goal of replacing SQLite for 10 minutes, because it's a dream etc. It's obviously a big prize. It's obviously this big prize etc. As I wrote in my blog post, were we bummed about it? No, because we built a pretty successful business on top of the code, on top of LibSQL, which is today our company, Turso. So for us, it was this exercise of -- if you heard the expression "Shoot for the stars, because even if you miss, you're going to hit the moon..." For us it was like "Yeah, it turns out -- I don't think it's going to be possible for us to replace SQLite with the fork." But that's us hitting for the stars. We didn't hit the stars, we hit the moon. The moon's great. Again, we built a super-great project, we're having fun. LibSQL brought something to SQLite that wasn't there, which was like the serverless, and running this on serverless environments, on the cloud platform, syncing databases... One of the things that we do with Turso is you can sync your databases between different devices, and backups, and all of that. And this is going great. I mean, this is going fantastically for us, and so we were never sad or disappointed that this thing that crossed our mind -- can you imagine how amazing it will be if people just start using our fork instead of SQLite? We were never disappointed at that.
+
+But when we put the rewrite out, it became very clear - and I've been using this analogy from chemistry, that the reaction was right. LibSQL was something that the market was clamoring for. I think what it became very clear seeing the reaction of Limbo now is just that the activation energy is not enough to make the reaction happen.
+
+\[20:15\] And after talking to a lot of people - and there's a whole story of what happened after we announced this... I talked to a lot of the people who came and started contributing to the project... And what it became clear is that they all agreed with us, that we need a better version of SQLite. I mean, it will be fantastic if we have something like SQLite, but that does this, and this. All our original vision. But the fork just wasn't differentiated enough. It just wasn't ambitious enough. It just wasn't bold enough to get people to come and contribute.
+
+Furthermore, it is actually very hard to do, because one of the things you did mention - the SQLite test suit is fully proprietary. So once you start making changes that are very deep into the core, you start hitting a bunch of issues. And some of them -- it's just like yourself saying "I know that the best way to solve this problem is this, but I'm going to solve it in this other way, which is almost as good, but not great, but it's less risky." So you start having this on your mind. So the fork never really unshackles you fully to go pursue this, which is why when we announced the rewrite, what we saw, again, was something I've never seen before. We got 8,000 GitHub stars in a week, essentially. We had what I consider to be for the first time in the history of computer science a Hacker News comment section that was mostly positive, with just a couple, I think maybe one or two comments only that were trying to denigrate us. But the rest very supportive. We had this fantastic, inspiring story of this individual who is in prison at the moment, and he was one of the first people in the United States to be granted access to the internet in prison. He can access Discord, but not X, by the way; that maybe say something about it. But he is now the fourth top contributor to the Limbo project...
+
+**Jerod Santo:** That's amazing.
+
+**Glauber Costa:** ...and many, many other things. I mean, the level of contributions that we started to get... We started getting like people running this on the browser. A week after we announced, it was running on the browser already; just like this, compiled to WebAssembly. All from contributions from third party people.
+
+Again, LibSQL, we did get a lot of contributors, but they were all contributing to the drivers, to the server... Things at the margin. Which - they're important as well for an open source project, but with Limbo, we actually managed to see our dream, the 10-minute dream come true. Like, we're going to build a community of people coming here and writing this database with us. So we revisited -- now it's no longer a 10-minute dream that went away. We truly believe now that once we get the momentum going, we can replace SQLite. And that's the goal.
+
+**Jerod Santo:** Well, I think you hit a great chord as you've given evidence to. I know my guttural response to Limbo was quite a bit more positive than mine was to LibSQL, just by reading the announcement post. I'm like "Okay, this makes a lot of sense." And then I realized later "Oh, it's the same people that tried LibSQL a couple of years back." And it seems like with a rewrite there's just so much more to do. It has its own identity, there's fresh and new ideas... Whereas a fork almost comes from a different place.
+
+**Glauber Costa:** Exactly, yeah.
+
+**Jerod Santo:** So I think you're \[unintelligible 00:23:31.05\] describing it. On the testing front though, I would imagine it's gotta be just as hard, or maybe I would expect it to be harder with a rewrite. What makes it easier with a rewrite than with a fork if you still don't have a test suite?
+
+**Glauber Costa:** \[23:47\] Your intuition is correct by the way, which is one of the things that all things considered we went with the fork. Because at least there's some level of guardrails that come with the fact that at least you know that this code that you're importing was tested by this proprietary test suite --
+
+**Jerod Santo:** It started off right. Yeah.
+
+**Glauber Costa:** Right. So I don't think our intuition is incorrect. But something amazing happened... And by the way, the way I look into the situation, another thing that I want to make clear - I don't actually think that we made a mistake with the fork. Because if we were rewriting back then, there were many decisions that we made today that we would not have made. But now we did, because we had the experience of running the Turso platform, and we learned a lot of things. So I almost see it as a prototype, and it actually allowed us to make tremendously good decisions with the Limbo project. And one of them is exactly the testing. What happened is that Pekka and I got hooked into something called deterministic simulation testing
+
+And deterministic simulation testing is a very niche thing, that most people have never heard about. I had heard about it a couple of times, but I didn't truly quite understand what it was, until I met Joran. So Joran is the founder of -- and I can say this... I mean, even though I am the founder of a database company, Joran is the founder of the most amazing database company I've ever seen... Although now I think with Limbo we have a chance to reclaim that title. Joran and I are going to fight about it. But ask me a month ago, I would say "Man, I have to give it to him. This is the most impressive database company of all." TigerBeetle.
+
+TigerBeetle is a database - just to keep it very brief - that is designed for financial transactions. So it's completely written in Zig, and their goal is to replace entire systems, the bespoke systems that most financial institutions have to process transactions. So if you're selling to banks and financial institutions, if your goal is to replace the backbone of world's commerce, it's pretty hard to do. And one of the challenges that you have is just this, "How do we trust this thing?" And Joran found a solution for that. The solution that Joran found was to write TigerBeetle entirely with a technique called deterministic simulation testing.
+
+Deterministic simulation testing is essentially a fuzzer. So it will generate a bunch of inputs. It randomly generates as many inputs as it can on the space of possible inputs... And this is the disadvantage. You kind of have to write software in a special way to lend itself to deterministic simulation testing. It's very hard to bolt it to an existing codebase. But you write it in a way that every single operation that you do on IO, on thread scheduling, everything that happens goes through a special interface that abstracts all IO. And then you plug a simulator -- when you're testing this, you plug this into a simulator. And the simulator is included in the codebase.
+
+So what the simulator does is that it's going to explore the space of possibilities, create the most arcane, impossible situations ever, and then when something breaks, it gives you the exact steps deterministically that everything in the system had up until that point, so debugging those problems becomes very easy. So in record time, in record time, TigerBeetle managed to create a system based on deterministic simulation testing, which is their database. And the stories that Joran would tell is "Look, we've found this bug that would only happen if you would call fsync on a disk, and then fsync would return an incorrect result, and at the same time, a packet will come from the network." He would describe the most complicated scenarios... Oh, and the end result of that - the simulator gives you a seed, and you type that seed into the simulator, and now you have every single step that happened to make that happen.
+
+\[27:43\] So this is one of the things that led us to believe, "What if we try the rewrite?" Because we knew that we could now try to rewrite using deterministic simulation testing. And then we also partnered with a company called Antithesis, that offers almost like the integration test analogy for deterministic simulation testing, which is a full system version of that, that will simulate things between machines, it will simulate the network, it will simulate hardware failures, and stuff like that. So whatever bugs our simulator does not catch, usually we just give it to Antithesis. The next day, Antithesis catches the bug.
+
+So we knew that that was the missing part of the puzzle, and then with that, we will be able to actually create something that probably even surpasses the level of testing that SQLite has. But here's the catch. It is easier to do this on the rewrite, because you have to write the system with these in mind from the get-go. It's not something that is easy to bolt upon.
+
+**Jerod Santo:** Is that popular enough that you can go out and there's Rust crates that will give you this functionality? You have to write all this DST stuff yourself?
+
+**Glauber Costa:** All of it, yeah. So for people who are interested in DST, but don't want to go through the pains of just rewriting all sorts of things like that, I truly recommend taking a look at Antithesis, because... I mean, Antithesis is amazing. It's the second -- it's not the second, next best thing. We use it in conjunction. It's not an either/or. The analogy that I have is that our own DST is like unit tests; we can run centuries of possibilities in two days. And it's very fast, all things considered. And Antithesis is like integration tests. So you want to have both, and we do have both, and Antithesis has been a great partner for us.
+
+But the problem with deterministic simulation testing is that once you start importing other crates in Rust, for example, you have no idea what those crates are doing. Those crates are calling IO. So you kind of have to write everything. So it's not even that you have a crate for deterministic simulation testing, it's that you might not even want to import something super-simple, 100%, but you don't want to import anything that does IO, that has a timer... You want to write those things.
+
+So it's something very hard to do for general purpose testing. For a database, like TigerBeetle, for a database like Limbo, it's worth it. And the scope is quite limited. And you don't want to be importing crates all the time anyway... TigerBeetle is way more insane than we are. They have a policy that they just don't have dependencies. That's it. They write every single piece of code. We try to be a little bit more flexible. But we will not import a crate.
+
+And again, it also depends, because SQLite comes with a CLI. If you want to import a crate that does whatever crazy stuff to implement callers in the CLI, that's fine. But for the core of the database, we try not to import anything that could potentially do IO, because we want to make sure that everything goes through the simulator.
+
+**Break**: \[30:52\]
+
+**Adam Stacoviak:** This DST sounds like magic magic.
+
+**Glauber Costa:** Yes, man.
+
+**Jerod Santo:** \[laughs\] It does.
+
+**Adam Stacoviak:** If it's like writing unit tests, what exactly -- and you said that the rewrite is better because you have to write... Assuming this DST is in place, what actually is the writing and the creating of this deterministic simulation testing? What is that? How do you do it?
+
+**Glauber Costa:** Yeah, so again, the simulator itself, you just write a couple of scenarios. And again, it's not truly magic, because you do have to write the simulator, right?
+
+**Jerod Santo:** Sure.
+
+**Glauber Costa:** So the simulator just writes -- imagine for example, a workload, and the workload is like generate a couple of queries, and then write etc. But the simulator will include a fuzzing element to that. So instead of like generate this query that I wrote in the unit test, you give it almost like a query generator, and then it starts generating random queries. And then in the simulator itself, what happens -- imagine this... If you want to write to a file in any software, you would call a operating system API, like, write. If we talk about Rust, it's like fs write, and then you write to the operating system. When you're writing software for DST, you don't do this. You have your own IO interface abstraction, and then all of your IO goes through this.
+
+So when you're running this in production mode, your abstraction for write just called the operating system write. But when you're running deterministic simulation testing mode, your abstraction for write runs the simulator code that will start injecting failures into this. And again, injecting failures in a deterministic way, because then if the query fails because you injected an IO failure at that moment, you will be able to replay that session piece by piece.
+
+**Adam Stacoviak:** As my friend that runs -- well, he used to be on \[unintelligible 00:34:46.01\] I'll just leave it there. \[laughter\] I'm sure any fans will get it.
+
+**Jerod Santo:** \[laughs\]
+
+**Glauber Costa:** Is that the guy that --
+
+**Adam Stacoviak:** "It's barely an inconvenience", yes.
+
+**Glauber Costa:** "Very easy, barely an inconvenience."
+
+**Adam Stacoviak:** Yeah, that's right. Super-easy, barely an inconvenience.
+
+**Glauber Costa:** I love that guy, man. It's super-easy, barely an inconvenience.
+
+**Adam Stacoviak:** Wow, wow, wow. That's what I'll say right now. That's how he'd respond.
+
+**Glauber Costa:** So how do you rewrite SQLite with the confidence that it will actually have the level of trust that SQLite acquired? It's super-easy, barely an inconvenience, right? All you have to do is write the deterministic --
+
+**Adam Stacoviak:** I love it.
+
+**Jerod Santo:** Just throw a DST at it. So when you talk about SQLite-compatible, that's Limbo's goal. There's a lot of different fronts that that has to be on; are you talking file structure, are you talking syntax language? Are there performance compatibilities? There's all kinds of things that SQLite is. What's Limbo's goals with regard to these different areas of compatibility?
+
+**Glauber Costa:** First of all, compatibility in a project like this - and this is the experience that we gained at Scylla, when we were rewriting Apache Cassandra. Compatibility has to be like a one-way street, right? You don't want to shackle yourself and say "I will always..." Because sometimes, for example, to implement a new feature, you have to do it in a different way. So it's usually something, "Hey, we're going to offer you the same feature set as SQLite. We're going to read your SQLite files, we're going to execute your SQLite code... And if you're not using any of the specific features that we have, we can generate SQLite files as well." But the moment you start using new features that are only present in your implementation, it's impossible to be compatible. It's just not possible.
+
+\[36:26\] Limbo so far - and this is something that we want to keep - the language is the same. Again, we want to keep the language, we want to keep the API, we want to be able to load SQLite extensions. The file format, obviously - it's the thing that defines SQLite, so obviously... Again, we might add to it in the future, but we will be reading SQLite files normally. And we are bytecode-compatible as well.
+
+So one of the other ways in which we test Limbo is by now out of the simulator just generating random SQL statements, and then seeing that the bytecode that is generated by Limbo is the same bytecode set of instructions that is generated by SQLite. That doesn't catch bugs in the implementation of the bytecode, but at least you know that the query plan is the same etc. So this is yet another way in which we've been testing to make sure that it's up to standards.
+
+**Jerod Santo:** So I mentioned at the top that it also has some ambitious goals. One of those - fully asynchronous IO, which is quite a bit different than SQLite, right? So how do you accomplish that, but maintain this compatibility?
+
+**Glauber Costa:** Yeah, so async is something that a lot of people misunderstand. Async doesn't mean -- for the Rust-minded people in the audience, to be clear, async does not mean that we're going to use an async runtime like Tokyo, or like others. I actually personally wrote an async runtime when I was a Datadog, that's still around, called Glommio. We don't use any of that. And again, part of the reason we don't use any of that is that we want all of that to go through our simulator.
+
+So if you look at the Limbo code, it's just really sync Rust. It's not async Rust. So all that async means is that when you call an operation, if that operation is not ready, it will return to you, instead of blocking. And now I'm getting a little bit too technical, but let me just give you the full context. The SQLite C API has a bunch of functions, but the most important of them is called SQLite Step. And SQLite Step is essentially like take another step, take another step in processing this query for me. And this is a blocking function. So what that means is that if you call SQLite Step, it will block until it resolves -- let's say in that step you want to execute four, five, ten, however many bytecode instructions. Until it does that, it will block. In Limbo, if you detect that you're not ready to execute those bytecode instructions at that time, you just return something saying "No. Call me back later." So that is essentially what async means.
+
+And then with that, it becomes very easy to plug something like Tokyo on top in Rust, or run it on the browser, or whatever, because you can call those async functions. And if it's not ready, it will not block. That's all that means. But it doesn't mean that we have to use the async runtime internally. In fact, quite the opposite way.
+
+**Jerod Santo:** So what does that unlock for you? Because SQLite historically is sync, but super-fast, because it's right there, in process, and stuff. But...
+
+**Glauber Costa:** Well, it isn't that fast. I ran a query yesterday, for example, that took 10 minutes to run. That was a query for five joins, looking at all the user data about users who are signed up to the Turso platform. It takes a long time, just because the query is complicated. SQLite is very fast for CRUD style. Just like "Hey, here's the key. Give me the value."
+
+**Jerod Santo:** \[39:59\] Sure.
+
+**Glauber Costa:** But it's not super-fast for analytical stuff. So the first thing that it unlocks is just this, queries that are much more complex, that you can run in dashboards, and things like that. And also, for the serverless environments, Turso being one example... And for clarification, I'm going to say - our product will be renamed to Turso Cloud, and we have all the intention in the world to rename Limbo to just Turso. Limbo is just not a great name. Just a parenthesis here. The story, it's all in the blog posts, but we never expected this level of support, so we just came up with a made up name for it, that means nothing. But now we will address that.
+
+But our product, Turso Cloud, allows you to do serverless SQLite on the cloud. You have a bunch of HTTP requests there in the middle. It allows you, for example, to host your data on S3. Because if you have your data hosted on S3, your query now is not necessarily super-fast, because you may be hitting a page that is not local. So it allows you, for example, to run SQLite with partial storage, like with most of your data on S3, and some of the data locally. It allows us to do all sorts -- it allows us to run on the browser, because the browser, as you know, is a very sync-unfriendly environment. If you block everything, the page just doesn't load. So in the browser you have to be async.
+
+So there's a lot of things that it unlocks. In fact, the one thing that -- Limbo doesn't even support transactions. It's very, very early, which for a lot of people who came to contribute, they actually saw as an advantage, because you come with this energy and there's so much to do... But we're not even yet at the point where we support transactions. And the one thing that is already different and is resonating a lot is the fact that it is async.
+
+**Adam Stacoviak:** Well, you've put a lot of thought behind this; not expecting it to be so well received, you've put so much, it seems like just so much thought in the design of it, from the DST, to async, to all the things... But why were you so surprised it was so well received?
+
+**Glauber Costa:** Technically, we've put a lot of thought in the product. And the full story there is that, first of all, we consider as - just revisiting, especially if people are joining now - we considered rewriting SQLite from the get-go anyway. So this was always on our mind. This was always on our mind. And we added to LibSQL. So one of the things, if you download LibSQL today in the fork, LibSQL comes with vector search out of the box. You don't have to install any extension. You can do your RAG pipelines of SQLite out of the box, with vector search. It was very, very hard to do. It was an ordeal to actually get vector search working. And there are lots of things in the syntax - earlier when I was saying you might do it one way, but then you end up doing this other way, which is more conservative, because vector search was the thing that I had in mind... Like many things in the syntax, they're okay, but they're not great, because we had to be a little bit more conservative. And then Pekka started thinking -- and this is all on Pekka. He started thinking "Okay, how would that look like? How would that look like if we were to rewrite it?" And our goal was something like "Hey, look, if this project keeps going and it goes well, and it's an experimental thing, maybe in two years, in three years we can make something out of it."
+
+So we were very thoughtful about the technical decisions that we were making, but everything around the presentation was just like "Yeah, whatever, man. Whatever." So again, this was not on our company's GitHub, this was on Pekka's personal GitHub... He spent 10 seconds thinking about the name, and then just said "It's going to be Limbo, because it's a state of confusion. We don't know what we're going to make of this." The whole story of the name Limbo was essentially "Pekka, what are we going to make with this?" "I have no idea. I just want to experiment with this concept." The logo - he just asked ChatGPT to generate. We took another five seconds...
+
+\[43:58\] So there was not a lot of thought on the presentation and publishing, but there was a lot of thought about the technical side, because we thought that "Hey, maybe in two years, maybe in three years, there is a play for us once we are a much more established company to tackle rewriting SQLite." What we did not expect is that as we were -- what we decided to do is just make it an official Turso experiment. But if you read the posts in which we were announcing this, we were very clear. "We don't have a roadmap for this. Don't ask us what our intentions are. There is nothing that we want to do with this one in the short term." This was an experiment on Pekka's GitHub that actually did pretty well. And it did pretty well based on two metrics. We got a thousand GitHub stars, and we never talked about it. We never talked about it. And we got two incredibly good engineers that started -- one from Red Hat, who is now with us...
+
+And so we hired those two engineers. We got two incredibly good engineers that started contributing to the project. Again, it was just a project on Pekka's GitHub.
+
+So we knew that there was something there, and we decided "Let's just now publish it as an official Turso experiment." We wrote in the blog post "Maybe, if that goes well, in six months we can start pouring some of our resources in, and then we're gonna see what happens."
+
+So we had some idea that this will be useful at some point, we just did not know it will be so well received, to the point that it was... To the point that we changed our entire company strategy now to be able to pour our whole weight of our resources behind the project. And that's what we're going to do.
+
+**Jerod Santo:** So you're all-in now.
+
+**Glauber Costa:** We're all in. And in fact, we wrote a blog post last week, telling the story, a lot of what we're discussing here. Look, in my wildest dreams, in my wildest dreams, I would expect maybe this to gain like 2,000 more stars in a month or so, then from 1,000 it goes to 3,000... Maybe a couple of other engineers that would come and contribute as well, and start slowly, but we would see some potential on it... That was my definition of success. And every single metric that we thought Limbo could be successful at, we saw three times more, four times more than what we anticipated. So we decided to go all-in.
+
+And there is a blog post that we wrote recently with all the changes that we're going to make to the platform to allow us to do this, with all the changes that we're doing to the company... We had a lot of reorganizations internally. And this is really something that we decided in a couple of weeks in January, because we're just like "How can we ignore this?" I mean it seems very clear to us now that the world at large really wants a evolution of SQLite. The signals are very strong, so I think we just need to get behind it.
+
+**Adam Stacoviak:** What do you think made those two contributors that were so pivotal contribute? If this was an experiment on Pekka's personal GitHub, it wasn't from a technical level well thought out, but just an experiment... What was compelling to them?
+
+**Glauber Costa:** Yes, and again, those two individuals in particular, they are now working for us, and we hired them already last year, because they're great. Pekka and I were like "Man, if Limbo has no other value, it's a great hiring tool for a company, because it's attracting this kind of people." But those are people, of course, that we have direct access to, so we talked to them extensively. And all the others that came as well after. We had 32 contributors to Limbo the day we announced it as a Turso experiment. Days later we had already 60, and we're reaching 70 now. And again, a lot of that, very core contributors, doing a lot of great work.
+
+So \[unintelligible 00:47:45.27\] those two contributors, are obviously the first ones that we talked to. And the story is always the same. The story is always the same, and it's exactly the thing I said. They were very excited about the prospect of a better SQLite. They knew about LibSQL, but LibSQL never enticed them, because it just wasn't - in \[unintelligible 00:48:07.00\] word, it just wasn't crazy enough, which I translate to bold. But it just wasn't it just -- it kind of caps where you can go, because you're still shackled by the fork.
+
+\[48:19\] So when they saw this thing that -- and I think you add this sprinkle to that as well too - this is a new project, so there's so many things to do, which engineers get attracted to to some extent as well. But if you talk to them, again, it's always the same story. "I love the vision, and I already loved the vision when I heard of LibSQL. But this thing here is just the right amount of ambitious that I want to start devoting my time to that, and I want to be a part of that as well." So that's exactly -- and one of the things that we said a lot in the early LibSQL days is that we said that a lot because we heard from a couple of people, a couple of companies, even. "We love the idea of what you're doing, because it gives us a seat at the table." And then my interpretation of that is it turns out it's actually pretty hard to give people a seat at the table if you don't own the table. So limbo gave us this -- I mean, it's our table now. Now, it's a table that is modeled after another table, but we own the table, and now we can really, truly give those people a seat at the table. I mean, they're coming, they're writing code, they're contributing, they're reviewing code, they're helping with the direction, they come up with ideas, things that we never considered, things that would not be a priority. Limbo already limbo runs on the browser. Somebody showed up today, "Oh, I got LibSQL to run the limbo CLI in WebAssembly, and now I'm distributing the CLI." It's not even something we would have done, but the person took an interest because it doesn't have a ceiling. You essentially can take this anywhere.
+
+**Jerod Santo:** So does LibSQL just go away then, or be replaced outright?
+
+**Glauber Costa:** LibSQL came to be a little bit -- and a part of that because, again, as I said, after... We had this dream of replacing SQLite for 10 minutes, and then it became clear that it wouldn't happen... So we had this something that was very valuable. We knew it was valuable, but we had to struggle a little bit more with what do we do with that. So the LibSQL project today, again, is two things. It's a fork of SQLite, with some of those changes, like vector search, but it is also -- and it's all in the same repository. It's all part of the same project. It is also a server implementation to do serverless SQLite over the wire. So that is what LibSQL is.
+
+**Jerod Santo:** Okay.
+
+**Glauber Costa:** It's an open source project, so we're still going to be maintaining that, but the client-side part is going to be full Limbo.
+
+We intend to eventually get all of those things redone, like vector search, all of that. LibSQL is still maintained for the time being, because again, we have a business around it, we have features that we depend on, but our goal is to eventually port all of those features to Limbo, which again, we plan to rename to Turso. And then the client is that, and LibSQL can become just the server implementation, as an open source alternative to do that.
+
+**Break**: \[51:13\]
+
+**Jerod Santo:** Is there trailblazing left to do? Or have you done the trailblazing and now it's just a matter of work?
+
+**Glauber Costa:** Oh, there's so much... Because one of the things that we heard as we launched LibSQL - for some reason, I'll never forget that individual. So we create this Discord community, we announce LibSQL, people came, and one person in particular, but others said the same thing as well... "Give me better schema changes and I'm switching to this tomorrow."
+
+So one of the areas, for example, the SQLite is really bad at is schema changes. You can make schema changes, but you cannot alter the type of a column... There's all sorts of limitations like that. Which is something, by the way, that we made better in LibSQL. So LibSQL does better schema changes than SQLite, but it's not better enough. Same story, because we were limited in what we can do.
+
+So we want to run replication, like native replication to the browser, which is something people have been asking us for a long time. Imagine you have this SQLite database running on your browser, that can then sync with an external server \[unintelligible 00:54:41.06\] or whatever, and just get pages on demand. That's one of the things we want to do. We want to tackle the problem of schema changes, we want to tackle the problem of write throughput, because SQLite is a very bad database for write operations... We want to make Turso much better for analytical workloads... So there's just so much that we believe we can do. But together with all of that, there is also the boring work of - you know, we don't even support transactions yet. And again, some of that you can do in parallel, like async, like browser, but some of them it's just time in the saddle.
+
+**Jerod Santo:** How much time, do you think? Are we talking months, years?
+
+**Glauber Costa:** Nine months to a year.
+
+**Jerod Santo:** Nine months to a year. Okay.
+
+**Glauber Costa:** For two reasons. First of all, because of the deterministic simulation testing. Because it allows you to move with a lot more confidence. Imagine trying to get into the thought process of a database writer. It's always like "I want to make this change, but I don't fully understand the impact that this change will have in all of those environments that I don't control, like running on an embedded device here, running in this other..." So making changes to systems like that... And those are the kind of systems that we dealt with at Linux.
+
+Sometimes Linux has a lot of weird stuff in the process of development around the idea of like "We know that this thing can break on a processor that only three people have in the world, and it's not supported anymore, and we don't want to break those things." So we always move very, very, very, very carefully. And the deterministic simulation testing just allows it to make a lot faster; paired with Antithesis, which is our integration testing, in this analogy, it just allows you to move so much faster.
+
+The second reason for why we believe a year is the reasonable timeframe here is that SQLite, as it turns out - and that's why one of the reasons we believe in forking, and I mean rewriting - is not the biggest codebase in the world. SQLite is not that complex. If we were rewriting Postgres, it's a completely different story. But rewriting SQLite - it's actually doable in a year. And also, as I said, we're going to put the whole -- we're not going to do this in Q1, because we're treating our first quarter of the year as a transition quarter, to allow us to finish other work, so that we can be free to do this. But we're going to have like at least seven, eight engineers just working on it full-time. So imagine that you have seven people working on a codebase that is less than 200,000 lines big, for nine months, with a deterministic simulator that catches all possible bugs that you can muster. We believe it's a very reasonable timeline, nine months to a year.
+
+**Jerod Santo:** So next January we'll be talking about - is it production grade? 1.0? What does it look like?
+
+**Glauber Costa:** We want to release a 1.0 much earlier than that; much earlier than that, because we want to -- we'll see. We still believe in release often, just put it out there.
+
+**Jerod Santo:** Get it out there, yeah.
+
+**Glauber Costa:** A lot of people are going to be early adopters. A lot of people, I think like -- our thesis is that it doesn't even support... Even without transactions, there's a lot of people that with a little bit more support on the read side, they can already use it, because a lot of workloads for SQLite... Like, you get the file; you don't write to the file. You get the SQLite file from somewhere, and then you just run a bunch of stuff on it. So even without writes being fully supported, we believe there's a lot of use cases that this unlocks.
+
+\[58:06\] So we want to be very aggressive with making releases, but the moment we're going to say "Hey this is stable..." And look, we can actually even give us more time, because it's a great commitment to say it's something is stable. I mean, people trust you that this is going to work. So we can take even more time than that, to say "Hey, this now has our seal of approval. We're going to support it", all of that. But much earlier than that, it can already be production-ready.
+
+**Jerod Santo:** How does this all affect Turso the business, and how does it fit in? This is going to -- I assume it's going to be an MIT licensed thing, open source...
+
+**Glauber Costa:** Yeah, so our business is going to change. We announced recently a lot of -- what we need to do is that we need to simplify our product a lot. So we had very hard decisions to make. Again, it was a tough time for us to go through those decisions. But we kept the mission in front of us. We kept this idea, like, we need to make those changes, because I think the community now trusts us to rewrite SQLite, which is like the biggest prize, we believe. So we have to make those changes.
+
+A couple of features that we have, some of them that I love... A lot of users, unfortunately, got quite upset with that, obviously. I mean, features that they use, and came to trust, we will discontinue. So the way we're doing this is that if you're a paid user, after a cutout date, you're going to be allowed to keep using those features. But new users, new signups, anybody else who is not a paid user at that point will no longer have access to those features. With that, we believe our platform will become a lot simpler.
+
+We're still going to maintain our platform, but again, we think that having a single person just running the platform will be enough. And we're not going to be investing in having new features into the platform. So the platform will essentially go -- which is the Turso Cloud. It will essentially be a place where you can still run -- like, if what we do today is good for you, great. And what we do today is two things very well, especially after those features will be discontinued. You can access SQLite over the wire, from serverless environments. So you have a serverless managed SQLite database. And we have our features, point in time restore for backups, branching... We have all of that that a serverless database needs. And you have syncing of databases between devices, servers etc. So you can start with your SQLite file, and you can upload that file to Turso, and then replicate that to other SQLite files that you own. That's the platform.
+
+We're not going to be investing too much in new features, in frontend features, in quality of life features... The platform will essentially go into a freeze for a year. And the platform is pretty good. It's at a position that we can afford to do this. And then for one year, we're going to go all-in in getting Limbo to replace SQLite. And that's it. And then at that point, Limbo then becomes the thing that we run on the platform, and we have the money, we have the runway to do this. It's all accounted for. So if we manage to, which I believe it's very doable, to get Limbo in a production-ready state in 15 months, our plan still works. But 9 to 12 is what we're thinking here.
+
+**Adam Stacoviak:** When you get to that place, whenever Limbo becomes Turso and Turso Cloud goes away, how will you differentiate this new launch?
+
+**Glauber Costa:** No, Turso Cloud doesn't go away. What is Turso today becomes called Turso Cloud. And then Turso, the embedded database just becomes Turso. And again, this is a lesson that we learned, and not that we haven't heard this before, but I think we're just stubborn... People kept telling us, it's very hard to create one brand. Two brands is even harder. But the reason we kept those things separate, and the reason LibSQL wasn't called Turso was exactly because we wanted to give people the sense of "This is a welcoming community. We have our business, but your interests will be heard." Like Linux. We've modeled a lot of that on the experiences we had at Linux. Very few people care about this. And that's the lesson that we learned. Almost nobody -- of all the reasons that we heard for people not contributing to LibSQL, this was never something that showed up. Never, ever, ever, ever.
+
+\[01:02:29.09\] And then when we announced our changes to the Turso platform, we had many questions from many people about "Hey, this feature they're going to discontinue, this other thing they're not going to support", and what happens is nobody asked about the name. Nobody said, "Well, with the name Turso--" Nobody cares. So this idea that we had, that we have to keep things very separate to create a welcoming community, we learned that nobody cares. So we're going to take the opportunity here to just fix it. Because again, it is true what we heard from many of our advisors... Creating a brand is already incredibly hard. Creating two brands is very hard. And the people in the know, they know, but people who are just hearing about it - now you have to explain that this is this, and this is that. So we're going to just consolidate the names.
+
+And Turso today, again, is the cloud offering that is going to be renamed to Turso Cloud. And then what is Limbo, which is the client offering, will just be renamed to Turso, and that's it. And we're still welcoming. We still want people to come in and build this with us. We just learned that the name doesn't matter.
+
+**Jerod Santo:** How do you manage that relationship between the open source Turso, the interests of Turso Cloud, third party contributors who may want Turso to go a different direction, that maybe you don't want it to? Success brings all kinds of challenges.
+
+**Glauber Costa:** It does. And that's why what we decided to do is never port the server code to Limbo. Because then the server can be kept completely separate. And again, I think the reason we had the server and the client together was exactly because it was clear to us after a month or so that we will not, with LibSQL the fork, have achieved our goal of really just replacing SQLite, this goal that we had briefly. So the best strategy then became have the server code here as well, because now there are other things you can do with this open source project.
+
+But with the success of Limbo - again, it's back on the table, and it's a very realistic goal that this will replace SQLite. So we just revisited the strategy and said "This has to be just the client-side library exactly because of this, because we don't want to be in a position--" And the name was a part of that. We just learned that it doesn't matter. But we don't want to ever be in a position -- we want to design things in a way that we are never, ever, ever in a position to even think about "There's a contributor coming with this code, and maybe we don't want to merge it, because this might affect our business." You never want to be in this position. In fact, we want to make people in the community, and maintainers as well, we want to have people, hopefully like Preston, the person that I mentioned that is in prison, and others, that are doing fantastic work. I mean, if they keep working for another six months or a year, and you see that they're committed to the project, they should be maintainers as well. And they should have the ability to just merge code, without talking to us. That's what we want to see.
+
+So the way we're going to structure things to make that happen is that our business now has to be just the cloud. And then all the code for the cloud - that is a separate project; that's going to be its own thing. And then the client... SQLite, for example, they don't have they don't have encryption at REST, they don't have a bunch of things, because they do try to build a lifestyle business around it, as far as I understand, by selling those things. Because they're very specific. We want all of this in the open source for us. We want all of this. We want encryption at REST, we want all sorts of extensions. We want everything that runs on the client to be 100% open source.
+
+**Jerod Santo:** \[01:06:05.06\] Very cool. So how will you know when Limbo/Turso has made it, has arrived, has replaced SQLite? Will it be on Mars? Will it be on trillions of devices? When will you think it's here, it's arrived?
+
+**Glauber Costa:** I think if this was a consumer business is when your mom calls you to say that she's using it, without knowing that this is the software they work. I don't think my mom's going to use Limbo, or Turso, or SQLite, or whatever, so I can't use that. So the criteria -- for us it's really when we can see somehow, through some fuzzy metric, that we have a billion databases out there. So SQLite has a trillion databases, and we don't think we're going to replace SQLite in a year. That's not the goal. But we think that in a year we can get to this point where "Hey, we got our first billion. Like, this is clearly working. If we do this for another two years, and we do this for another three years, we get there."
+So the first milestone that we have in mind is when we can, through some proxy, because a lot of that you cannot measure... But we have a good level of confidence that we have a billion databases in the open. Then I think "Okay, that's success."
+
+**Jerod Santo:** If people wanted to play with it today, how much is there? What could you play with, if anything?
+
+**Glauber Costa:** A lot of the read stuff works. A lot of the read stuff works. And one of the things that Pekka -- this was entirely Pekka's idea, that he did very well, and he was very praised by a lot of people, including those contributors... Some of them said "This is one of the reasons I contributed." It's that he wrote "On day one", because I mean, technically, this was very well thought. We thought about everything very well on the technical level; just on the presentation that we didn't have a lot... But there is a compatibility matrix. So if you go to the repository, there is a file there, compat.md, that is linked in the readme. And there is a full compatibility matrix with everything. So what you're going to see there is like the read stuff from SQLite already works pretty well. So you can, in a lot of ways, if your queries are not the one I run today, with 10 joins, and very complicated use cases - the basics of like reading from a SQLite file, they work already.
+
+**Jerod Santo:** Adam, any other questions for Glauber before we let him go?
+
+**Adam Stacoviak:** I'm just trying to understand this future, I suppose... And I'm trying to piece it together, and I don't know how to ask this question necessarily, but I'm reading your most recent roadmap where you mentioned "Changes to our server implementation." And sorry for my stuttering, because I'm still drastically behind on my breaths because of getting over the flu... But client-side, open source, server-side closed source. And so LibSQL is Turso Cloud. And if you want to self-host...
+
+**Glauber Costa:** More or less, because -- yeah, just to add to that; it wasn't something very relevant, so I didn't bring it up... But we already had -- for other reasons, we already had a proof of concept of a new server implementation, that is not based on LibSQL. And the reason we did this is that we also want our server to have deterministic simulation testing.
+
+**Adam Stacoviak:** Okay.
+
+**Glauber Costa:** So that new server will be closed source. So that's one of the changes in strategy that we're making. And again, we understand that in a perfect world, everything will be open source. I would love that. But the reason we're doing this is that we want to have this very clear separation, without concerns, of what goes where. And we think the best way to do this is if the server is fully proprietary, and the client is fully open source. So LibSQL is open source, the protocols are all the same, so you would be able to run your Turso Cloud databases on LibSQL. That's the goal. But the new server, that runs deterministic simulation testing, that has a lot of additional features, that is designed to be multi-tenant, the design goal that we had is that we will be able to run queries to a billion SQLite databases in a single box. So all of that, that level of scalability, we're going to keep it closed source.
+
+\[01:10:01.14\] So that's going to be the dividing line. And then LibSQL still exists as a reference implementation. If you truly want to run an open source version -- I think it's more accurate to say if you want to run an open source implementation of our cloud platform, LibSQL will do this for you, with limitations. But our new server, we decided not to -- seeing the success of limbo, we decided not to release that as open source, to keep things clear. And then Turso, the embedded database is 100%, 200% if we can, open source.
+
+**Adam Stacoviak:** 200%, gotcha. So in the future, when somebody runs Limbo, which will be renamed Turso, they can run that on their own. But if they want this massive scale multi-tenant world, they've got to host with you.
+
+**Glauber Costa:** Yes, or write their own implementation. Or use LibSQL, again, which is still going to be there, but we believe our implementation is going to be better. That's why we did it.
+
+**Adam Stacoviak:** So you're long betting on hosting a lot of databases in the future.
+
+**Glauber Costa:** Yes, yes.
+
+**Adam Stacoviak:** Okay. You said you don't need money, right? So you have the runway for the next 15 months. You get in danger if --
+
+**Glauber Costa:** No, we have runway for more than that. Everybody needs money.
+
+**Adam Stacoviak:** No need to raise more now though, to get to this goal.
+
+**Glauber Costa:** We don't need to raise more money now, yeah. That's right. For more than two years. I've been talking about 15 months, not as a cutoff point for that. It was more about when do we believe that this can become... But we have capital for a lot longer than that.
+
+**Adam Stacoviak:** Okay, good. Let's talk in a year, I guess. I mean, that's all the questions I have for now, for the most part... Everything else is speculation, but there you go.
+
+**Glauber Costa:** Send me an invitation today, for a year in the future.
+
+**Adam Stacoviak:** January 22nd, 2026...
+
+**Glauber Costa:** Same time, same date... How about that?
+
+**Adam Stacoviak:** Exactly.
+
+**Jerod Santo:** Let's do it.
+
+**Glauber Costa:** Let's do it.
+
+**Jerod Santo:** Alright. Thanks, Glauber. Hey, we're rooting for you. This is really exciting stuff.
+
+**Glauber Costa:** Thank you.

Vite documentary companion pod (Interview)_transcript.txt

@@ -0,0 +1,404 @@
+**Jerod Santo:** Today we have a special treat, it's Evan You from Vite. Welcome, Evan.
+
+**Evan You:** Hello. Nice to be here.
+
+**Jerod Santo:** Happy to have you. Now, as we record this, we're together, but as it ships out to our audience, you are going to be at the first ever in-person ViteConf. That has to be exciting...
+
+**Evan You:** Yeah, we're looking forward to it.
+
+**Jerod Santo:** Vite's come a long way, I think, since you first started it, to a live in-person conference with a CultRepo - formerly Honeypot - classic documentary as well shipping about Vite... You probably didn't have such grand ambitions when you first got started, did you?
+
+**Evan You:** Definitely not. Yeah, Vite really just started as -- similar to Vue, it started as this random idea, "What if I just do this?" And then just started hacking on it, and got something interesting, and then kept pushing on, and eventually, more people started using it.
+
+**Jerod Santo:** What was that idea? This is going back to the WebPack days... I think WebPack was the bundler of the day. And of course, that allowed a lot of things, but it was also complicated, and sometimes frustrating, and lots of configuration... What was the idea you originally had with Vite that was worth hacking on?
+
+**Evan You:** Yeah, so the original idea -- so obviously, I worked on Vue for a long time, and for Vue we built this whole CLI that wraps WebPack, called Vue CLI... And one day, I was just trying to reproduce a Vue bug that a user reported, that came with this whole reproduction that requires running Vue CLI... And it was installing all the dependencies, and getting the reproduction started... And it was taking so long, I was like "There's got to be a better way of doing this." And the idea was really about loading a Vue single file component directly over the network from the dev server without an actual build step. That was really what I was looking for initially.
+
+So the first prototype of Vite was actually back in 2019, it was called vue-dev server, and it was really just a simple HTTP server that looks at the incoming request. If it's a Vue file, it'll just compile the Vue file on the fly, and send it back to the HTTP request. I had that prototype done very early on, played around with the idea... I liked it, because the experience I wanted was really something closer to a normal HTTP, local HTTP server, that I could just start on demand. And then when I import a Vue file, I want the file to be compiled into JavaScript automatically when I request it. Later on, I realized "Okay, this idea actually applies not just to Vue files. If I load a TypeScript file, if I import a TypeScript file, it should come back with the compiled JS, too."
+
+That was also the time when more browsers started to support native ESM imports. So previously, if we were to write something like import foo from a file, we had to first bundle it with WebPack, because browsers couldn't just load native ESM. But in 2019, more browsers started to ship that. And that really got me started thinking "What if we can just use this to build a whole dev environment?" You just import things on demand, it gets compiled, transpiled on demand on the server side, and you get the compiled JavaScript back.
+
+So in 2020 -- the first prototype in 2019 actually got sort of... I dropped it aside because I couldn't figure out how to do hot module replacements with it. So I kept thinking on about it... I was also working on Vue 3 at the time, so that idea kind of stayed in sort of a prototype status for a very long time, until in April 2020 it somehow clicked. I was like "Hey, maybe we can do hot module replacement over native ESM, just by doing this and that." So I started hacking around the idea... I pulled out the old prototype, tried to add the hot module replacement logic on top of it... And it kind of worked. So that got me really excited. And that was the real start of Vite as a real idea.
+
+**Jerod Santo:** There were other contenders at the time, right? There were other people working on this problem. I remember Snowpack...
+
+**Evan You:** \[08:17\] Yeah.
+
+**Jerod Santo:** At some point there was one called Wmr... I'm not sure the timeframe of that one.
+
+**Evan You:** Yeah, I think -- yeah, Wmr was around the same time. Jason Miller worked on that. I think he also kind of worked on it in stealth for a bit of time. He kind of teased about it on Twitter. When it came out, I believe it was modeled after Rollup's API. That's also the source of inspiration for Vite to support Rollup's plugin API as well. And Snowpack also came around at the same time.
+
+The interesting thing with Snowpack is we kind of went from very different directions. Vite started as this sort of "If we can load modules with native ESM and do hot module replacement..." And then we realized, "Okay, if we load every file this way, it will actually be slow." Because if you have a npm dependency that has like hundreds and thousands of modules, it actually gets quite slow. So we were looking for a way to sort of, first of all, combine these individual modules and Node modules into fewer files. And also, some of the Node modules shipped were in CommonJS, which we can't load natively over ESM. So we had to pre-transform them into ESM, so that we can load them in the browser. So we were looking for something that would handle that part.
+
+Snowpack started out as something that does exactly that. The first version of Snowpack was explicitly saying "Let's convert all your CJS dependencies into ESM." And then they later on added the capability to, say, transpile on the fly, and to do a hot module replacement... Essentially, they added on a dev server to that. And we started out as a dev server, but then tried to later on add this dependency compilation part. So we went from like two opposite directions to eventually become very similar in scope. So there was a period where we were pretty much just shipping features left and right. We were kind of doing the same thing in parallel.
+
+**Jerod Santo:** And at a certain point, Vite kind of busted out from the notion, at least the one that I had, and I think others shared, that it was for Vue. Like, if you're not using Vue, Vite's usable, but really it's built for Vue. And that may have never been true, but the reputation was that, or just the association was there.
+
+**Evan You:** Yeah, so the very first version of Vite was definitely for Vue only... But I think we got it to -- I was working on it halfway. So we got all the way up to -- there was a long period of like 0.X releases, and we got all the way up to 1.0 RC.
+
+At that time, the internals were still very specifically designed for Vue. For example, the whole compilation logic of handling Vue single file components was directly inside Vite itself for version one. And it was kind of an afterthought that we added this additional way of transpiling, say, TypeScript. And in version one, we also did that through middlewares. So you need to add your own server middleware that intercepts a custom file format, then transpile it into JavaScript.
+
+And I was then looking-- before we shipped 1.0 I was looking at this whole API design and I realized, "Okay--" That was the point when I also started looking into Wmr a bit more, and I realized, "Look, Wmr is actually having this plugin container, that abstracts away the plugin hosting interface for Rollup..." So you can actually run a Rollup plugin without Rollup itself; not in the browser, but in Wmr. So Wmr is able to run process Rollup plugins, both during the development server and in the production build.
+
+\[12:33\] So we realized, okay, if we do that too in Vite, we can support not just Vue components; we can support TypeScript, we can support TSX, JSX... We can support even Svelte. So that got me started thinking about refactoring the way we handle these custom file formats, and that eventually lead to a complete rewrite. So we never actually shipped 1.0. We were already at 1.0 RC25, or something...
+
+**Jerod Santo:** Really?
+
+**Evan You:** Yeah. I was like "We've got to just completely rethink this." So I started a whole rewrite. That eventually became v2. And that rewrite took another three to four months, I believe.
+
+**Jerod Santo:** So you never shipped one, but then you shipped two?
+
+**Evan You:** Yeah, we just went directly to v2.
+
+**Jerod Santo:** \[laughs\] And how long did you say that the difference was in timing between the two, the RC and the v2?
+
+**Evan You:** I can't remember the exact date we started working on RC, but I first talked about the idea in like April 2020. That was like the first 0.x version of Vite. And Vite 2 was shipped in February 2021.
+
+**Jerod Santo:** Okay. So I remember this part from the documentary... So I was able to see a rough cut of the doc, and Fred Schott, who's the Snowpack author - and I'm sure a friendly competitor at the time; you know, you have the spirit of open source competition, but everyone's kind of rooting for everyone... At least that's the way I play the game. He's telling the story of this December time period where he spent time with his family, and all of a sudden Evan starts committing like crazy. I don't know if that's dramatized for the doc... And of course, they do a great job with these documentaries. They have this close-up view of your contribution graph, and all of a sudden, little green dots start to show up in December, and then into January. And he's like "Evan cracked the nut", or something like this. He figured something out that got you excited to like really start going on it. Is that approximately true?
+
+**Evan You:** Yeah. Actually, looking back at the days, I think the idea of dropping RC and just go directly to 2.0 was like -- it started exactly in December, 2020.
+
+**Jerod Santo:** Oh, okay.
+
+**Evan You:** At the start of December.
+
+**Jerod Santo:** So people were still waiting for 1.0...
+
+**Evan You:** Yeah, people were waiting.
+
+**Jerod Santo:** I mean, extensively. They're like "Man, this is the longest release candidate ever." \[laughter\]
+
+**Evan You:** Yeah. It never landed.
+
+**Jerod Santo:** Yeah. Never landed. And then here comes 2.0. So at that point, did you get the groundswell of adoption and support, once 2.0 hit? Was that when Vite really started to -- because I mean, the adoption story since then has been astronomical, hasn't it?
+
+**Evan You:** Yeah. I think 2.0 really was the real starting point of -- real momentum in terms of adoption. Because people saw -- people knew about Vite 1, before 2.0. Some people were already using it, excited about it, but 2.0 was really the version that made people realize "Okay this is more than just for Vue. It can actually be a proper tool for other frameworks as well." I think that also started -- we started seeing other frameworks actually adopting Vite, and building on top of it. So that really kicked things off.
+
+**Jerod Santo:** \[16:04\] That's a beautiful thing, because there's fractals of frameworks in the JavaScript community, and each one for a long time had their own build tools, their own Webpack thing... Or like Vite did. You guys were like "Well, we're going to build one for Vue." And then eventually, it's nice to see some coagulation and some really rallying around a tool so that everybody can benefit, and helps with that paradox of choice to a certain extent. It's "Yeah, well, everyone's just using Vite now, so let's just all do that." And one less decision to make, right?
+
+**Evan You:** Yeah. I mean, that definitely took some time. I think Svelte was the earliest sort of non-Vue framework to embrace Vite... Rich Harris was working on SvelteKit at that time. SvelteKit was initially built on top of Snowpack. So when we were working on Vite, I think at one point the Svelte team eventually decided to switch from Snowpack to Vite... And then the Snowpack team themselves started working on Astro, and later on they also decided "Okay, we want to focus more on Astro. We want to focus on the framework layer, instead of maintaining our own bundler, our own dev server/bundler/everything that does more or less the same thing as Vite." So Astral eventually switched over from Snowpack to Vite as well. So these are probably two of the earliest frameworks that decide to just build on top of Vite, instead of reinventing the wheels.
+
+**Jerod Santo:** So what was your own trajectory like with regard to Vue.js and Vite? I don't know your history around -- did you leave the Vue project, or are you still --
+
+**Evan You:** We're still working on it.
+
+**Jerod Santo:** Are you still working on the Vue project?
+
+**Evan You:** Yeah. I spend a lot less time directly on Vue myself. We do have other team members still working on features and maintaining it... But the reason we created Vite was because I wanted to use it with Vue. So obviously, we switched all the default tilting of Vue over to Vite.
+
+So the Vue plugin was actually exactly what I used to validate the design of our plugin interface. And later on, we also have Nuxt, which was the main meta framework in the Vue ecosystem. So Nuxt 3 was in development when Vite 2 was introduced. Originally, it was, again, built on top of WebPack. So when they found out that Vite is coming out, I think what they did is they -- I think a smart thing they did with Nuxt 3 in the beginning is they made the core logic of Nuxt 3, the framework, bundler-agnostic. So it can actually work with different builders. So it can work with WebPack, it can also work with Vite. So when they saw Vite was out, they immediately started refactoring things so it can work with Vite. And eventually, the default version of Nuxt 3 was also shipped with Vite as the default build tool.
+
+**Break:** \[19:17\]
+
+**Jerod Santo:** One of the things that I'm very impressed by you, Evan, is your ability to foster communities around your projects. And you've had so many -- I mean, even Vite has so many amazing contributors. Anthony Fu, Matias Capeletto... I don't remember his last name, Patak... Those are just two examples of a lot of people.
+
+**Evan You:** Yeah.
+
+**Jerod Santo:** Is that something you do intentionally? Did that just kind of happen? How do you get these amazing devs to come alongside? And then Vue is maintained and continued being worked on largely by people that aren't yourself. That's an accomplishment. How do you do that?
+
+**Evan You:** To be honest, with the two projects I kind of handled things a bit differently. With Vue, I would say we're really lucky to have -- first of all, for any project to have capable and ambitious contributors stepping up. It's something you kind of -- you're lucky to have any kind of these people, regardless what you do intentionally or not. So first of all, you've got to have these people. If you're lucky enough to have some people like this, it's important to sort of encourage them, acknowledge their efforts, and just make them feel that their contributions are being valued.
+
+In a lot of ways, I think for the Vue project it's very organic. We don't really do any sort of -- we don't do any sort of like trying to ask for contributions, or try to find people explicitly to do these things most of the time. We noticed a pattern. If someone is already very, actively involved, they seem to be passionate about the things they're doing... So what we do is basically acknowledge them and empower them, and make them feel like "Okay, you're now part of this, and your contributions will now impact a lot of people."
+
+Obviously, there's the financial part of it, too. Vue is still independent. We have a decent number of sponsorships, and we redistribute them among the team members. So many of the maintainers who work on Vue now, they are sponsored through our Open Collective, or directly through GitHub Sponsors. So we basically redistribute the sponsorships coming into Vue to people who actually work on Vue. That helps, but again, Vue is a very loosely organized team. So people would come and go.
+
+\[24:00\] We have quite a few people on the team who've been with us for close to a decade. So Edoardo, who maintains the router, and Pinia, our official statement management library - he's been with us like forever. One of the earliest to join, and he's still working on things. And we also have people who've maybe contributed very actively for a year or two, and then they just disappeared... Which is totally fine, because that's how open source works.
+
+**Jerod Santo:** That's how it works, right?
+
+**Evan You:** Yeah. So it's a combination of people who joined, and just stick around, or people who joined and left, but they still left something really valuable to the project. So for Vue, it's very -- I would say it's very unintentional. Like, we're really lucky to have these people.
+
+**Jerod Santo:** It kind of just happened.
+
+**Evan You:** Yeah, it kind of just happened. For Vite, I would say it's a bit different, because I think once it shipped to version two, I was aware that now I have to basically split my attention between Vue and Vite. So my original plan was "I'm going to go back to focus on Vue", because we still hadn't shipped Vue 3 at that point. I basically took this huge detour. People were waiting for Vue 3, but I ended up building Vite, as kind of a -- maybe as a way of procrastination.
+
+**Jerod Santo:** Was anybody mad about that?
+
+**Evan You:** I don't know. I think it turned out to be a net positive. You wait a bit longer for Vue 3, but you end up also having Vite. But the idea was "Okay, I need to go back and focus on Vue 3 now, so I need someone to sort of help me take care of Vite, in a way." So I made an explicit post, I said, "Okay we're setting up a team for Vite. If you're interested, you should join." And that was, I believe, when -- at that time I think Matias was already very actively involved. He started -- I think his first few PRs were mostly documentation. He was just helping us on like adding additional details in the documentation, started digging into the internals a bit... And he was very, very active with the community. He was on the Discord server... He was basically everywhere, answered all the questions. So I was like "Well, this guy seems very enthusiastic about Vite. I should probably encourage him to kind of help us with the long-term maintenance of it." And he was really happy to join, too.
+
+So yeah, so Matias became the very -- I think he was literally the first team member, and also played a very critical role in the initial sort of growing the community and making these additional team members feel they are being welcomed and appreciated.
+
+**Jerod Santo:** Yeah. I've met Matias. He is very good at that. Lots of enthusiasm, lots of care and kindness. It just kind of oozes out of the guy.
+
+**Evan You:** Yeah.
+
+**Jerod Santo:** When you go in to offer him something like that, what is the Vite side? What's the offer? What does he get? Do you get your name on the core team? Do you get this, that, or the other thing? Is there some sort of formalization of what that looks like, or is it just like "Hey, why don't you do more of what you're doing?"
+
+**Evan You:** I think for the initial -- when we set up the team, it was purely voluntary. We were basically saying, "Okay, you get to call yourself a team member, and that's it."
+
+**Jerod Santo:** Right. Which is not nothing.
+
+**Evan You:** \[27:42\] Yeah. But yeah, it was very much like "You're already doing this, so we just want to want you to keep doing this for the recognition of being a team member." We have no plans around how it will turn out in the long run, but I think a very lucky -- also, a very lucky event happened, because StackBlitz, which is now the company behind Bolt, they actually haven't come up with the idea of Bolt back then. That was like in 2022. They were mostly still working on their web containers technology. StackBlitz, the dev environment. So it's an in-browser development environment, and they were looking for a dev setup that just starts off faster than WebPack, and they found Vite. So they were really excited, and they started looking for people who are working on Vite. So they actually hired Matias to work on Vite full-time. So that was what allowed Matias to be able to spend a lot more time on Vite. So that's more three and a half years ago.
+
+**Jerod Santo:** Quick side story on StackBlitz - we had them on JS Party a couple of times, because doing cool stuff in the browser... And the second time, I think, or maybe the third time that one of them came on, it was to talk about a tutorial kit. Tomek Sułkowski came on. And after that show, he's like "Hey, we have this cool thing that we're just about to launch called Bolt. Do you want to try it out? Do you wanna see what it is?" And this was before -- I mean, they had no idea. Because I mean, Bolt just was like a rocket ship; just stratospheric launch.
+
+**Evan You:** Totally.
+
+**Jerod Santo:** But they had no -- I mean, he's just like "This is a little side project we're doing, check it out." And I'm like "This thing's pretty cool." And he's like "Yeah, we're gonna ship it next week. We'll let you know how it goes." And now they've renamed. You know, it's formerly StackBlitz -- I mean, it's still out there, I think, but... Talk about a pivot. That's an amazing \[unintelligible 00:29:42.05\] I'm sure.
+
+**Evan You:** Yeah. They actually debuted Bolt at ViteConf. They also showed me like a week before. And I was like "Well, this looks cool, but I'm not sure how far this is going to go." Because I think initially, the -- at that time, I was still a bit skeptical about the general capabilities of LLMs in general. But everything changed.
+
+**Jerod Santo:** Yeah. That's funny, because when he showed it to me, I thought the same thing. I'm like "This is neat." It seemed like a kind of a cool toy, was my thought. "This seems like a good example of something that the StackBlitz team could do."
+
+But I never expected it to be like a product that would just take over their entire company. So congrats to them, of course... And I'm sure they were pleasantly surprised that it went so well.
+
+**Evan You:** Yeah.
+
+**Jerod Santo:** So what about this documentary? This is the second time you've been documentary-tized. Didn't they do a Vue.js documentary?
+
+**Evan You:** They did. It's actually from the same team.
+
+**Jerod Santo:** Yeah, exact same team. Round two. What is that process like, as like -- an everyday hacker, all of a sudden it's like, put on some nice clothes, going to get interviewed in front of hot lights, and stuff...? Is this something that you enjoy, you put up with because it's good for the projects? How do you approach these things?
+
+**Evan You:** Yeah, again, the two times - it's the same team, but it feels quite different. For the Vue one, the team was largely sponsored by Honeypot. And when they came to us -- when they came to me about the Vue idea, I was like "Sure, sounds great." Because Vue was -- I think the Vue documentary was done... I can't remember exactly which year it was, but I was still living in suburban New Jersey. And they actually managed to fly all the way to New Jersey, come to my house to film it. But the vibe was still quite casual. They filmed me like making coffee, and everything...
+**Jerod Santo:** Yeah.
+
+**Evan You:** And then we I think we met again, when I was speaking at LaraCon in New York, 2019. They went to the LaraCon venue and filmed a bunch of stuff there. Taylor Otwell was actually in that documentary, too.
+
+**Jerod Santo:** Oh, nice.
+
+**Evan You:** \[31:58\] But they were very professional back then. I mean, the director, Josiah, he's amazing. The thing is, they are able to not only make fancy edits, but also, they actually understand the story behind these things. They ask the right questions, and they're able to properly tell the story behind the technology. I think that's the impressive part. So this time when they approached about another documentary -- actually, this time... This is after Honeypot actually folded, so they now have to work on these documentaries by securing their own funding. So they now have a new company. They do need to basically secure sponsorships or sponsors to make these documentaries.
+
+So they approached me, and basically just tried to ask me, "Do you have any ideas? Can you introduce us to some companies or projects that would be looking to make documentaries about their technologies?" And I was like "Hey, we can make one about Vite."
+
+**Jerod Santo:** Gotcha.
+
+**Evan You:** And they were like "Wait, this sounds like a great idea. We should totally do this."
+
+**Jerod Santo:** Yeah. So you and/or VoidZero - you bankrolled it, basically? Or how does it work?
+
+**Evan You:** Yeah, so VoidZero obviously is supporting the documentary. We are the main sponsor, but we also managed to secure a few sponsorships from other companies that are either using Vue or actively involved in the Vite ecosystem. StackBlitz is one of them. Sentry, Shopify, Bolt, and Supabase.
+
+**Jerod Santo:** Nice.
+
+**Evan You:** Yeah. So we have quite a few interesting companies backing the documentary, too.
+
+**Jerod Santo:** Yeah, that's really cool. I didn't know that side of it. I knew that they had renamed. I wasn't sure about the rename, Honeypot to CultRepo. And that's a really cool model, because they do a great job of telling you stories, and making them interesting. That's always the thing; you're like "It's the story of an open source project. It's a bunch of nerds writing, code. How could that be interesting?" But actually, it is interesting, and they do a good job of making it that way. So I think these things need to exist and continue to be made. And so hopefully this works out for them and for you all, to be worth it as an effort.
+
+**Evan You:** Yeah. I think they are working on quite a few additional ones in parallel. I think they did one with Rails, I believe...
+
+**Jerod Santo:** Yup, I think that one's out there. And then we just helped promote the Python documentary as well.
+
+**Evan You:** Yeah, the Python one, the Angular one... They're doing a lot of great stuff, and I think it's definitely needed for the dev community.
+
+**Jerod Santo:** Yeah. It's so cool to be featured as a person who writes software, to have your story told... Especially because it's never the story of one person. Like, maybe it starts with Evan You, but then the story actually weaves in and threads through all these people's lives, you know?
+
+**Evan You:** Totally. Yeah.
+
+**Jerod Santo:** It's really neat.
+
+**Evan You:** Yeah. I believe this is the great part. In a way, every time I see these documentaries starting with a big shot of me, I feel a bit uneasy... \[laughter\] I feel like I don't want these documentaries to be about just one person. In a way, open source is about people, so these documentaries should be about people.
+
+**Break:** \[35:46\]
+
+**Jerod Santo:** So I've been curious about VoidZero, since you made the announcement this is a new direction for you, I suppose... Because prior to Vue, weren't you working at a big tech company? Was it Facebook, o...? Where were you working?
+
+**Evan You:** I was working at Google when I created Vue.
+
+**Jerod Santo:** Right. And then you went indie, and open source...
+
+**Evan You:** Yeah. I left Google in 2014. But I left Google to work at a startup called Meteor, which was also a startup working on open source JavaScript framework.
+
+**Jerod Santo:** Meteor.js?.
+
+**Evan You:** Yeah.
+
+**Jerod Santo:** I didn't know you worked there. Okay...
+
+**Evan You:** Yeah. I spent less than two years. So I left Meteor in 2016. So I officially became fully independent in 2016. That's when I started working on Vue full-time. Which means I've been working independently for like nine years now.
+
+**Jerod Santo:** Almost a decade.
+
+**Evan You:** Yeah. That's a long time.
+
+**Jerod Santo:** And you're kind of a poster child for "This can work." The open source independent thing with sponsorships can work. Now, it doesn't mean it's going to work for everybody... And in fact, for most people it doesn't work in that particular model. But for you, it was working. So you're kind of inspiring to a lot of us, to say "Okay, well, if Evan can do it and he can get his commercial support enough to stay independent and still provide for his family etc. maybe the rest of us can, too."
+
+**Evan You:** Yeah, I've always considered myself lucky to make that work. Because to be honest, when I started doing it 2016, I had no idea if it was going to work.
+
+**Jerod Santo:** Yeah.
+
+**Evan You:** I was like, YOLO... And was I like "I want to just work on this thing full-time, and let's see how long it lasts. Worst case scenario, I'm just going to go back and look for another job." But it was just so fulfilling to be able to work on something that -- you just work on that one thing, you know?
+
+**Jerod Santo:** Yeah. Yeah. It's really cool. That being said, I think we can probably count maybe on two hands - maybe not just one hand, but two hands - the people that have made that work, like you have. It is the exception, it's not the rule.
+
+**Evan You:** It is definitely an outlier kind of thing. In fact, a lot of people, a lot of open source maintainers have came to me over the years, asking for advice, "How can we make this work?" And in a way, my response was always "Look, it's hard to replicate exactly what I did. There's no real formula that you can just follow."
+
+**Jerod Santo:** Yeah.
+
+**Evan You:** It depends a lot on the kind of project you have... Even if your project is like wildly popular, what kind of layer it sits on also kind of determines a lot... Because for a framework like Vue, you get massive exposure and constant interaction with your users.
+
+**Jerod Santo:** Yeah.
+
+**Evan You:** And people also kind of -- it's easier for people to feel more closely associated to a framework that they use every day, compared to something that sits lower level. The more lower level it is, the more difficult for you to get this sort of connection with your users. Like, an example I've always used is Babel, because I was friend with the maintainer of Babel for a while. Henry Zhu, he was the lead maintainer of Babel for a while. And they've always struggled with funding, because even though almost everyone used them, used Babel, because it's just one of the dependencies that kind of, once you set it up, it just sits there... You don't think about it anymore. It was extremely difficult for them to do the same kind of model Vue is doing, because they just don't get enough eyeballs.
+
+So yeah, over the years we've seen a few other examples successfully sort of making this work, but... Yeah. I've also seen so many cases where they're trying to do what we do, but it's very challenging.
+
+**Jerod Santo:** Yeah. Interestingly, I think WebPack was one that was making it work as well, maybe because again, they're right there, in your face. It's a build tool that you use every day. And for a long time -- well, they had a good evangelism as well, and large corporations using them, and so it became kind of cool to support WebPack, because it was an example of like supporting open source back in the early days of that movement. What do you think about a lot of these efforts to somehow calculate out dependencies, and maybe distribute out, redistribute from the top down, like from the Vue.js down to, or from the Vite to the Babels of the world, and kind of do it that way? Because there's so many fundamental aspects of software that are completely invisible to the end user, but they aren't usually invisible to the one who pulled them in originally, and so maybe we can do it that way. I know there's been lots of efforts to do that, I wonder your thoughts on it.
+
+**Evan You:** \[42:18\] Yeah, there is a project called thankyou.dev, I believe. Vite and Vue are actually on there too, but we receive such a small fraction compared to what we receive from, let's say, GitHub Sponsors. And given the amount of Vue's usage in the wild, I can kind of imagine how little maintainers of smaller packages actually get.
+
+I think the intention is obviously good. The idea of consumers open source should try to support the dependencies you're using... In theory, that's what everyone should do. I think the fundamental problem is the donation or sponsoring open source is just not a thing in many of these companies. So if we think about like who are actually getting the biggest value out of open source, it's directly proportional to the scale you're operating at.
+
+An individual developer, if you're just making a sort of side project, you're using open source, you're not even creating value yourself. So asking them to donate to the open source project - maybe they'll like donate one or two bucks a month. In the grand scheme of things, that's negligible. But if you think about a big corporation that's making billions of revenue, they're also using so much open source.
+
+I guess for larger projects like Linux or Kubernetes, they can operate it in a very, very different way. They have foundations, they have bigger companies just backing those foundations... But many other open source projects kind of fall into this category where you're not big enough for these big companies to say "Hey, we need to somehow have a strategic sort of thing to support it." But you're also large enough that you have such a maintenance burden that it's no longer sustainable if you're just one person; or people just like spending after work hours, spare time. Because you could have been spending that time with your family. But now you're somehow obligated to support all your free open source users.
+
+**Jerod Santo:** Right.
+
+**Evan You:** So I think that's the unfortunate sort of state of things for a lot of open source projects, is that they are creating value for these commercial entities/businesses that's actually sort of using their open source projects. I wouldn't say it's -- it's close to freeriding in a way, but it's also because in a lot of ways we start a lot of these open source projects just because we want to share with people, we just want to put the stuff out there. But then people unintentionally find themselves in this awkward spot where your thing kind of blew up. It's making a lot of impact. Now you have a lot more responsibilities. But it's also very difficult to make the companies that are getting the most out of the thing you create to sort of contribute back properly. I think that's a systematic challenge in the way a lot of independent open source projects work nowadays. Obviously, some of them eventually start companies to sort of turn it into a real business. But it's just not a thing that everybody wants to do, and it's also maybe not feasible for a lot of the projects.
+
+**Jerod Santo:** \[45:59\] Yeah. I think the corporate open source support movement was had some steam up until the market correction, when money became tight. And then it was like "Well, now we're doing layoffs, and now we're tightening the belt. We're not going to give money away that we don't have to." And so it's kind of lost steam. But was your then idea with VoidZero? Because probably Vite could just continue like Vue does. Like, you probably just do the support -- I'm not saying Vite is VoidZero, but you yourself and your team have made it in terms of support, I'm assuming, continuous.
+
+**Evan You:** Yeah, I think for Vite - it actually has a similar challenge, because it's a very high visibility project, but it's not as... It's still a little bit different from Vue, in the sense that it's not a framework in itself.
+
+**Jerod Santo:** True. But it has Evan You, and it's a command that you type into your terminal. Like, you're using it actively. It's not hidden from you. And so do you think -- I guess, where was your thinking with VoidZero and Vite? Was it like "We're not going to do the same model that we did last time?" Or did you want a new challenge, or were you bored? I wonder where you were coming from.
+
+**Evan You:** Yeah, it's a combination of multiple factors. So one thing is my experience with the sustainability side of Vue is that the success of Vue is an outlier, and has a lot luck involved. The timing - the fact that Vue was able to become now the second largest framework has a lot to do with doing the right thing at the right time. And for Vite, we noticed that we were getting some sponsorship for Vite as well. But I guess for Vue the model was that I was still doing most of the work. I'm not trying to take all the credit, but over the years, I was still the main person that was driving Vue forward: making the main decisions, shipping the main features, and all that. So I was also making most of the financial decisions for the project. But for Vite, I essentially wanted to "Okay, how about let's make this thing a bit more decentralized, so that more people can actually work on it full time? So it's not me having to be fully responsible for the financial side of it." That was the initial hope. But then we realized, "Okay, this sponsorship model really, really has its limits." Over the years, we are able to call Vue sustainable. But if you think about the amount of users we have, Vue has like 2 million weekly active users based on how many people are using our DevTools extension.
+
+**Jerod Santo:** Right.
+
+**Evan You:** If we were a business and we were only getting the amount of sponsorships we get today, which is like less than a million a year... That's like -- right? If it's a business, it's not a very successful business.
+
+**Jerod Santo:** Right.
+
+**Evan You:** Yeah.
+
+**Jerod Santo:** Successful open source project, not successful business.
+
+**Evan You:** Yeah, in terms of conversion rate. So it's all relative, but... Imagine if you're a business with like 2 million users and you're only making this much. That's not a lot.
+
+So for Vite, what I realized is -- actually, in terms of like sponsorship conversion, Vite is is doing worse than Vue, because it sits a layer lower. It's not a direct framework people interact with. So people have less sort of this idea, "I need to support the tool I'm using", because a lot of people are still using Vite through a framework. For example, they're using SvelteKit, they're using react with Vite, they're using Vue with Vite... So they're still thinking primarily about the framework, instead of this tool.
+
+**Jerod Santo:** \[50:10\] Well, it's perhaps the Vue people also the people who are using Vite with Vue; they may already be supporting view and they think "Well, I'm already supporting these people."
+
+**Evan You:** Yeah. Right. Exactly.
+
+**Jerod Santo:** So that's possibly a part of it as well.
+
+**Evan You:** Yeah. So for me, I was thinking, "Okay, I want to make Vite independently sustainable, aside from Vue. I never want to sort of just mingle/mix these two projects." Because when we decided to make Vite framework agnostic, I was like "Okay, Vite is going to be its own thing. It needs to find its own way, own path to sustainability." And on the other hand, I was also witnessing this trend where Vite is becoming this shared infrastructure between a lot of frameworks. So multiple different frameworks are building on top of Vite. So I saw this opportunity to say "Okay--" Because when we talk about the JavaScript ecosystem, it's always like churn, fragmentation, things come and go. Right? But it's quite rare for multiple frameworks to be betting on the same thing at the same time. And a lot of these people are smart people, that I also really, really respect, like Rich Harris, Fred, you know... So I was like "Look, people are betting on Vite. Vite is serving a much bigger responsibility. So I do foresee it also needs more substantial resources to support its long-term viability."
+
+And then we started thinking about "What can we do to make Vite better?" And that's a lot of work when I think about it. And if we keep going the way Vue is doing, what I really worry about is what if we can't make Vite sustainable enough, and eventually people will just have to switch away from it when it is no longer well maintained... And that's another big churn for the ecosystem, because so many things already depend on it now. And I just don't want that to happen.
+
+So I was thinking, "Can we find a way to a) leverage Vite to serve as this sort of gateway to more unification in the ecosystem, so we can maybe build more fundamental stack from the bottom up, to sort of leverage Vite to unify more things in the ecosystem?" At the same time, if we can make all the work on these parts sustainable through a business, I think that's better than seeing so many people just working independently on a bunch of stuff, supporting tools, and then they burn out, they give up the project, they walk away... And then the next group of people come in and come up with a new alternative, and this cycle just keeps going on for the JS ecosystem. Because I've seen so many of such instances happen in the past...
+
+**Jerod Santo:** Sure.
+
+**Evan You:** ...and I just don't feel like it's -- in a way, what I'm seeing is people keep redoing the work, and then they try to -- somehow, they try to figure out sustainability, and many of them fail. And then they give up, they move on to other things, and now we are stuck with an ecosystem of -- every problem has like five solutions, four of them abandoned, and now you have to try to find new maintainers, or new people, or a new developer will come up with a new project.
+
+I think in the early days of the JavaScript ecosystem, having so many people just stepping up and creating solutions to problems - it's the strength of the ecosystem, and that's what made the JS ecosystem great.
+
+\[54:05\] But over time, a lot of the problems have, in a way -- the ecosystem has matured. We are arriving at more conventions and best practices on some of these problems, that we can more or less consider solved problems. So we want to kind of find a way to provide a very solid solution to these soft problems, and we want make it well-sustained through a good business model... Maybe not a good business model, but at least the business model needs to be good enough so that people can stop worrying about like "Hey, will these people just one day burn out because they can't find enough funding and just walk away?" And I think the key to cracking that problem is to make the right people pay for open source. And that won't happen through donations. That is kind of the conclusion I sort of arrived at after all these years of working on independent open source.
+
+**Jerod Santo:** Yeah. Well, that's a big thing to say, coming from you, because you're probably one of the best donation receivers out there. And if you say that, then it's got to be true. Okay, so VoidZero then... How does it work? What's the business model? You talk about these fundamental pieces underneath Vite, or like where are those lines draw... I have so many -- it's a cloudy thing to me right now. Please help me understand it.
+
+**Evan You:** Yeah. So at VoidZero we're building a unified tool chain for JavaScript. So the scope is quite large, so we would cover -- so obviously, there's Vite, which is at the center of it... It's a dev server, it's also a production bundler... And then we have Vitest, which is the test runner. Then we have the linter, which is based on Oxc, and inside Vite, we were previously relying on ESBuild and Rollup; we rely on two different bundlers. There's a lot of reasons we did that, because of the technical limitations on both sides, and we just kind of have to combine the strength of both. So now we built our own Rust-based bundler called Rolldown, that kind of combines the strength of ESBuild and Rollup and WebPack. In a way, we needed the best bits of those different solutions that we just couldn't get them working together, but now we actually built something that would perfectly suit the needs of Vite itself, while being more powerful. And in order to support Rolldown itself, at the lower level, we have all the parser, transformer, resolver, all written in Rust, in the form of OXC, which is a collection of these lower level tools.
+
+So essentially, we want to have this vertical, unified, integrated stack that supports from the parser all the way up to Vite, and then extend into test, lint, so that you can have a Cargo-like experience. If you've worked with Rust, when you start learning the language, the tool chain you use is Cargo, and it covers most of the things you need out of the box. For JavaScript, we never had that kind of experience. And in a way, we're so used to -- so when people get into web dev, the first thing you need to do is to decide which framework you want to use, and then which build tool, which linter, which test runner... There are a bunch of conventions and best practices, but still, it's a lot. People will feel lost when you just try to set things up for the first time. We want to see if we can provide a sort of more coherent and unified experience to that, so one unified tool chain to do all these things. And at the same time, we want to make sure this thing is well-funded, which means we are planning to do a business model where we want to charge from businesses actually making decent amount of money, instead of individuals.
+
+\[58:30\] So we'll probably discuss more when we announce Vite+ as a product, which - the name of the tool chain is called Vite+, so... We'll share more details at the conference. But the general idea is this is not a thing that can be sustainable if you're only trying to get donations from individuals. We want to make sure that we are still keeping a very low barrier of entry.
+
+Like, if you're open source, if you're an individual, if you're even maybe a small business, or educational, academic entities, or governmental entities, you should be able to just use this thing without any restrictions. But if you are a business that's having a decent amount of revenue, these are the targets that should be actually paying for the majority of the shares to support these efforts. So this is something we're trying to explore. And in a way, it's also my hope of finding a model that can sort of -- because we haven't really seen other projects doing this kind of thing. I think there are maybe a few examples in the wild, but this is something we're trying to explore, to see if we can make the right people pay for the -- essentially, we want the businesses to be subsidizing the wider use of these infrastructure-level pieces in the JavaScript ecosystem.
+
+**Jerod Santo:** And so to start this, did you raise money?
+
+**Evan You:** Yes, we did. Yeah.
+
+**Jerod Santo:** And the reason for that was because you needed time to explore this, or...?
+
+**Evan You:** Yeah, so the scope of this whole vision is obviously much bigger than what we can cover with just the sort of sponsorships or donations, like we did with Vue. If you want to build this whole thing from the ground up, you need a dedicated team of people, full-time efforts. And it also needs to be more structured than the voluntary model that Vue did. Because with Vue, we never had any sort of goal -- like, for Vue, there was no concrete goals of like what the end form of Vue should be. In a way, Vue's development was very much sort of user-driven or community-driven, an a way like we work on Vue, we fix issues, and then we discover problems users are facing, and then we propose solutions to it, and then we just evolve organically.
+
+So Vue doesn't have sort of a fixed scope or a very concrete roadmap of what you need to achieve, say, in five years. But for this, I guess when I started having this idea, I realized "Okay the things you want to do... It's a lot. It's a pretty big scope if you want to build a unified tool chain for JavaScript. And it's something you need a team of full-time members, that's fully dedicated to this goal, and work closely together." And the best way to do that is still to form a business, and have them properly compensated, so that in the short term we don't need to think about "How can we negotiate sponsorships, or how can we get enough individuals to donate to us?"
+
+**Jerod Santo:** Gotcha. And so when you received the money, did you go out and then start hiring members of the Vite community, basically?
+
+**Evan You:** \[01:02:02.28\] Yeah, so we hired quite a few team members from Vitest, from Vite, but we also had a few folks from ByteDance. They were working at the web infra team on ByteDance, so actually a few of the earliest members of the company worked on the RSPack project at ByteDance, so they have a lot of experience and expertise in the Rust4js tooling, which helped us kick off the initial efforts in this area. So the folks working on Rolldown actually worked on RSPack.
+
+**Jerod Santo:** Gotcha. Funny story, I used to call it RSpack because no one ever told me how to pronounce it, and I had to say it on the show, and I'm like "It looks like RSpack." They lowercased the S; it doesn't make any sense. But RSPack makes way more sense than RSpack, for sure.
+
+Vite+ - is it Vite plus other features? Is it Vite, but you pay for it? Is it bundled? Is it Vite is multiple things, and we put them together and package them, and that's what you use? What is Vite+? I know you haven't actually announced the details, but give me a shape of the object.
+
+**Evan You:** Sure. Yeah. So Vite+ is a drop-in superset of Vite.
+
+**Jerod Santo:** A drop-in superset. Okay.
+
+**Evan You:** So if you're using Vite today, you can basically alias Vite to Vite+, and everything will continue to work the same way, but in addition to ViteDev and ViteBuild, you can now also do ViteLinked, Vitest, ViteBench, ViteFormat...
+
+**Jerod Santo:** I see.
+
+**Evan You:** Yeah. So it comes integrated with a lot more capabilities. So it's a tool chain, instead of just a build tool.
+
+**Jerod Santo:** And where and when is Vite+? Is Vite+ out there? Is it going to be out there at ViteConf, or...?
+
+**Evan You:** So we're going to -- yeah, we're going to talk about more details. We're probably going to give a first demo of what Vite+ looks like in action at ViteConf.
+
+**Jerod Santo:** Okay.
+
+**Evan You:** Right now it's still in active development, so we probably want to save the sort of big reveal at the conference.
+
+**Jerod Santo:** Sure. Absolutely. So after the conference, if I'm Jerod, the open source hacker, and I want Vite+, I'll be able to use Vite+ without any money, with money involved...? I'm just Jerod, the dev.
+
+**Evan You:** Jerod the dev, if you're just an individual hacker, you can definitely use Vite+ without any restrictions.
+
+**Jerod Santo:** Okay. Now I'm Jerod, the well-funded Fortune 500 company. I cannot use Vite+ without money, correct?
+
+**Evan You:** So you can still try it, you can still just experiment with it, but if you're using it in any commercial capacity, you'll need to talk to us.
+
+**Jerod Santo:** Right. Are the individual components of Vite+ separately open source, or no?
+
+**Evan You:** Yes, they are.
+
+**Jerod Santo:** So it's the packaged deal that is --
+
+**Evan You:** Yup, yup. So the underlying components, which includes Vite, RollDown, OXC, Vitest, they remain open source, remain exactly what they are today. Vite+ is a combination of these things, plus a few more Vite+ only features. So we are actually working on an integrated task runner/caching system for monorepos, which tightly integrates with all these underlying tools. So that allows us to smartly infer the input for each task that you run through the system. So we smartly cache the outcome of the -- for example, if you run vitebuild through Vite+, we know exactly what the inputs are and outputs are, so we just cache it automatically. So if you edit a file, but it doesn't actually affect the task you're running, you would just get a cached output automatically.
+
+**Jerod Santo:** Okay. And that will be a Vite+ thing. So does that require a VoidZero server somewhere in there?
+
+**Evan You:** It doesn't. Right now it's local.
+
+**Jerod Santo:** Okay. Do you imagine a world in which VoidZero servers are involved?
+
+**Evan You:** \[01:06:15.19\] In the future maybe, but we don't want to make that a requirement. So these will be probably more like value-added services, where it will make your... Like, remote caching with the system definitely will require some sort of server setup, but we'll definitely make it feasible to use your own storage backend.
+
+**Jerod Santo:** Gotcha. So as Jerod the Fortune 500 software team, and I want to use Vite+... Do you have it figured out in terms of my understanding of that whole deal? Where do you interface to me when I know I have to actually come give Evan and his team money? How does that -- have you got that figured out yet?
+
+**Evan You:** So I don't want to get into too much details, but...
+
+**Jerod Santo:** Okay.
+
+**Evan You:** Yeah, yeah. But we'll have a more formally documented "How does this license work?", the FAQ and everything, we'll have that up on the website by the time we announce it.
+
+**Jerod Santo:** Yeah. But Vite and its subcomponents, they all remain open source licensed?
+
+**Evan You:** Yes.
+
+**Jerod Santo:** And so this new Vite+ will have its own proprietary license that you will tell more details about at ViteConf, right?
+
+**Evan You:** Yeah, yeah. So we'll probably call it source-available still. So you'll be able to look at the source code, open issues on GitHub, and everything. But we're trying to come up with a license that would make as many people able to use it freely as possible, but still allow us to capture value from the right group of users that we believe who owe the most to supporting such infrastructure tooling for an entire ecosystem.
+
+**Jerod Santo:** Mm-hm. And if this project succeeds, you believe or you hope that it will be reproducible by other people, right? Whereas the previous one, Evan succeeded with Vue and with that sponsorship support, but it wasn't really a pattern that people could necessarily follow, unless they were also outliers...
+
+**Evan You:** Yup.
+
+**Jerod Santo:** But your hope with VoidZero, it seems, and with Vite this time around, is "What if we can build something in such a way that you could take that pattern and perhaps in your neck of the woods or your little open source area of the world, reproduce it and create a sustainable project"?
+
+**Evan You:** Yeah. If we can sort of be the first example of that, we can also normalize the acceptance of such a model in the corporate world, where... You know, there will probably be some challenges for companies to sort of realize "Okay, this thing is actually using a different license." And if we can make this model sort of acceptable in the corporate world, where next time they see it, they go "Oh, they're doing the same thing that Vite+ is doing..." It makes it easier for other people to replicate this.
+
+**Jerod Santo:** So Vite is open source and Vite+ is source available.
+
+**Evan You:** Yeah.
+
+**Jerod Santo:** Cool. Well, I hope it works. I'm excited to hear more as you announce the details at ViteConf. I appreciate all the efforts you've put in over the years, and I guess the trailblazing that you've done with Vue, and now with Vite; hopefully, VoidZero is yet another trail blazed, and one that people can follow after... And I wish you all the success in the world.
+
+**Evan You:** Thank you.
+
+**Jerod Santo:** Anything I didn't ask you that you'd like to get on the record, or say to the open source world before I let you go?
+
+**Evan You:** Uhm, probably not. I mean, I probably said more than I was planning to, but...
+
+**Jerod Santo:** \[laughs\]
+
+**Evan You:** But yeah, we were planning to announce all of this at ViteConf anyway, so... Yeah. Well, there will be a lot more details on the technical front; how Vite+ feels, how it works and everything at the conference, so make sure to tune in.
+
+**Jerod Santo:** Well, I will have to wait, but nobody else will have to. By the time this is out to our listener and to our viewer, check out ViteConf, check out the Vite documentary, because it's out there, it's to be watched... I've seen a rough cut and it's very good, you're going to enjoy it. So definitely tune in and see what Evan and his team are up to. Alright, Evan, thank you so much. This has been awesome. Talk to you soon.
+
+**Evan You:** Thank you.

Voices of Oxide (Interview)_transcript.txt

@@ -0,0 +1,1260 @@
+**Adam Stacoviak:** So Cliff, what is it you do here, man?
+
+**Cliff Biffle:** I'm responsible for parts of the really low-level firmware on the computer. So everything from the machine turning on, up through fans and power management and all of that. Basically, all the stuff that happens before the thing our customers think of as the computer turns on.
+
+**Adam Stacoviak:** Right.
+
+**Cliff Biffle:** That's all -- well, me and my colleagues now, but originally it was me.
+
+**Adam Stacoviak:** Is that like BIOS stuff, or is it --
+
+**Cliff Biffle:** Well, before then, actually.
+
+**Adam Stacoviak:** Before BIOS.
+
+**Cliff Biffle:** From --
+
+**Jerod Santo:** What's before BIOS?
+
+**Cliff Biffle:** What's, indeed?
+
+**Adam Stacoviak:** Okay...
+
+**Cliff Biffle:** So these big AMD and Intel processors, and even the big ARM nowadays that you see in something like an iPad - there's actually a lot of work that has to go on to allow them to turn on. They need a bunch of different -- like, voltage supply is stable, they need a bunch of clock signals set up and devices set up, and you've got to get their flash ready for them before they wake up, so they can get their code out of it... And so there's almost always one or more small processors in the machine that are responsible for doing all of that dirty work, that nobody ever thinks about, because it happens in the moment between when you hit the power button and when the screen wakes up. So that's our job.
+
+**Adam Stacoviak:** Pretty fast. Like 20 seconds? 15 seconds?
+
+**Cliff Biffle:** Ours is milliseconds.
+
+**Adam Stacoviak:** To me as a user, it's like 15 seconds, you know?
+
+**Cliff Biffle:** Sure.
+
+**Adam Stacoviak:** They mentioned PCBs, so you're building your own boards...
+
+**Cliff Biffle:** Yeah.
+
+**Adam Stacoviak:** So you're going first principles on a lot of this stuff.
+
+**Cliff Biffle:** Yeah, absolutely.
+
+**Adam Stacoviak:** And you're a first principles kind of guy?
+
+**Cliff Biffle:** Yeah.
+
+**Adam Stacoviak:** Do you like that?
+
+**Cliff Biffle:** I think I actually designed the first board we made...
+
+**Adam Stacoviak:** No way.
+
+**Cliff Biffle:** Back before we hired actually qualified electrical engineers.
+
+**Adam Stacoviak:** Okay... It was an MVP, or was it a --
+
+**Cliff Biffle:** Yeah, it was just for testing out some of our circuits we proposed for the bigger, expensive board... Because it's nice to make cheap things that if they're wrong, you can throw them away.
+
+**Adam Stacoviak:** Right.
+
+**Cliff Biffle:** You know.
+
+**Jerod Santo:** And so Oxide has this writing culture... I assume that you were kind of attracted to that, or kind of helped formulate it...
+
+**Cliff Biffle:** No, it was pretty well established when I got here.
+
+**Jerod Santo:** Okay.
+
+**Cliff Biffle:** But the fact that -- we make people write a bunch of stuff during the application process; the materials, and like that whole packet of "Why do you want to work here? What's an interesting problem?" Because it really saves time in the interviews. But for a lot of people, that's daunting. For me, it was like "Oh, you want me to write a bunch of stuff? I like writing. I can do writing."
+
+**Jerod Santo:** Nice.
+
+**Cliff Biffle:** Yeah.
+
+**Adam Stacoviak:** Was that a process you liked? This request for discussion, righ? RFDs... Had you done it before? Is this something that was --
+
+**Cliff Biffle:** Previous companies I'd worked for had a process around design docs that kind of was similar, but not exactly. We are way more writing-focused here, and in particular, most of the docs are living. So as we learn what we're doing, we go back and fix the docs that we wrote when we were done, to try to better reflect... So that way they can also serve as documentation for the next person that comes through. So it's not perfect, but I honestly couldn't really point to what I would change. Like, it's working pretty well, as long as you've got people that are invested in the process, and that are comfortable expressing themselves in writing.
+
+**Adam Stacoviak:** Would you call it document-driven? Do you think you start with docs or an idea, or not so much a spec, but something fleshed out first before you do --
+
+**Cliff Biffle:** I actually feel like one of the ways people sometimes have a hard time starting here is if they treat it too much like that. So it's not like you need to write a thousand paragraphs in English before you can write a line of code. A lot of things here start as a prototype. But then if you want to build consensus, if you want to get other people involved, if you want to try to get feedback, that's when you need to write everything down and share it.
+
+**Adam Stacoviak:** Okay.
+
+**Jerod Santo:** So when you were writing on your blog, this was about Rust, right?
+
+**Cliff Biffle:** Lately, yeah.
+
+**Jerod Santo:** But back then, when they founded it...
+
+**Cliff Biffle:** Since like 2015. It's been mostly Rust.
+
+**Jerod Santo:** Okay. So how'd you find Rust and why'd you like it?
+
+**Cliff Biffle:** So I was working in firmware at Google, doing high-altitude balloon tracking and communication stuff, and we were using C. And I've been using C since I was a kid, because my dad just flew up in it. But it's really hard to produce correct software that doesn't contain bugs, particularly on a team with different experience levels, working in C. And I was bringing people in at intern level, up through experienced developers, and trying to get everybody working together and productive as a team... And the problems with a lot of work in process, you can manage it. Car companies do this all the time, but it takes them a tremendous amount of overhead, and I just wanted some way out of that. So there were a couple of different alternative languages I was watching at the time, and Rust was the one that matured at about the right time, and got enough things right to be worth spending time on.
+
+**Adam Stacoviak:** \[08:16\] This is 2019 you said?
+
+**Cliff Biffle:** This would have been 2015, originally.
+
+**Adam Stacoviak:** 2015, okay.
+
+**Jerod Santo:** Way back. So you said "recently", but that was a decade ago...
+
+**Cliff Biffle:** Yeah. I'm old.
+
+**Jerod Santo:** Okay. I just wanted to point that out. \[laughs\] Alright, so you're writing Rust blogs, so you're writing Rust... Here comes Oxide, you start working on Oxide, you're writing more Rust...
+
+**Cliff Biffle:** Yeah. I think Oxide was an interesting opportunity, because I don't super-care about our product.
+
+**Adam Stacoviak:** Ooh. Tell me more.
+
+**Jerod Santo:** \[laughs\]
+
+**Cliff Biffle:** Like, I care that it's good. I want to make a good product.
+
+**Adam Stacoviak:** Sure.
+
+**Cliff Biffle:** But I'm never going to buy one of these. You're probably not ever going to buy one of these.
+
+**Adam Stacoviak:** I can aspire to buy one...
+
+**Cliff Biffle:** Yeah, you can aspire to buy one.
+
+**Jerod Santo:** We'll pitch a home lab version. We can set that aside for now.
+
+**Cliff Biffle:** Yeah, now we're talking... But it's like a fancy McLaren sports car, or like a Ferrari. I'm probably never going to buy one.
+
+**Jerod Santo:** Right.
+
+**Cliff Biffle:** So this product isn't for me. So I had to find other ways to really get motivated around it, and the main things are this is a team that really wants to try to do things right from the ground up, which I can get behind. That sounds like a hell of a challenge. The team is amazing. My co-workers are amazing. You should talk to more of them.
+
+**Adam Stacoviak:** We will.
+
+**Cliff Biffle:** Good. So you have to look for all the other ways to do this. And sort of the "How do you build an engineering org from a team that fits around a small restaurant table, to this size, and be able to bring in people that don't have relevant experience, and be able to bring in people that maybe had a career change, but are enthusiastic... And building a framework in the software, but also in the processes and the documentation to support expanding the team like that" - that's the thing I got really passionate about.
+
+**Adam Stacoviak:** In the firmware world, is it written in Rust, or is it written in C?
+
+**Cliff Biffle:** Our stuff's all Rust.
+
+**Adam Stacoviak:** All Rust. So there's no -- there's one thing in C, right? Wasn't it like some operating system you had that's not Hubris, that still is C?
+
+**Cliff Biffle:** Yeah. Helios is our sort of version of the Illumos operating system, which is descended from SunOS Solaris. That's mostly C. It's also older, pretty well-tested C. It's not new, potentially buggy C, so we think it's a lower risk.
+
+**Adam Stacoviak:** Does it change much?
+
+**Cliff Biffle:** It doesn't change that much, although we've obviously had to extend it a bunch. But we've been doing a lot of the extending and changing in Rust. But other than that, all the stuff on top of that, all the stuff below that is all Rust.
+
+**Adam Stacoviak:** That's a good thing.
+
+**Cliff Biffle:** Yeah, we think it's good.
+
+**Adam Stacoviak:** I was talking to somebody that said if they had to not write Rust here... Like let's say Go, for example.
+
+**Cliff Biffle:** Alright...
+
+**Adam Stacoviak:** They were like "Nah, nah... Can't do that." How do you feel about that?
+
+**Cliff Biffle:** I don't love Go specifically, but there's other languages that --
+
+**Adam Stacoviak:** Nothing goes bad, but like compared to Rust for some of the things you solve...
+
+**Cliff Biffle:** Not for what I'm doing.
+
+**Adam Stacoviak:** ...Go's not in the picture there.
+
+**Cliff Biffle:** No, it's not. I mean, there really aren't a lot of options on the systems I work on, which are like the 50 cent microcontroller that's inside of your credit card. There's just not a lot of resources.
+
+**Adam Stacoviak:** I never thought about that. A computer in my pocket. What makes Rust uniquely --
+
+**Cliff Biffle:** There's probably a bunch of computers in your pocket right now, actually.
+
+**Adam Stacoviak:** At least four. What makes Rust uniquely positioned for firmware?
+
+**Cliff Biffle:** So Rust --
+
+**Adam Stacoviak:** Or your lower-level things.
+
+**Cliff Biffle:** The thing that Rust took from C and the C family is like - the C family gives you really fine-grained control over what the computer's resources are being used for at any given time. So you have tight control over how much memory is being used. You have tight control over whether memory is used at all, or if you try to solve a problem through some other way.
+
+**Adam Stacoviak:** Or even size.
+
+**Cliff Biffle:** Or size. Size of code, size of flash required...
+
+**Adam Stacoviak:** Strings, numbers, bigger, smaller... Yeah.
+
+**Cliff Biffle:** \[11:51\] ...Rust replicates that control pretty well. Languages like Go are less focused on that, and don't come out of the box with as much help in that area.
+
+**Adam Stacoviak:** Do you mess with Go at all?
+
+**Cliff Biffle:** I have a little bit, yeah.
+
+**Adam Stacoviak:** How do you feel when you do that?
+
+**Cliff Biffle:** It's okay. I mostly just kind of feel like a foreigner. Like, it's not really my native territory. I could get more comfortable with it if I needed to.
+
+**Adam Stacoviak:** And what are you doing with Go when you do play with it?
+
+**Cliff Biffle:** I do some like -- periodically I'll do like... They have like the Advent of Code exercises, or like programming exercises that people put out annually... And I try doing them in other languages, just to kind of keep my brain stretchy.
+
+**Adam Stacoviak:** Yeah.
+
+**Cliff Biffle:** So I've done stuff like that, but I've never used it in anger. Never for anything real.
+
+**Jerod Santo:** Tell us about Hubris.
+
+**Cliff Biffle:** Yeah, so... \[laughs\] So I tried really hard not to write Hubris. When I got here, they were -- there's this other operating system called Tock. It's also in Rust...
+
+**Jerod Santo:** Okay...
+
+**Cliff Biffle:** It targets the same sort of very low-level, deeply embedded --
+
+**Adam Stacoviak:** How do you spell that, Tock?
+
+**Cliff Biffle:** T-O-C-K.
+
+**Adam Stacoviak:** Okay.
+
+**Cliff Biffle:** Like you'd expect.
+
+**Jerod Santo:** Like Tick-Tock.
+
+**Cliff Biffle:** Yeah. But not like TikTok --
+
+**Jerod Santo:** Not the platform. The way clocks work.
+
+**Cliff Biffle:** Yeah. The original Tick-Tock.
+
+**Jerod Santo:** Yeah.
+
+**Cliff Biffle:** They were trying to use TickTock, or --
+
+**Jerod Santo:** Sorry...
+
+**Cliff Biffle:** You fried my brain. So they were trying to use Tock when I got here. And I had some previous experience with Tock,
+
+I know the people that wrote it... So I got in line and tried really hard to make it work for our application, but we just kept hitting areas where their design intent and the things we needed didn't really overlap. Like, Tock is mostly at the time being written with educational use cases in mind, so they wanted kids to be able -- kids... University students who are adults...
+
+**Jerod Santo:** Okay...
+
+**Cliff Biffle:** I'm old. To be able to dynamically reload programs on it as they're working, and nice use case stuff like that... But we really don't want that for security reasons. We want -- any code that runs on this better be what we shipped, when we shipped it. So we put a bunch of work into trying to work around that, and then finally in May of 2020 I think I wrote an RFD that was like "Guys, I think we're going to have to do our own thing. Here's a rough sketch of how it might look like", and there were enough people here that had been involved in operating systems work before, that they all kind of... We pressed our heads together and they said, "Okay, this might work. Like, take a week, and see how much we can prototype." And we got a thing working, and then it seemed compelling enough that now it's... There's... What - 64 to 70 computers running it inside every rack. All the little service processors you don't think about.
+
+**Jerod Santo:** They're all running that.
+
+**Cliff Biffle:** They're all running Hubris.
+
+**Jerod Santo:** So how big or small is Hubris in terms of like line count, or whatever?
+
+**Cliff Biffle:** So the core kernel is like a thousand lines of code...
+
+**Jerod Santo:** That's not very much.
+
+**Cliff Biffle:** ...but there's a bunch of other stuff you want to make it useful. But it runs right now on everything from like sub-50 cent microcontrollers that you wouldn't even spot on a printed circuit board, because they're just a tiny fleck of silicon, up through the big service processor that we use to run the Oxide rack, which is like a... It's basically a computer you would have been really excited to own in 1999 or 2000, but now it's $3, and inside of another chip.
+
+**Jerod Santo:** And does that job.
+
+**Cliff Biffle:** Yeah.
+
+**Adam Stacoviak:** Are there multiple instances of Hubris on a given full rack system?
+
+**Cliff Biffle:** Yeah, so every slide has at least two. There's the service processor that is responsible for basically caring and feeding of the big AMD chip. Then there's a root of trust that handles security and crypto. That's a separate copy. And then there's... So there's those two on every compute sled, and there's two in every switch, there's two in every power supply... And that's everything in the rack. But then a bunch of our manufacturing tools are also running Hubris. So all the little boards we plug into a thing to program it, or to interpose in an interface for testing... It'll be like, you know, I need to remove this fan and simulate the fan controller... All of our tools for that are all Hubris-based.
+
+**Adam Stacoviak:** Why so many copies? Is that hard to manage? It's like...
+
+**Cliff Biffle:** Yes and no.
+
+**Adam Stacoviak:** ...multiple updates, different versions potentially even...
+
+**Cliff Biffle:** That's true. It has pluses and minuses. So the SP versus root of trust split, which is the main source of the many, many copies... There has to be at least one of these on each board...
+
+**Jerod Santo:** Security...
+
+**Cliff Biffle:** ...because the board's got to be able to power itself on. Because it might be the first one powering on and responsible for driving all the other ones.
+
+**Adam Stacoviak:** Yeah, it's a specialized case. That makes sense to be a copy, but...
+
+**Cliff Biffle:** So one makes sense. Why two? And the honest reason why two is we can buy one chip with the features we need to do the service processor. We can buy one chip with the crypto security features we need to do the root of trust. We can't get them both in one chip right now. And we can't afford to make our own chips.
+
+**Adam Stacoviak:** \[16:10\] Okay. So when you can--
+
+**Cliff Biffle:** We might merge those. Or we might not. There's advantages to having a \[unintelligible 00:16:14.11\]
+
+**Adam Stacoviak:** Are you alluding to making chips in the future?
+
+**Cliff Biffle:** I mean, we'll probably have to.
+
+**Adam Stacoviak:** Yeah. Would it be a collab with like AMD, or an existing...
+
+**Cliff Biffle:** AMD returns our phone calls now. It's very exciting.
+
+**Adam Stacoviak:** Nice.
+
+**Jerod Santo:** Is that new?
+
+**Cliff Biffle:** Yeah, it's pretty new.
+
+**Adam Stacoviak:** Like, "Hey, Oxide, we'll take that phone call." That's cool.
+
+**Cliff Biffle:** Yeah, it's nice. Probably not. So we have some FPGAs on the newest generation server board, which are basically...
+
+**Adam Stacoviak:** What's FPGA mean?
+
+**Cliff Biffle:** Yeah, so Field Programmable Gate Array is the full nerd expansion. But the purpose of the chip is it's basically a bunch of... It's a Lego set for integrated circuits. You've got a bunch of generic logic circuits that you can then program to act like another chip... And it's slower and more expensive than the other chip would be, but if you can't afford the million-plus dollars to get started making your own chip, this is like a way to fake it, essentially. We have one of those on the next generation server kind of playing around with some things we would do if we made our own chip.
+
+**Adam Stacoviak:** And Hubris is open source...
+
+**Cliff Biffle:** Yeah.
+
+**Adam Stacoviak:** Does anybody else use Hubris?
+
+**Cliff Biffle:** I've heard from about five other companies that are using it in production.
+
+**Adam Stacoviak:** Really?
+
+**Cliff Biffle:** Yup. And I can't remember which one of them lets me say that publicly. Also, I can't remember them anyway right now... But I could check my notes.
+
+**Adam Stacoviak:** Sure.
+
+**Cliff Biffle:** Volvo is really interested, but we don't have the certifications that they would need as a car company... But I've been talking to somebody about what that would take.
+
+**Adam Stacoviak:** They can contribute though, right? It's open source.
+
+**Cliff Biffle:** They could. The thing they would need to contribute, unfortunately, is a bunch of money for consultants to go through the certification process... Which they're not excited about. So that's fine.
+
+**Adam Stacoviak:** Well, they probably have -- maybe they have more money than you all now? \[unintelligible 00:17:56.03\]
+
+**Cliff Biffle:** I hope Volvo has more money than we do. I hope so.
+
+**Jerod Santo:** I bet they do.
+
+**Adam Stacoviak:** I have no idea.
+
+**Cliff Biffle:** It's Volvo.
+
+**Adam Stacoviak:** Yeah, but car companies - they burn money. Don't they burn money?
+
+**Cliff Biffle:** Yeah... There's like three other startups I've heard from that are using it in products. So that's pretty cool. I'd like to get more people using it, but there's some work we need to do to make it more friendly to people that aren't Oxide.
+
+**Adam Stacoviak:** Right.
+
+**Cliff Biffle:** Because right now, if there's a trade-off we have to make, and one thing would make us ship faster, and the other thing would make it more general for other customers, we almost always have to pick the "We ship faster" option.
+
+**Adam Stacoviak:** Right.
+
+**Jerod Santo:** One of the fun things about being here in the building with everybody is every once in a while a fan just goes crazy. This is not a person who enjoys -- like, a literal fan just starts... Like, what's going on when that happens?
+
+**Cliff Biffle:** That's a great question. So one of two things is going on.
+
+**Jerod Santo:** Okay.
+
+**Cliff Biffle:** The good thing is somebody just ran something on that machine that's boosted all the CPUs. It's like when your laptop starts trying to take off.
+
+**Jerod Santo:** Sure.
+
+**Cliff Biffle:** So something made the CPUs go really fast, everything ramped up, the machine got hot, fan turns on, cools the machine back down. That's sort of the working as intended. What's probably happening here is that something's crashing.
+
+**Jerod Santo:** Something's going wrong.
+
+**Cliff Biffle:** We have a chip on the board that's a hardware watchdog that if it doesn't receive regular instructions on what to do with the fans, it assumes the worst, and it ramps the fans up to make sure we don't overheat.
+
+**Jerod Santo:** To avoid... Okay.
+
+**Cliff Biffle:** This means that if you're doing a firmware update on the service processor, the Hubris-based service processor that's responsible for sending those messages, and it's gone for more than watchdog, however the setting is seconds, then the chip kicks up, ramps the fans up, and then it wakes up and finishes doing its update, and then the fans go back down. So this is why -- I don't know if you saw earlier, but whenever those fans go up, I'm like...
+
+**Jerod Santo:** You're like the watchdog for the watchdog --
+
+**Cliff Biffle:** I get yelled at in chat, but one of the computers is crashing --
+
+**Jerod Santo:** Because you're one of the people that works out of the office generally...
+
+**Cliff Biffle:** Yeah. And I might have wrote the code that's messing up, so...
+
+**Jerod Santo:** So you'd better get out of your seat and go do something?
+
+**Cliff Biffle:** So frequently I'm like "It's not me. It's not me. This one's not me."
+
+**Jerod Santo:** So as being somebody who's regularly in the office, but most folks aren't, what does OxCon do for you in terms of your camaraderie with your colleagues, or the excitement level? How do you feel about it?
+
+**Cliff Biffle:** I get to find out how tall everybody is.
+
+**Jerod Santo:** \[laughs\]
+
+**Cliff Biffle:** \[20:13\] You can't tell that on the computer.
+
+**Jerod Santo:** That's true.
+
+**Cliff Biffle:** Like Aaron... Aaron's like damn near seven feet tall, but he looks normal on the computer. So that's been interesting. Other than that, gosh... I don't think the company would work without this, honestly. I mean, how would you even -- it doesn't even feel like you're at a real company. It's like you're watching a TV show of a company.
+
+**Jerod Santo:** Good point.
+
+**Adam Stacoviak:** Like The Office.
+
+**Cliff Biffle:** Like The Office, or Silicon Valley.
+
+**Adam Stacoviak:** Oh, gosh.
+
+**Jerod Santo:** Ding!
+
+**Adam Stacoviak:** Are you a fan?
+
+**Cliff Biffle:** I worked there, so I couldn't really make it through the show... Like, my boss was personally a parody on the show.
+
+**Jerod Santo:** "I worked there..."
+
+**Cliff Biffle:** And he loved it.
+
+**Jerod Santo:** Did he?
+
+**Cliff Biffle:** Yeah. So Astro Teller at Google X was the inspiration for the Hooli X guy in the show... And he thought it was hilarious. He had a showing of the episode where --
+
+**Adam Stacoviak:** With the monkey guy. The guy that --
+
+**Cliff Biffle:** Yeah.
+
+**Adam Stacoviak:** Okay. Gosh...
+
+**Cliff Biffle:** So... He did have a monkey. The real guy.
+
+**Adam Stacoviak:** Oh, man... That was good.
+
+**Cliff Biffle:** But yeah, so I was watching the show a little bit, and I'm like "Why are you making me watch this? This is my day job. I'm going to watch Game of Thrones, or something."
+
+**Adam Stacoviak:** Yeah, I can understand that. So you've never gotten past season one, or even one episode.
+
+**Cliff Biffle:** No, I made it like three episodes in.
+
+**Adam Stacoviak:** How about now? You feel better about it since you're free of that world?
+
+**Cliff Biffle:** I'm good.
+
+**Adam Stacoviak:** You're in a different world now?
+
+**Cliff Biffle:** I'm good.
+
+**Adam Stacoviak:** Sadness...
+
+**Jerod Santo:** He lived it.
+
+**Cliff Biffle:** Yeah, I'm alright.
+
+**Adam Stacoviak:** I was listening to Brian tell one of his stories and I was like "Oh my gosh, that was literally in the show." Something he described from the stage today, regarding money and funding... And I was like "That was literally copied from it." I mean, it's real life, but it's there. And so I can understand that.
+
+**Cliff Biffle:** Yeah.
+
+**Adam Stacoviak:** For me it's entertainment because I haven't worked for Google, and I haven't done your life... For you, I can imagine how it's PTSD.
+
+**Cliff Biffle:** Yeah. I also don't watch a lot of TV, so I'm kind of picky about what I spend my time on. But... Hopefully we don't do anything to get a show made about us.
+
+**Adam Stacoviak:** Or you do, and it's good.
+
+**Cliff Biffle:** That could happen. That could happen.
+
+**Jerod Santo:** Like Severance.
+
+**Cliff Biffle:** I just started watching that... That's a hell of a show.
+
+**Jerod Santo:** That's worth your time.
+
+**Adam Stacoviak:** One thing I've been thinking about is churn. Is there any churn here at all?
+
+**Cliff Biffle:** We've had people leave... Yeah.
+
+**Adam Stacoviak:** Without being, you know TMI, what are some of the reasons? Have they been negative, or has it just been just mutual separation...?
+
+**Cliff Biffle:** I actually really like basically everybody who's left. For some people, this work environment doesn't work. Honestly, it barely works for me. Like, the all remote thing... I actually took this job because I turned down two other offers that were fully remote, because Oxide at the time wasn't...
+
+**Jerod Santo:** Yeah.
+
+**Cliff Biffle:** And this was in February 2020, so you can guess what happened next.
+
+**Adam Stacoviak:** Yeah.
+
+**Cliff Biffle:** So... Oh well.
+
+**Adam Stacoviak:** Literally next. The very next month, yeah.
+
+**Cliff Biffle:** It's a good thing my doctors are great, but the whole remote thing doesn't work for some people. And sometimes you just can't ever really get in the swing of things.
+
+**Jerod Santo:** Yeah.
+
+**Cliff Biffle:** We've had folks where this turns out to set off other past work trauma... Like, we've all got like work PTSD from some shitty former boss. And if things are happening here that's too much like that, you might get freaked out and decide you need to leave, which I totally respect. We've had folks like Arian - Arian joined before me, and was involved in bringing me over here, and he just left last year because he's like "I've been here for five years. I feel like I've done all the startuppy stuff I can do here. I'm going to go do a new startup."
+
+**Adam Stacoviak:** So just time.
+
+**Cliff Biffle:** Yeah.
+
+**Adam Stacoviak:** So mostly good reasons. Not like "This place sucks. I'm out of here."
+
+**Cliff Biffle:** I think there's -- you could probably find people that think this place sucks.
+
+**Adam Stacoviak:** Yeah?
+
+**Cliff Biffle:** Yeah. I'm not totally sure who, but I'm pretty sure you could.
+
+**Adam Stacoviak:** One person?
+
+**Cliff Biffle:** Probably one or two.
+
+**Adam Stacoviak:** Two people.
+
+**Cliff Biffle:** I don't know.
+
+**Adam Stacoviak:** Out of 80, that's pretty good odds.
+
+**Cliff Biffle:** It's not too bad.
+
+**Jerod Santo:** How do you feel about the uniform compensation stuff?
+
+**Cliff Biffle:** I think it's amazing it's worked this long. It was part of the reason I joined. Because I came in all like knives out, expecting to negotiate...
+
+**Jerod Santo:** Yeah.
+
+**Cliff Biffle:** \[24:02\] ...and Steve's like "So what we're doing is we give people stock according to this formula, and we pay everybody the same amount of dollars." I'm like "Well, that'd save me a lot of stress. Sold."
+
+**Jerod Santo:** Yeah. Probably the opposite of your time at Google.
+
+**Cliff Biffle:** Yeah. A hundred percent. Oh, my God. My only real concern there is that one of the things I really like doing is bringing people in who either don't have a lot of experience in industry, or like are just out of school, or just out of some other job...
+
+**Jerod Santo:** Right. Mentoring them up.
+
+**Cliff Biffle:** Right. And like, are we comfortable bringing in people that are basically interns, and paying them pretty good Bay Area salaries? Maybe we are.
+
+**Jerod Santo:** Yeah.
+
+**Cliff Biffle:** But I do want us to be able to bring those people in, because that's how we get the next generation of us.
+
+**Jerod Santo:** That's a good point.
+
+**Cliff Biffle:** But if we can do that while keeping the compensation uniform, or at least fair, that would make me really happy.
+
+**Jerod Santo:** Yeah.
+
+**Adam Stacoviak:** What does it do for your personal ability to show up to not worry about compensation...
+
+**Cliff Biffle:** A lot.
+
+**Adam Stacoviak:** ...as much as you had to before?
+
+**Cliff Biffle:** Oh, my God. So not having to worry that some of my coworkers are getting screwed over by having not like hardball negotiated in their interview... So I became a manager at Google, of a team that I had previously been on... And at that point, at the next promotion cycle, I was able to see everybody's salaries for the first time, and that was how I found out that there was like a hundred thousand dollars a year difference in salaries among people at the same level on my team.
+
+**Jerod Santo:** Wow...
+
+**Adam Stacoviak:** That's a lot of money.
+
+**Cliff Biffle:** Yeah. And it was mostly us guys that had the higher numbers. It was kind of crap. So I don't have to worry about that here, which is great. And I feel like people are a little more comfortable talking about both job conditions, and also kind of like financial stress. Some people have been really open about like "My husband lost his job, and we've got the new kid, and so it's kind of rough right now..." I feel like people are a little more comfortable sharing that, because we all know what we each other make. So it's not like you're going to reveal that "Oh, wow, you're being really overpaid", and now all your colleagues are mad at you. So it's got its perks.
+
+**Adam Stacoviak:** This may not be accurate, but one thing I thought about was the fact that it seems like you all are owners of the company. Like, everyone in there that works here owns --
+
+**Cliff Biffle:** Small owner, yeah.
+
+**Adam Stacoviak:** You all have equity. Some may have more, some may have less, and the compensation is the same across the board, but what changes is you have a different job. One person has a CEO job.
+
+**Cliff Biffle:** Yeah.
+
+**Adam Stacoviak:** Now, that person may have more equity, but that's because they also started the company.
+
+**Cliff Biffle:** It's because they were here early.
+
+**Adam Stacoviak:** The compensation is a little different. But the day-to-day, the check, the reason you show up, it seems like an even playing field, and you all are sort of owners in a way...
+
+**Cliff Biffle:** Yeah.
+
+**Adam Stacoviak:** ...as much as you can be owners, equitably.
+
+**Cliff Biffle:** I really don't like hierarchy. When I'm managing people, I view -- manager's just a job. Like, it doesn't mean that I'm above you, or like in control of you, or better than you. It's that I'm going to do the manager things, you can do the engineering things, and we're both happy. And I'll develop skills here, you develop skills there. But I kind of feel like we've got the same thing going up through Steve. Steve is the CEO. He does the CEO work that we don't want to do.
+
+**Adam Stacoviak:** Who wants to do that job, right?
+
+**Cliff Biffle:** The sales guys --
+
+**Adam Stacoviak:** That job he described on stage... I was like "Wow, that is a hard job."
+
+**Cliff Biffle:** Yeah. Like, our sales team - they do the sales that I don't want to do. They're good at it, clearly. They actually make different amounts of money... They're the one corner case. Because salespeople - we have an incentive thing, where the more they sell, the more money they make...
+
+**Jerod Santo:** Yeah. Compensation.
+
+**Cliff Biffle:** ...which we're excited about, and we don't have a problem with. So yeah, I do think it helps with the sort of sense of like we're all in this together...
+
+**Jerod Santo:** Yeah.
+
+**Cliff Biffle:** Which is good.
+
+**Adam Stacoviak:** 100%.
+
+**Jerod Santo:** Well, we'll let you get back to it.
+
+**Cliff Biffle:** Alright.
+
+**Jerod Santo:** Thanks for chatting with us. It's been awesome.
+
+**Cliff Biffle:** Yeah, my pleasure.
+
+**Adam Stacoviak:** We appreciate it.
+
+**Break**: \[27:28\]
+
+**Adam Stacoviak:** What is the update on the update?
+
+**Dave Pacheco:** Well, let's see... We're working on shipping the first version of what we call a self-service update for the Oxide rack. So today, the process for updating the Oxide system, including the control plane and everything, involves a support process where our support engineers are getting on the system through a debug interface... And that works pretty well for a lot of things and it's very simple, but it doesn't work for a lot of customers to have an Oxide person involved in the whole upgrade operation. And so we want that to be something that they can do through the API, just like they can do all the rest of the infrastructure stuff. But that is kind of a big deal, because that means that the control plane is driving the update, which means the whole control plane is online during the update. People use different metaphors for this. It's replacing all the parts on the car while you're driving down the freeway, or replacing all the parts on the plane while it's in the air, or whatever, but... It's just a lot of work to make that work.
+
+**Jerod Santo:** How long have you been working on that?
+
+**Dave Pacheco:** I've been working on it for about two years.
+
+**Jerod Santo:** How long has Oxide been working on that?
+
+**Dave Pacheco:** We needed an update process for our MVP. And so that started before that and finished before that, because we launched before that... But that's the process that we use today, the support-based process.
+
+**Jerod Santo:** This is like an upgraded rewrite.
+
+**Dave Pacheco:** Kind of, yeah. So the idea behind the first version of update was what we called the minimum upgradeable product, which was Mupdate... And so the idea here is that, you know, it's an MVP, we're shipping it really as fast as we can... But there's a lead time between when you deliver software to the factory and when you actually get it at the customer site... And that's a period in which we can continue working on the MVP. And so what we needed was for there to be enough in that first thing that we could update it to whatever software we wanted once we got to the customer site. And that became the minimum upgradeable product, which is Mupdate, and that's the procedure that we have today.
+
+So the priority there was about having a robust support procedure for recovering the software on any one of our compute sleds. And that's something that we knew we needed even separately from update. And so that's why it came first, was this idea - like, we need the ability to recover a sled (one of our compute sleds) no matter what state it was in, and we can use that to do our initial updates. And that's what we've been doing for the last two years. So in that sense, we've been working on it for a while. But in terms of being able to have the control plane do a more operator-friendly update, that's been about two years.
+
+**Jerod Santo:** Gotcha.
+
+**Dave Pacheco:** And actually, even that - there are a bunch of building blocks involved in that; what I've come to call dynamic reconfiguration of the system. So having the ability for any component to come and go while the system is running is kind of a prerequisite for that, but it also allowed us to deliver other important things that customers would expect to be there... Like the ability to remove a sled, replace a sled, add a new sled to the system. The first year of the update project was really building this foundation that we use for these other support procedures as well.
+
+**Jerod Santo:** Gotcha.
+
+**Adam Stacoviak:** \[31:47\] What exactly is an update? Is it big? Is it small?
+
+**Dave Pacheco:** That's a really good question, because --
+
+**Adam Stacoviak:** What's happening?
+
+**Dave Pacheco:** ...you think of the control plane as like "It's the control plane. It's just like one big piece of software", or something like that.
+
+**Adam Stacoviak:** Sure.
+
+**Dave Pacheco:** But actually, every single one of our -- so in a rack we've got 32 sleds, we've got two switches, and we've got a couple power shelf controllers. Every one of those has a service processor, a root of trust, and the root of trust has its own software, and the bootloader software. Then, on all the 32 sleds, we also have a host OS, which comes in two parts, for historic reasons around bootloaders and stuff. Then we have all the control plane software on top of that, including storage software, which is one per disk. What all this means is that when you update the software in an Oxide system, you're updating literally hundreds of components, and we're kind of doing it one at a time. And you're also going through all these intermediate states where you're running some of the old software and some of the new software. So you ask "What is an upgrade?" We're replacing all of the software running on everything in the system. And it's a lot of different things.
+
+**Adam Stacoviak:** It's a big deal then.
+
+**Dave Pacheco:** Yeah. And that list I just gave doesn't even include a lot of stuff that for us gets bundled. CPU has microcode, NICs have their own firmware. For the update system, that's simplified, because it all gets part of the host OS, but there's a lot of software in the system, and it's updating all of those things. But the whole idea of what we're doing is that operators don't have to think about any of that stuff. So our release process puts together a giant zip file; it's like two or three gigs of data. You download that from us, you can look at it if you want, you can validate it, whatever... But then you upload it to the API, and you hit Go, and then the system goes and churns for probably a couple hours up front, and then you come back, and then the whole thing's updated. So the idea is that the operator is only thinking about this policy. They're not thinking about all those other things that are involved.
+
+**Adam Stacoviak:** That's nice it's a zip file. I mean, several gigs, though...
+
+**Dave Pacheco:** It's a lot of software, yeah.
+
+**Adam Stacoviak:** I mean, even the bandwidth cost on that. Do you measure that? Does it matter, as you grow your customer base? I mean, obviously not, because you're getting paid lots of money... But you've got to worry about those things, right? Speed to get it, accessibility...
+
+**Dave Pacheco:** It's a good question. It's not something we've been focusing on for the most part so far. It's the kind of thing where the customer is currently going to be responsible for getting that from whatever our download site is, whether that's GitHub, to the rack.
+
+**Jerod Santo:** \[unintelligible 00:33:58.20\]
+
+**Dave Pacheco:** Yeah. So they might get that on their laptop, downloading it from GitHub, or something like that. And then they'll upload it to the rack.
+
+**Adam Stacoviak:** \[unintelligible 00:34:02.17\] Nice.
+
+**Dave Pacheco:** Yeah, that part is GitHub's problem.
+
+**Adam Stacoviak:** It's a release. Alright, that makes sense.
+
+**Dave Pacheco:** Yeah. And then uploading it to the rack is over their network. And we've kind of been assuming it wouldn't be an issue... That's a fair question, whether they would consider that an issue.
+
+**Adam Stacoviak:** I was thinking about size, the file size. I guess most people can download a couple of gigs pretty easily, without it being a major problem... But some people can't.
+
+**Dave Pacheco:** It's definitely a lot.
+
+**Adam Stacoviak:** I've been told that before. I guess if you're --
+
+**Jerod Santo:** Most of their customers are not...
+
+**Adam Stacoviak:** Yeah, they're probably not having this problem. I'm solving different problems here, Jerod.
+
+**Jerod Santo:** What about air gaps? So some people opt for an air gap.
+
+**Dave Pacheco:** Yeah, that's really important, too. This model works pretty well for air gaps, because you, the customer, are downloading it to your laptop, and then uploading it...
+
+**Adam Stacoviak:** No connection required.
+
+**Dave Pacheco:** No connection to the -- the rack definitely doesn't care. It doesn't even know if it's connected to the internet. That's fine. It doesn't care.
+
+**Jerod Santo:** Right.
+
+**Dave Pacheco:** You could imagine a nicer experience where the rack was connected to the internet, and could see, "Oh, Oxide's just published a new thing, and I'm going to download it." And maybe I only download certain parts that I need, or I download it one part at a time... Maybe that helps with some of the bandwidth stuff. But actually, it's not something most of our customers are interested right now. Most of them are actually more interested in the "I'm really not connected to the internet, and I'm doing this because I care about my security, and my privacy, and my data. I definitely don't want the rack talking to the internet", so that's why we've done it the way we've done it.
+
+**Adam Stacoviak:** We talked to Cliff earlier about Hubris... So when you look at Hubris, that's the operating system. You mentioned a couple of chips, every different device on there... When the update comes through, is that on top of the API of Hubris? Or is that -- how are they compared to each other, this update and Hubris?
+
+**Dave Pacheco:** Yeah, so Hubris is a sort of... It's an operating system. We use it in a couple of different components in the service processor and the root of trust, on all of these systems: on the sleds, the switches, and the PowerShell controllers. So it's one of the things that we update. It is also true that in order to update everything, we talk to the service -- or in order to update much of the system, we end up talking to the service processor, which is talking to that Hubris thing. So we end up using the current version of Hubris through the service processor to be able to update the service processor itself, the root of trust, the root of trust bootloader, and the host OS as well. All the control plane stuff is on top of that and doesn't go through Hubris.
+
+**Adam Stacoviak:** \[36:12\] Is this a novel problem that you all invented, given your architecture?
+
+**Dave Pacheco:** That's a good question. Yeah, a lot of the details are specific to our architecture and pretty novel. It's the sort of thing that I expect cloud providers today have their own bespoke software for. And in fact, large deployments of on-prem stuff will have their own bespoke software to do a lot of this stuff. But a lot of it is also stuff that people kind of just don't update. Like, how often do you update your BIOS? How often does a company running on-prem software update their BIOS? Probably not all the time. But our model involves delivering a lot of value through stuff like that, and we do need to be able to update that stuff.
+
+I remember back at one of my past jobs, we did have to go update the BIOS on 64 systems, or something like that. And you've been in the BIOS thing, right? You're clicking through the thing, and like how do you do that on 64 systems?
+
+**Jerod Santo:** Oh yeah, \[unintelligible 00:36:59.09\]
+
+**Dave Pacheco:** And at the time --
+
+**Jerod Santo:** One at a time?
+
+**Dave Pacheco:** This wasn't a productionized thing, but at the time I didn't know that iTerm or whatever has this mode which is like "Send all my keystrokes to all the other panes."
+
+**Jerod Santo:** You just taught me that, right now. \[laughs\]
+
+**Dave Pacheco:** Someone just opened up 64 panes, and was just like "Enter, Enter, Tab..."
+
+**Adam Stacoviak:** Oh, my gosh. I would be that person.
+
+**Jerod Santo:** That's a 64X developer right there.
+
+**Dave Pacheco:** It's a real problem, right? ...when you have this software at such a low layer; it's not really designed to be interacted with by automation. Everyone that's had to do this has had to come up with their own way to do it, basically.
+
+**Jerod Santo:** Yeah. Well, one reason why you don't update your BIOS very often is because you've got to reboot your machine. And I know that ultimately, your guys' goal is like "No, reboot update", right?
+
+**Dave Pacheco:** That's right.
+
+**Jerod Santo:** That's not what you're working on now, though.
+
+**Dave Pacheco:** That's going to be the next phase, and it's going to be a much smaller part of the problem for us. At least we expect.
+
+**Jerod Santo:** It's going to be easier...
+
+**Dave Pacheco:** You never know until you're done, but...
+
+**Jerod Santo:** It's not going to take you two years...
+
+**Dave Pacheco:** Right. So we've done a lot of the pieces involved in that. All the stuff we've been doing so far is like the orchestration... It's foundational stuff. So our system is based on what's called the plan-execute pattern, which means that before taking any action, the system generates a new intended state of the world, which we call a blueprint. And then it goes and executes that blueprint. And all that was really important foundational work for building a system that can be operated autonomously, which is also really important for the air gap thing... Because we can go test all kinds of things that can happen with just the planner part, without even worrying about the execution stuff.
+
+**Jerod Santo:** Right.
+
+**Dave Pacheco:** And then we can go test all the execution stuff given whatever plan we want, without having to have gotten a system into exactly that state. And it also lets you do all kinds of things, like ask the system "What are you going to do next?" before it does it. And "Why are you doing that thing?" You know what I mean?
+
+**Jerod Santo:** Yeah.
+
+**Dave Pacheco:** So these are really important operational things that we just need to have. That's the kind of stuff that's taken the first two years.
+
+**Jerod Santo:** So you laid a lot of groundwork.
+
+**Dave Pacheco:** That's right. And so now we're talking doing what we're calling non-disruptive updates. So this is doing updates without rebooting the customer VMs. We're still rebooting the sleds.
+
+**Jerod Santo:** So do you move the VMs?
+
+**Dave Pacheco:** Exactly. We're going to live-migrate the VMs. And now, I don't mean to oversimplify it, but that should be just a question of policy, which is like we're flipping a bit in that blueprint that says "This sled needs to be evacuated \[unintelligible 00:39:03.08\]
+
+**Jerod Santo:** So let's say you have an Oxide rack, 16 sleds... Is that typical?
+
+**Dave Pacheco:** Yeah, \[unintelligible 00:39:08.07\]
+
+**Jerod Santo:** Alright. So we've got 16 sleds, and we need to run an update. This is in the new world, when this exists. And each sled has, I don't know, 30 VMs on it. Now we're doing math in our head... "Uh-oh, he's going to do a math problem." No, I'm not going to. And a couple of times run update. So I go put it on my thumb drive, off my laptop or whatever, plug it into the rack, and it's going to run. It's going to live-migrate VMs off of a sled one at a time, update that sled, reboot it, and then move some stuff back to that one, probably.
+
+**Dave Pacheco:** Something like that. Yeah, that's where actually --
+
+**Jerod Santo:** Distribute that load evenly across the other ones, in the meantime...
+
+**Dave Pacheco:** That's where a lot of the complexity does come in with the non-disruptive update, is first of all, how do you mechanically move these things around? It's like a bin packing problem. But then there's also how do we make sure that we have the capacity to do that? If we're going to start doing -- if you've totally filled every sled...
+
+**Jerod Santo:** If you're running your thing at max...
+
+**Dave Pacheco:** Right, there's no place to put it.
+
+**Jerod Santo:** ...there's nowhere to move them.
+
+**Dave Pacheco:** Right.
+
+**Jerod Santo:** Buy another rack.
+
+**Dave Pacheco:** \[40:11\] Right. \[laughter\] Well, that's the thing. At scale, people actually don't care about this problem. Because keeping a couple of sleds capacity free when you've got 100 racks is like a very small fraction of your cost. And it actually makes sense for a lot of reasons. It also allows you to sustain failures, and put that stuff over there. But when you've only got one rack, that might be more of a problem. So then there's the question of how do you create an experience for the operator that communicates clearly what the trade-offs are, but also gets this input from them, which is like "What do you want to happen?" Do you want me, the rack, to prevent you from using all your capacity, so that you can update it? Or do you want to have the possibility that you go do an update and we just say "Sorry, we're paused right now until you can tell us just like "Reboot all these VMs", or whatever you want to say."
+
+**Jerod Santo:** Right. That reminds me, or that makes me think of failed updates in the self-service world, when this version's out; not the... What do you call it? Undisrupted?
+
+**Dave Pacheco:** Non-disrupted.
+
+**Jerod Santo:** Yeah, non-disrupted. With the current iteration you're working on now, can you guarantee that an update will finish?
+
+**Dave Pacheco:** That's not what I thought you were going to say.
+
+**Jerod Santo:** \[laughs\]
+
+**Dave Pacheco:** Finish? No, because there's things outside of our control. For example, one of the biggest challenges in self-service update - because the control plane is running, we have these intermediate states I mentioned, where you've got new version of software talking to old version of software. And how do you avoid that becoming a problem on our ability to change our own software, because you have to do backwards compatibility forever? And one of the ways we've addressed that is to say that there will be an order to the updates. So we will always update, for example, the host OS before we update the control plane that talks to it.
+
+**Jerod Santo:** Okay.
+
+**Dave Pacheco:** Because the reverse never happens right now. And so that's fine. But that means that if you're doing an update and one of the sleds is like out to lunch and we can't talk to it, we don't know if we've been able to update it, we can't actually keep going and update the rest of the control plane.
+
+**Jerod Santo:** You've gotta wait for that sled.
+
+**Dave Pacheco:** Yeah. So we've got to tell the operator "Look, you either need to do (what's called) expunge the sled", which means to remove it from the control plane, and we'll pretend like it's just caught fire; like, it's failed, and we've moved everything else elsewhere. Or you've got to figure out what's wrong with it and bring it back. And that would be a support call. Probably a support call. And then \[unintelligible 00:42:14.11\] just unplugged it and they just plug it back in, or whatever. Those sorts of things are always outside of our control.
+
+**Jerod Santo:** Now, what did you think I was going to say?
+
+**Dave Pacheco:** I thought you were going to say an update that explodes; that like you start doing the update --
+
+**Jerod Santo:** That was my next question.
+
+**Dave Pacheco:** ...and the control plane is now down, and - what do you do?
+
+**Jerod Santo:** Right.
+
+**Dave Pacheco:** And that's my nightmare. That's been my fear for the last couple of years.
+
+**Jerod Santo:** \[laughs\] That's why it's taken two years and you not done yet...?
+
+**Dave Pacheco:** Yeah, but seriously, that is why we've spent so much time on having the automation take these careful steps where every one of these steps we know is safe. As an example, we've got a CockroachDB cluster that's storing all the control plane database data... We've got five nodes. We definitely don't want to bounce a sled that's hosting a Cockroach node while that Cockroach cluster is already unhealthy. That's just like a thing we want to make sure we never do. Because that increases the risk that we actually lose quorum on the cockroach cluster, and it's dead, and we're in trouble. So we have all these kinds of safeties built into the automation, all this testing, this whole pattern, and all this stuff. So that's like one angle. Obviously, testing is another angle, but that's kind of a given.
+
+**Jerod Santo:** Sure.
+
+**Dave Pacheco:** But it's a hard problem really, because part of what is involved in an upgrade is making backwards-incompatible changes to data formats, like database schemas, and things like that. And once you've done that, the old software can't read the new thing. So rollback is really not possible. What a lot of software does is it'll have a point of no return. Some call it a finalizer, or a deferred update, or something like that, where you basically get the whole thing kind of working before you've committed in that way, and then you ask the operator "Does everything seem to be okay?" and then they hit the button and it's like "Okay, fine. Go."
+
+\[43:57\] But even then, there's still risk there, because whatever it is you're activating by taking that last step hasn't been tested before that. And there's kind of no way to get around that. And that's kind of a future problem for us right now, but it's something we're going to have to deal with.
+
+**Adam Stacoviak:** So it is your nightmare... Has it ever been, your nightmare? Meaning it's happened.
+
+**Dave Pacheco:** Like, in my career, yeah.
+
+**Adam Stacoviak:** Well, specifically for Oxide, and updates...
+
+**Dave Pacheco:** No, no. It hasn't.
+
+**Adam Stacoviak:** Any updates gone wrong?
+
+**Dave Pacheco:** No, but we haven't started doing that in production yet, so there's still time for that, I guess...
+
+**Adam Stacoviak:** How do you do it in production now?
+
+**Dave Pacheco:** The process we used --
+
+**Jerod Santo:** The manual way...
+
+**Dave Pacheco:** Yeah, the manual thing when we -- so how the manual process that I was talking about at the very beginning works, basically it shuts down the whole control plane, and replaces all the software and brings it back up again.
+
+**Adam Stacoviak:** How long is that process?
+
+**Dave Pacheco:** It's actually not that -- it's shorter than the other way.
+
+**Adam Stacoviak:** I feel like that's safer.
+
+**Dave Pacheco:** It is safer.
+
+**Adam Stacoviak:** I mean, I get the whole reason --
+
+**Jerod Santo:** Downtime, man. Can't have it.
+
+**Dave Pacheco:** It's downtime, and it's the self-service aspect. Although you could imagine a self-service version that looked more like that, but then the thing is like, if it's self-service, you're talking to an API, what's running that API while the thing is down? There's nothing, right? So that's why you've gotta do what we've done.
+
+**Jerod Santo:** It's gotta be done.
+
+**Adam Stacoviak:** So downtime's required now. Future is -- you said non-destructive?
+
+**Jerod Santo:** Non-disruptive.
+
+**Dave Pacheco:** Non-disruptive, yeah.
+
+**Adam Stacoviak:** Meaning the VMs get migrated around versus shutdown, but you still reboot sleds.
+
+**Dave Pacheco:** That's right.
+
+**Adam Stacoviak:** You're still rebooting control planes, and stuff like that.
+
+**Dave Pacheco:** Yup. But that's not visible to them, because we have enough --
+
+**Jerod Santo:** Is there a world in which that isn't even required?
+
+**Dave Pacheco:** Which part?
+
+**Jerod Santo:** What if you didn't have to reboot anything?
+
+**Dave Pacheco:** Oh. I think that's pretty dicey. It's definitely a thing that people have done...
+
+**Adam Stacoviak:** Hot swap?
+
+**Dave Pacheco:** Yeah. There's types of updates, or patches, running the latest kernel, where you kind of write the new one over here, and then you jump over there... But then the state --
+
+**Jerod Santo:** It's just a pointer... \[laughs\]
+
+**Dave Pacheco:** It is. It's all software, right?
+
+**Jerod Santo:** That's right. Like, who says we have to reboot, you know?
+
+**Dave Pacheco:** Yeah. That is definitely harder to do with stuff like \[unintelligible 00:46:02.02\] job is to attest to the software that's currently running.
+
+**Jerod Santo:** Yeah. If you're going to change it out from underneath it \[unintelligible 00:46:11.13\]
+
+**Dave Pacheco:** Yeah. You need to figure out what that means. And then there's also this risk that you're now in a different state than you would have been in if you had actually bounced that thing. So have you created a time bomb for yourself, where if that thing loses power and actually does power back on, is it going to do the same thing that it was doing? That's the thing I always -- bifurcated code paths, where you're like "This is the thing we do sometimes and this is the thing we do other times" is totally the kind of thing that results in something failing at runtime.
+
+**Jerod Santo:** A catastrophic failure.
+
+**Dave Pacheco:** Yeah.
+
+**Jerod Santo:** Alright, fine. Bad idea.
+
+**Dave Pacheco:** The thing you asked about, like my nightmare, right? The upgrade just explodes, and we're toast.
+
+**Jerod Santo:** Yeah.
+
+**Dave Pacheco:** Another thing that we've done there is try to create a lot of guardrails around the types of changes that we can make to the software, so that we know if we're making a change that's going to break things. So for example, if you're changing the database schema, we know that you're changing that, and we've operationalized that one. So that's kind of fine. If you're changing like an internal API, we make sure that you're doing it in a way where none of those intermediate states will expose us to a situation where those components don't speak a common version. And that's something I imagine other organizations do do. I haven't actually really seen that before, but I think it's really important. Because that's the way that I've seen this fail in the past, is like, someone goes and makes a change to the API; they're "I'm not changing upgrade." So like they test everything, everything seems to work...
+
+**Jerod Santo:** Right.
+
+**Dave Pacheco:** And you go deploy it and it blows up in the middle. It's like, we've tested the end point, we've tested the beginning point, and we just got unlucky in one of these intermediate states that wasn't tested... And so we've tried to identify the kinds of changes that would cause those problems, and then detect those at CI time for us, and at build time if we can.
+
+**Jerod Santo:** That's cool. Any newer novel testing strategies that you've had to come up with as far as this? I don't know, fuzzing, or deterministic testing, or anything that's...
+
+**Dave Pacheco:** \[48:00\] That's probably the biggest one. The other thing I would point to is that sort of distinction between the plan and execute stuff. We haven't actually gotten to this kind of thing, but one of the things we want to do with that is like property-based testing on the planner, where you're basically like sending all kinds of different inputs at it and putting constraints on what kinds of outputs can happen, and make sure it never does anything crazy.
+
+**Jerod Santo:** Make sure it does that, yeah.
+
+**Adam Stacoviak:** You were supposed to give a talk today...
+
+**Dave Pacheco:** Yup.
+
+**Adam Stacoviak:** It got rescheduled...
+
+**Dave Pacheco:** That's true.
+
+**Adam Stacoviak:** We're not going to be here for it...
+
+**Dave Pacheco:** Right.
+
+**Adam Stacoviak:** ...so you have to spill the beans.
+
+**Dave Pacheco:** Yeah, so let's see. What was I going to talk about?
+
+**Adam Stacoviak:** Updates... \[laughter\]
+
+**Dave Pacheco:** So I've been doing this for a couple of years...
+
+**Jerod Santo:** Update on the update.
+
+**Dave Pacheco:** We have a surprising number of new faces... And so part of that talk is literally the stuff we were just talking about, like "What is update?"
+
+**Adam Stacoviak:** What is it?
+
+**Dave Pacheco:** Well, we've got a couple hundred components that we've got to replace, and... Pick your metaphor, or whatever. And so part of it is just like laying that out. And what we do today... The stuff we've been talking about - why this is a problem for customers, what we're doing, current status of that, which is like we're planning to ship the self-service part very soon now, and then non-disruptive is coming after that... And then the rest of it was probably -- I don't know if that's really interesting to a broader audience, but it's kind of reflections on what it's been like to run a project for such a long term. And I don't know, maybe it is more generally interesting, but... I have a lot of fears about update at runtime, but my big fears about update as a project was that it would feel perpetually a year away, and we would make decisions day to day and week to week that ensured that it continued to be a year away. Because when something's like a month away and somebody asks you to do something else, you're like "Sorry, I've got to do this thing. We're shipping it in a month." But when something's a year away, it's very easy to be like "Well, here's a really important problem over here..." And it's hard to know what the next thing is to do on update.
+
+That's been really a challenge the whole time, is like "What's the next step?" There's so many steps, and there's so many circular dependencies in those steps that you're like "Well, I've got this other important problem over here. Maybe I'll just kind of solve that." And that's fine. Sometimes that's the right call. But if you make that call...
+
+**Jerod Santo:** Over and over again...
+
+**Dave Pacheco:** ...even 20% of the time... Right, exactly. You just never get there.
+
+**Adam Stacoviak:** You timeline stretches.
+
+**Jerod Santo:** How close is it?
+
+**Adam Stacoviak:** He said very soon. Didn't you hear him?
+
+**Dave Pacheco:** Yeah. \[unintelligible 00:50:07.27\]
+
+**Jerod Santo:** Within the next month, or...?
+
+**Dave Pacheco:** That's the plan... \[laughter\]
+
+**Jerod Santo:** That's not a year away, though.
+
+**Dave Pacheco:** No, no, no, no, no. This was two years ago I was worried about it. Two years ago, I think I was like perpetually a year or two away.
+
+**Jerod Santo:** Well, I saw the roadmap this morning, I saw last year's roadmap... It's not a roadmap --
+
+**Dave Pacheco:** Did you diff them and see \[unintelligible 00:50:26.02\]
+
+**Jerod Santo:** One of the things was Update, and then the priority this year is Update...
+
+**Adam Stacoviak:** \[unintelligible 00:50:32.28\] which was first.
+
+**Jerod Santo:** ...so ongoing...
+
+**Dave Pacheco:** Accurate.
+
+**Jerod Santo:** But almost ready.
+
+**Dave Pacheco:** Yup.
+
+**Adam Stacoviak:** Almost ready.
+
+**Jerod Santo:** That's cool.
+
+**Adam Stacoviak:** Well, why you? So we had a side conversation... He's been with Brian almost his whole career.
+
+**Jerod Santo:** Really?
+
+**Dave Pacheco:** That's true.
+
+**Adam Stacoviak:** Sun, to Oracle, the acquisition, and then Joyent... And then what makes you uniquely positioned for this task, this quest?
+
+**Dave Pacheco:** I don't know if I'm uniquely positioned, but right after we shipped the MVP, this was one of my big worries about the products... And I was like "This is something I think I have a lot of experience with", in terms of building distributed systems, and reliable automation, and things like that. And so I thought this was a good opportunity for me to swing in and try to create the vision that I want us to get to, and then be able to execute that. So I was interested, and I asked, and that's how a lot of stuff works around here, so...
+
+**Adam Stacoviak:** What has it been like casting the vision for it and getting feedback from the team? Because I'm thinking culture, I'm thinking the process of getting feedback on any idea... Because you're laying out all this groundwork, and you're doing all the work to kind of get to a direction. What is it like to put that idea out there, that vision out there, and get that feedback and start moving on it?
+
+**Dave Pacheco:** It's great. I mean, I don't know if you all talked about the RFD process, but...
+
+**Adam Stacoviak:** The RFD process, yeah.
+
+**Dave Pacheco:** So when I started on this, the first step was writing an RFD. I think it's 4.18, if you want to find that...
+
+**Adam Stacoviak:** Is this public, the RFDs?
+
+**Dave Pacheco:** Some of them are and some of them aren't. I don't know what that one is. It probably could be, if we want it to be.
+
+There's nothing particularly sensitive in it. But it's kind of laying out where we are. Lay of the land, this is where we are, these are the problems, here's where we're trying to go... And it was very specific in some ways. It was like this idea of plan-execute pattern, and the automation has to be safe, and all this stuff. But it was also very like "We have a lot of stuff to do, and I don't know what all the pieces are yet."
+
+\[52:24\] So that was the first step, is getting everyone aligned on the vision. And that RFD itself was a team effort. I drafted this first version, but people are looking at that... And I think broadly, there were no surprises there. Everyone was like "Yeah, this all makes sense." And then that process just keeps happening. You get more and more specific designs, and say "Okay, let's get some feedback on this." And I do enjoy that part of it, and I enjoy the collaboration... And it goes pretty well. Like, it's not the sort of environment where you're worried about what so-and-so is going to think about this, and is someone going to be unproductive about it, or something.
+
+**Adam Stacoviak:** Sure. We know that Rust is a foundational language here, obviously. We talked about this being somewhat of a novel problem... How is Rust uniquely positioned to help solve this problem? Like, what about Rust makes this problem easier than another language that you may choose to do this with?
+
+**Dave Pacheco:** Yeah. So the big thing for me about Rust, that I really love and I think that's been huge for Oxide, is its ability to help us ensure things, especially at build time, that need to be true of the system. That sounds really vague, but what I mean is you can catch so many problems early. And everyone talks about the obvious ones. We talked about this earlier, but everyone talks about the obvious ones, like, the borrow checker will help you find memory safety problems, but it also allows you to create abstractions for the rest of the team that can't be misused, right? So you can say "I'm creating this thing--" Maybe you represent it with an object, and you say "You can't do these two operations concurrently on it." Well, that's the thing that we can enforce in the type system, and you literally just can't compile the code that would do that. That's awesome.
+
+And it sounds so low-level, but that's what a lot -- like, it's extending that same idea and applying those same tools to do it, that allows us to say "If you try to evolve the API in a backwards-incompatible way that won't work at upgrade time, depending on how you do it, you might get a build failure." You'll at least get a CI failure. I don't know if we have a minute to talk about it... I mean --
+
+**Adam Stacoviak:** Gush, man. Gush.
+
+**Dave Pacheco:** So one of the first things that I built here with Adam Leventhal is something called Dropshot, which basically lets you write an HTTP server and then generate an open API spec from the code. And then we feed that into something called Progenitor, which generates the clients for it. But that alone means that if you make an incompatible change to an API, even before we'd done any of the versioning stuff, your client fails to compile now... Which is like hugely valuable. And that's true, not just because of it doesn't have the operation in it, but like you passed an enum with three variants and it now has four variants, and you need to like accommodate for that fourth one, or something like that.
+
+**Adam Stacoviak:** I call them enums...
+
+**Dave Pacheco:** Yeah.
+
+**Adam Stacoviak:** Just sayin'.
+
+**Dave Pacheco:** Yeah. So anyway, we ended up extending that, so now that's how we do this versioning stuff, is we have a bunch of these open API specs that are the ones that this thing supports... And then we know if you've changed it, because we know if it generates a different thing, then you've changed it incompatibly. So Rust has facilitated all this stuff. And sorry, the last thing I want to say about that with the Dropshot thing is you've got your rich, structured types in Rust, and you're like all happy, "Because I'm in Rust land and I've got my strong types." That carries all the way to the client, because of the way it goes -- like, Dropshot just takes those types and puts them in the open API spec, and then the client generates faithful things on the other side. So you basically get that strictness all the way through there.
+
+**Adam Stacoviak:** Gush a little bit more on this then. Speak to your confidence in the code you write because it's Rust. I'll just -- easy shot you that one.
+
+**Dave Pacheco:** Yeah. I mean, I'm reluctant to make any bold claims about the correctness of the code I write, lest I immediately walk back in there and there's some horrible thing happening...
+
+**Adam Stacoviak:** Sure. If the fans spin, you know that's why.
+
+**Dave Pacheco:** \[56:07\] But I mean, I'll say this... There's a lot of changes I've been able to make where I'll go work on the code for like four hours, and when it compiles, I know it's already correct. I know I haven't broken anything. And it's not as simple -- I know there's the cartoonish version, "if it compiles, it works." And I'm not talking about that. I'm talking about something that's either a refactor... This happens a lot with refactors. Or I'm building a new thing in terms of these things that already exist, and I'm plugging into the middle of it... And it's like, there's no way for this to be wrong at this point, because it fits neatly into the narrow interfaces on both sides of it. It's correct. And that is huge. It's so huge.
+
+**Adam Stacoviak:** No nightmares.
+
+**Dave Pacheco:** My previous experience was in Node.js, and it was the complete opposite of this. And part of the reason I love Rust is by the time we got to the end of the road at my last job when we were using Node.js everywhere, every single JavaScript function we had started with like 30 assertions about the types of all of its arguments. And I'm like "Why are we doing this? The computer can do this. It's what the compiler is for." And so I have so much more confidence in those things now.
+
+And that gets back to why I think Rust is so valuable for this problem space, is that these things are so complicated... By allowing us to encode all these constraints into build time constraints, that allows us to evolve the software so much faster, and with so much more confidence. Someone can come in here and make a really big complicated change and you're not wondering about "I wonder if they missed some callers", or something like that. It's like, no, you haven't. You've covered every single case. And that's huge.
+
+**Adam Stacoviak:** It's huge. Right on, man. What else? Anything?
+
+**Jerod Santo:** Tell us about OxCon.
+
+**Dave Pacheco:** OxCon. OxCon is awesome. So this has been something we've been doing for a while, and this is something we -- it's very similar to something we did at Joyent, these sort of company-wide meetups, engineering-wide meetups. For a remote company, it works really well to be remote basically all the time, but it is also so valuable to have that time in person together, to get to -- there's so many conversations that don't happen if you haven't scheduled a meet for it. And we have other ways of trying to have those conversations anyway, but it's just really nice to have the FaceTime with people, and to also -- like, the small talk, and you're going to get dinner, and you're just like talking to people about whatever it is, you learn more about them... It's a really good time.
+
+So there's a mix of the structured time that we have... You know, Brian and Steve talking about all the exciting stuff over the last year. It's inspiring for people, it sort of gets everyone ginned up to talk about everything that we got to go do... And then there are other good, company-wide sessions talking about important projects and stuff like that, but then there's also all this breakout time that I started by talking about, which - as with any conference, the hallway track is almost... It's at least as important as the rest.
+
+**Jerod Santo:** It's the best.
+
+**Adam Stacoviak:** It is the best.
+
+**Dave Pacheco:** Yeah, right? So that's huge...
+
+**Adam Stacoviak:** Conference talks, not really... Hallway track only. Keynotes only, and then hallway track.
+
+**Jerod Santo:** That's our move.
+
+**Dave Pacheco:** This is so much more -- it's getting more mature and professionalized. We never had a stage before...
+
+**Adam Stacoviak:** No?
+
+**Dave Pacheco:** We never had like pro A/V before... So this whole week is a little bit of a dream. OxCon already feels different. And then we've got the stage there, and then it like rained in the Bay Area in September this morning, and I was like "Am I dreaming, or what's going on here?"
+
+**Adam Stacoviak:** Yeah.
+
+**Dave Pacheco:** But it's a good time.
+
+**Adam Stacoviak:** Thanks, Dave.
+
+**Dave Pacheco:** Thank you. I appreciate it.
+
+**Jerod Santo:** Thanks. It's been awesome.
+
+**Break**: \[59:28\]
+
+**Adam Stacoviak:** One of the things that stands out about Oxide is its design. I think you can have a good company, a successful company, but every company that's successful is set apart by its design. Can you talk about how you came to be here at Oxide, and the design story behind the brand?
+
+**Ben Leonard:** Yeah. So I was working at a branding agency called Pentagram. Oxide was my last branding project. I was planning on leaving to work as a freelancer. I'd been at Pentagram for four or five years, I was ready to move on. And it just so happened that Oxide were looking for a designer just as I was leaving, and it felt like a unique opportunity.
+
+Because what would happen is you'd work on these brands and then you'd throw it over the fence, and then you'd check in on it like two years later, like "What have they done to my baby?"
+
+**Jerod Santo:** They destroyed it.
+
+**Ben Leonard:** "My boy, my boy...!" And so it was a unique opportunity to continue it on. And I'm kind of -- I was telling you about this before, I'm relentless. I like to fiddle with things.
+
+**Jerod Santo:** You keep working on it.
+
+**Ben Leonard:** Yeah, I kind of get bored of it as well, so I just want to tweak something... Recently, we made a modification to the logo that most people wouldn't even notice. It's slightly thicker, but it's something that's been bugging me this entire time...
+**Adam Stacoviak:** I noticed.
+
+**Ben Leonard:** Yeah, yeah, yeah.
+
+**Adam Stacoviak:** When Steve was up here, I was like "Steve's shirt has the thinner logo..."
+
+**Ben Leonard:** Yeah, yeah.
+
+**Adam Stacoviak:** And on stage here, it's much thicker. I noticed that.
+
+**Ben Leonard:** Right away, right away.
+
+**Adam Stacoviak:** I saw it. I noticed kerning... Oh, yeah.
+
+**Ben Leonard:** Yeah, so I'm always fiddling with stuff. But then the beauty of Oxide is I get to work across everything. So I'm working across -- generally, most of what I do is working on the product. So in my case, it's the web console. So I work on the design system, on the UI, but I'm also working on the brand, and the marketing, and the sales assets and all this stuff; on the industrial design... The beauty being is that those all inform each other. Because what tends to happen is you have your product team and your creative team, and they're very distinct. This is a bit in the weeds, but our --
+
+**Adam Stacoviak:** Go in the weeds.
+
+**Ben Leonard:** We have the same UI design system that drives every piece of design. We have the same colors, we have the same UI elements on both the website and the web console. They're kind of continuously informing each other. I mean, it's difficult, because I think there's been a shrinking of creativity in tech design, that is branding stopped being -- what it was is that the product itself, so the proliferation of SaaS, the product itself became the design. So the UI language of the product became the design.
+
+Linear is a great example of this, which is the design of the product and the design of the creative are one and the same. What that can mean is the world of design within product is much smaller. And so what that can mean is you have -- everyone's always complaining that all tech looks the same now. And that's probably because product is much smaller, and if your brand looks like your product, then your world is much smaller.
+
+I'm always figuring out how wide I can go with the branding, and keeping it still like it comes from the same world as everything else... And yeah, the industrial design, the creative, the product - there's so much to work on, so you have this variety.
+
+**Adam Stacoviak:** This rack behind us... Would this rack behind us look like that at all if you weren't here?
+
+**Ben Leonard:** It'd look a bit like that.
+
+**Jerod Santo:** Same shape, tall...
+
+**Ben Leonard:** Yeah -- if I wasn't here, it'd be squat, it'd be half the size... No, I think--
+
+**Jerod Santo:** Less green, probably.
+
+**Ben Leonard:** It might be less green...
+
+**Jerod Santo:** \[laughs\] Or maybe more green.
+
+**Ben Leonard:** We consulted a little bit with my old agency, Pentagram, industrial designers there, and we worked together on that a little bit. One thing that I'd learned from working on a hardware product previously is the first version was a small run that they distributed to people... And so they produced this thing, which is initially really magical, and then as soon as they mass-produce it, they think "Okay, we're going to have to make this cheaper, easier, more practical to make." And then the next version looked really, really bad. Because essentially, they were just trying to recreate what they've done, and "Okay, we can't do this complicated tile system, so we'll just print it directly on--" It was a PCIe card. And what it was - it was compromised in the worst way... So I think going into it from the beginning thinking "What are our limitations? We're going to make thousands of these, from the beginning..." I don't want to do things that just exist in this kind of initial run... And so figuring out what those compromises are. But industrial design, working with hardware is hard. I haven't done very much of it. But yeah, as soon as you're touching materiality, you're dealing with --
+
+**Jerod Santo:** \[01:04:46.09\] Things are different.
+
+**Ben Leonard:** Yeah. I mean, we were speaking about like color; color is the bane of my life. We have painted drive bays, we have powder coated metal, we have plastic pieces, and color matching the green I think is --
+
+**Adam Stacoviak:** Yes, that's probably got to be your thing, right?
+
+**Ben Leonard:** Yeah.
+
+**Adam Stacoviak:** Is that a little off sometimes?
+
+**Ben Leonard:** Oh, it's a little off all the time. It's off all the time. And some of it's outside your control, so you have to think about how to handle that. So what you do is you avoid putting elements next to each other that might be different. If there's a little bit of separation, then visually you can get away with a little bit of difference between those things. One example is originally the rack was black, but it was a slightly bluey black.
+
+**Adam Stacoviak:** Like bluey as in the cartoon bluey?
+
+**Ben Leonard:** Exactly.
+
+**Adam Stacoviak:** Okay, bluey black. Disregard.
+
+**Ben Leonard:** It was a cool black. Cool as in like color temperature, not like cool black.
+
+**Jerod Santo:** Oh, I thought you liked it.
+
+**Ben Leonard:** Yeah, yeah. But my thought was as soon as we start integrating multiple components, then I'm trying to make sure I'm using the exact right black. Because --
+
+**Jerod Santo:** Impossible.
+
+**Ben Leonard:** And yeah, so --
+
+**Adam Stacoviak:** That's hard.
+
+**Ben Leonard:** You're trying, yeah. Although I don't know how much of it is me trying to avoid upsetting myself. Like, these things that no one else notices...
+
+**Jerod Santo:** It sounds like a lot of it is internally-driven.
+
+**Ben Leonard:** Stuff no one else notices.
+
+**Adam Stacoviak:** Yeah. They're like "Whatever..."
+
+**Ben Leonard:** This is like trauma-driven design. You guys have TDD, I have the same.
+
+**Adam Stacoviak:** Do you get into the PCB design then, too? Or is it simply this -- and I don't want to simplify what you do, but...
+
+**Jerod Santo:** "Just the easy stuff back there..."
+
+**Adam Stacoviak:** ...do you get to step into that world where, like how is this -- because I think about Apple. Apple has done a great job of like branding everything.
+
+**Ben Leonard:** Yeah, yeah, yeah.
+
+**Adam Stacoviak:** From the look of their CPU, right? From their latest M1 chips, or the latest M series chips, it's a design to everything they do. Do you get into that as well?
+
+**Ben Leonard:** I mean, Apple's a good example, because they're stunning inside as well. A little bit... Occasionally I dip my toe in and I get shouted to leave... The PCBs are currently green, and I think once I came into a channel and I was saying "Hey guys, can we make the PCBs black?" And I'm thinking, "Okay, you change the color of PCB... I don't know, just order black ones instead of green." I was told that's not the case.
+
+**Adam Stacoviak:** I think it does have impact to like thermo, and stuff like that.
+
+**Jerod Santo:** What did they tell you?
+
+**Ben Leonard:** They told me some information, none of which I retained...
+
+**Adam Stacoviak:** "Sorry I asked."
+
+**Ben Leonard:** Maybe they weren't telling the truth, maybe they were just saying it because they were like "Yeah, yeah, yeah..."
+
+**Adam Stacoviak:** Black was cool. I've seen black PCBs, they were cool.
+
+**Ben Leonard:** Yeah, but there are manufacturing ramifications which mean that it's a bit more complicated than that. I think with the rack you have to take -- there's a balance. We can't go full Apple, and just invest so heavily in details, which costs money for little benefit... But what we can do is make sure that an item which is underappreciated... I like taking boring objects and boring designs and boring things and making them better. I think it's easy to come work with Nike and do some beautiful design for something that's already really compelling. But I think it's much better to take something for which design does not usually kind of feature and try and elevate that.
+
+**Jerod Santo:** \[01:08:24.08\] So Oxide is big on the values... You go to the website, there's values. I'm sure you know all the values. They inform all the decisions, how you act... Do you have a design language or value system or something that drives your design for Oxide, that's separate, or maybe even congruent with that?
+
+**Ben Leonard:** I think there are ideas. I think my approach to branding is you have one idea which should then filter through into everything, and that's how you make a holistic thing. It's not necessarily that this looks like this, it's that it comes from the same place. So with Oxide - Oxide is an old idea brought new. It's this old, crazy idea that you own your own hardware. And so you see that in the design language. You see the ASCII... I mean, the green is like this kind of terminal green... It has this kind of retro edge to it, and so the design language is like old meets new. And yeah, so I think nostalgia is a key part of it. I mean, you see over there, we have the logos, and nostalgia is a key thing. But you don't want to lean so much into nostalgia it's a kitsch. So yeah, I think that sort of referential thing to kind of old computing is a good thing.
+
+**Jerod Santo:** Maybe that's why it speaks so well to us, because we're all about that.
+
+**Ben Leonard:** Yeah, yeah, yeah.
+
+**Jerod Santo:** Where the old meets the new is interesting.
+
+**Adam Stacoviak:** Precisely. What gets you excited about what you do here? What makes you be like "Hell yeah!" Or however you say "Hell yeah."
+
+**Ben Leonard:** I don't know what the British equivalent of "Hell yeah" is. I don't think we're that -- yeah, we're not that emotive.
+
+**Jerod Santo:** \[laughs\]
+
+**Ben Leonard:** Yeah.
+
+**Adam Stacoviak:** What gets you properly excited? \[laughter\]
+
+**Jerod Santo:** Now you're talking in language.
+
+**Ben Leonard:** "I'm properly excited...!" No, I just I love to do a bit of everything. And it's the variety I think that really drives me. I've been here more than four years now, I think that's what -- variety is what keeps it interesting. And Oxide's the company where you get to do that; you have hardware, you have software, and everything kind of in between. And yeah, that's the thing that keeps me interested.
+
+**Jerod Santo:** I would expect the opposite. Because at an agency - you'd think that's where the variety comes, because you're on to the next project. New company, new design language, new brand.
+
+**Ben Leonard:** Yeah.
+
+**Jerod Santo:** Whereas you decide to settle down with one brand, and do that for years and years, I would expect it to be less variety. But you've found that it goes wide.
+
+**Ben Leonard:** I mean, I think certainly working at agency, I think it's always gonna be much more varied than working at a startup. But I think for me there's this interesting tension between what's staying the same and what's different. There are things in the Oxide brand which have retained; logo more or less... The colors... But I think there are things that are kind of constantly changing. And obviously, I think Oxide as a company is changing, so the design needs are changing...
+
+**Jerod Santo:** Sure.
+
+**Ben Leonard:** As an example, I think we are growing, and so our marketing and sales needs are growing... Previously, there was very little collateral outside of the rack, outside of the web console and the website. That was it. So you have that kind of small world. But then, when you're kind of entering the world of sales and marketing, then that kind of opens the door to a bunch more design --
+
+**Adam Stacoviak:** That thing in the door when we came in - was it a cart, with like a chip on it? Did you design it?
+
+**Ben Leonard:** Yeah.
+
+**Adam Stacoviak:** Does that coin come off the cart?
+
+**Ben Leonard:** It does. It's got two sides.
+
+**Adam Stacoviak:** \[unintelligible 01:12:08.16\] You designed that, right?
+
+**Ben Leonard:** Yeah, yeah.
+
+**Adam Stacoviak:** That's fun stuff, right?
+
+**Ben Leonard:** Yeah. I mean, I think those are the palette cleanses. So the big pieces of merch tend to get made just as I finish something that's like drained my creative energy. Website is the big one, because -- I mean, I'm usually coding the thing, too; so I've designed it, I've coded it... It took so long that I'm ready for something else.
+
+I don't know if you saw those, we have the little kind of rack stickers. Those I designed right after I'd done like a run of the website... And so those things are a nice, little --
+
+**Adam Stacoviak:** A little cherry on top.
+
+**Ben Leonard:** They're a way to kind of refill my creative cup a little bit.
+
+**Adam Stacoviak:** Yeah. Are you excited about the growth? I imagine with new footprints, we're doing video... We talked to -- a little fourth wall breaking here behind the scenes... We talked to \[unintelligible 01:13:02.16\] about being able to put motion graphics into place, and collaborating on that... Is that exciting to you, to see the different areas you can do?
+
+**Ben Leonard:** Yeah, for sure. I have a list as long as my arm of things that I will eventually get to. Motion is one of those things. There are things that I want to do, but you don't necessarily have the reason to do.
+
+**Jerod Santo:** Yeah. You need an excuse.
+
+**Ben Leonard:** I was chatting about this before, which is like the size of the gamma of like creativity, or things that you can do... Product design is much smaller, because it needs to be functional, and people are using it every day. There's also the zeitgeist of the way that people use products, and you need it to be designed in a way that's expected... And a website's a little larger, and then the kind of creative space is even larger than that. And then you just think about what are these one-off excuses to do something different. So like product launches, all those things that you can get a bit more experimental around.
+
+**Adam Stacoviak:** We love the work you do.
+
+**Jerod Santo:** Yeah, it's great work.
+
+**Adam Stacoviak:** It's beautiful work.
+
+**Ben Leonard:** Thank you very much.
+
+**Adam Stacoviak:** I think the company is awesome, and I think the design really, in my opinion, is -- it's the glue. You can have a great product, you can have great software, but it's the final piece that says "This is just super-awesome." Basically, the design to me is what sets it apart.
+
+**Ben Leonard:** Yeah. I mean, the design is what says you care about everything.
+
+**Adam Stacoviak:** Yes. It shows intention, it shows trust, it shows all these things, and it's just really important.

We're all Builders now (Interview)_transcript.txt

@@ -0,0 +1,689 @@
+**Jerod Santo:** I want to read your title, because -- I mean, you just can't memorize that thing.
+
+**Amanda Silver:** I know. Sorry. \[laughter\]
+
+**Jerod Santo:** No need to apologize. It's a spectacular title.
+
+**Adam Stacoviak:** I love it. Let it loose. Let it loose.
+
+**Jerod Santo:** Well, today we're honored to be joined by Amanda Silver, CVP - that's Corporate Vice President - for Microsoft's Developer Division. You're the head of product design, user research, general manager of engineering systems... That's a lot.
+
+**Amanda Silver:** It is. It's incredible.
+
+**Jerod Santo:** What does it all mean?
+
+**Amanda Silver:** What does it all mean? I mean, I think at the end of the day there's a group inside of Microsoft that's focused primarily with developers as our primary customers.
+
+**Jerod Santo:** Okay.
+
+**Amanda Silver:** And so when you think about what does Microsoft actually deliver to customers? Visual Studio Code, Visual Studio, .NET, TypeScript, our Azure application platform, our DevOps solutions... We work very, very closely with the GitHub team, do a lot of product integration across our products... So that's kind of what the gig is.
+
+**Adam Stacoviak:** How do you feel about leading people? Is it fun for you?
+
+**Amanda Silver:** I love leading people. I mean, I've actually done it since fairly early in my career. I think maybe two or three years in I started to be a manager of people. When I first started at Microsoft, I was working on the interop layer between .NET and unmanaged code... And so I kind of think about it as like I started at the systems level, and then I started working more and more on programming language design and API design. And then from there, I got more involved in the editor experience, the debugging experience... And that's about when I started to become a manager.
+
+And for me, I think initially, it was a way for me to have more control and more influence over the product, so that was exciting. But I think over the years, I think anybody who's been in the industry long enough recognizes that software is 90%, 95% about people, and how do you construct the team, and how do you motivate them day to day, how do you have the right balance in terms of trying to push them to do what they may not be ready to do on their own, versus when do you take the temperature off and let them recoup from an intense period.
+
+**Jerod Santo:** Right.
+
+**Amanda Silver:** So I find a lot of joy in the act of management, and I also will say, I've come across a lot of managers in my history that are much more self-serving. That's their primary objective. And empire building, a lot of people call them... And for me, I think it's really important that I maintain my personal integrity, and I have to check my ego a lot to make sure that I'm not putting myself before my people.
+
+**Adam Stacoviak:** That's tough, sometimes, right?
+
+**Jerod Santo:** I don't know, I've never led quite as many teams as she's led, I'm sure... So you said empire building, and I date all the way back to when I, as a young college student, used to refer to Microsoft as evil empire. So just full confession. You know, the M dollar sign people... And over the years, I've changed. Microsoft has changed, it seems. And the embracing of open source, which has been kind of a decade-long story, and perhaps more, has been an amazing thing to watch from the outside, and see my relationship to Microsoft change over the years... And I think you've been along for the entire ride.
+
+**Amanda Silver:** Yeah. I mean, that's been kind of core for my career at Microsoft, in a lot of senses. And even when I started at Microsoft in 2001 - let's back up for a second.
+
+**Jerod Santo:** Okay, that's my college years right there.
+
+**Amanda Silver:** When I started at Microsoft, I had two older brothers who were both in the tech industry, and were part of the dot-com bubble bust.
+
+**Jerod Santo:** Sure.
+
+**Amanda Silver:** They were each on their third job or something like that by the time I graduated from college. And so I thought, when I was graduating, like -- I thought I was going to be a scientist, because my dad was a scientist... But I thought, "You know, if I'm going to go get a PhD or whatever, maybe I should try industry a little bit." And so on a long shot, I just went to the tech career fair and handed out my resume to different companies that I thought would pay me decently in cash, because I wasn't interested in stock at the time... And Microsoft seemed like a relatively stable company. And Google at the time was a startup, and I was like "Not going to apply."
+
+**Jerod Santo:** \[08:05\] \[laughs\] Too risky.
+
+**Amanda Silver:** Too risky.
+
+**Adam Stacoviak:** That's funny.
+
+**Amanda Silver:** So I ended up at Microsoft. And I think my first decade or so I was really focused on enterprise software, .NET primarily. And it was like the Java/.NET tension, and that was kind of the main primary competition that we were thinking about at the time... But open source wasn't in the vernacular, it wasn't a thing at the time...
+
+At the same time, I think it was ASP.NET that was the first to actually include open source in the product jQuery. Everybody had to use jQuery to be able to manage the different browser experiences, and so we first started to ship jQuery as part of ASP.NET, I guess in the 2008, 2009 kind of era. At the time, I was moving more -- I kind of moved into the JavaScript space. I started to work on - around 2009, 2010 I started to work on the Chakra JavaScript engine that was inside Internet Explorer at the time... And that was like a really big change in terms of the day-to-day competitive atmosphere that we were working in. Enterprise software moved much more slowly than the pace of the web at that time, and so it really changed the cadence for what we had to think about... And that's actually when we started to work on TypeScript.
+
+TypeScript, originally, when we first started, was really trying to answer the challenge that we had inside Microsoft, which was that we were building what we now call M365, which is like the web experience for Excel, and PowerPoint, and SharePoint, and everything... But we had this challenge that we had a lot of developers inside Microsoft that had deep, deep, deep familiarity and decades of experience with C++ and C\#, but they really didn't know how to build for the browser. They didn't know how to build complex applications in JavaScript. And actually, at the time, the industry didn't really, either.
+
+**Jerod Santo:** Do we now? \[laughter\]
+
+**Amanda Silver:** We're a lot better.
+
+**Jerod Santo:** We're getting there.
+
+**Amanda Silver:** But a lot of the challenges were really about encapsulation and modularity, and how do you create modules. And so that's where TypeScript came from. And that's when we -- TypeScript was really our first open source project that we did fully open source from the get-go. And I remember in that era it took me like six months to convince the muckety-mucks that we should --
+
+**Adam Stacoviak:** The muckety-mucks...
+
+**Jerod Santo:** How did you finally convince them? What was the winning argument?
+
+**Amanda Silver:** I think the argument was... So our objective at the time was to make sure that our internal developers didn't end up on a different path than the broad open source ecosystem that was benefiting from the evolution of JavaScript. We had decades of experience at that point of building our own C++ compiler, that kind of became really more the internal Microsoft C++ compiler, and was divorced from other paths of C++ compilers that were being used more broadly in the industry... And there was a challenge in terms of trying to keep them aligned. So over time, the internal Microsoft developers didn't get to benefit from what was happening in the broad industry on the C++ compiler. So with that experience, looking at this problem of "How do we address this large-scale JavaScript solution challenge?", we decided first of all we have to kind of stay in line with what the broad web industry is going to end up using...
+
+\[12:01\] But then secondly, we also thought that if that was going to be the case, if we wanted to create something where we call it "first party equals third party", meaning that our developers internal to Microsoft use the same tools that our third party developers use, our external developers use.
+
+**Jerod Santo:** Okay...
+
+**Amanda Silver:** If we wanted to accomplish that, then we had to build something that the JavaScript community would actually use and like. And at that time, there was still a fair amount of hostility towards Microsoft in that community.
+
+**Jerod Santo:** Sure.
+
+**Amanda Silver:** And so we absolutely had to launch it as open source, to be able to introduce TypeScript to the world and start to get traction.
+
+**Adam Stacoviak:** From the inside, describe the hostility.
+
+**Amanda Silver:** The hostility from the community?
+
+**Adam Stacoviak:** Enumerate over how hostility shows up and manifests...
+
+**Jerod Santo:** Yeah, how do you see it?
+
+**Adam Stacoviak:** ...from the developer community at that time.
+
+**Amanda Silver:** Well, I mean, there's many different forms. There's some folks who would never even consider anything from Microsoft, because there's just some kind of halo effect of history, or something like that, that they would refuse to use anything from Microsoft in their stack, period. And that they won't even look at how good the technology is.
+
+Then there's kind of like the indifference, or treating Microsoft products as though they are irrelevant... And they just, again, wouldn't use it or consider it because Microsoft couldn't come up with it. It's kind of a disbelief, right? Microsoft couldn't come up with anything useful. And then there's the more kind of common conversations that we have inside the industry - I think everybody has them - which is more of the debates, where it's like, you end up with one developer who likes the technology, and can speak about the advantages of the technology, and another developer who has another argument, and they dislike the technology, and they will enumerate all of the ways that they dislike the technology, but really, at the core of it it's really some other kind of emotional thing. It's not actually on the technical merits.
+
+**Jerod Santo:** Sure. I think in the past that was a lot clearer lines to draw, because you kind of couldn't live entirely in Microsoft's world, or live entirely in the open source or Unix-y world... And now it's just much more of one world, where it's like -- even if you are skeptical of Microsoft, try not to use some Microsoft open source. It's going to be used around you, and probably forced upon you perhaps by your teammates or something... And at the same time -- I mean, I think Azure made that change, to a large degree. Between open source and Azure in the cloud, it's lik, yeah, kind of ubiquitous at this point.
+
+**Amanda Silver:** Well, that was the next thing that happened after TypeScript. We started to get a little bit of traction with TypeScript. And actually, there was a fantastic partnership that we had built with the Google Angular team at the time, that actually kind of got TypeScript... You know, in some senses, no programming language starts to get traction until it has frameworks. And so it was the Angular team at Google that we had a really close relationship with in building TypeScript in that era. And this was, again, 2011-2012... There was a lot of fickleness in the web community in terms of different frontend stacks. It was Angular, then React, and then Vue, and so on and so forth. But six or seven different frontend frameworks kind of made it through in those four or five years, that were very, very popular. And what became fairly obvious to us was we needed to create something that was a little bit more durable, that would be able to survive those different epochs of frontend frameworks. And I think over time, TypeScript kind of became that thing - which was great - and started to get more of the frontend community to use something in our stack, almost to their chagrin or reluctance or whatever... But I think that started to open the door. And then --
+
+**Jerod Santo:** \[16:15\] And then VS Code kicked the door open. \[laughter\] Am I right?
+
+**Amanda Silver:** And then we introduced VS Code. And I think that, in a lot of senses, TypeScript and VS Code actually went really hand-in-hand... Because part of what TypeScript was doing was creating static types over JavaScript. And the tooling for JavaScript wasn't particularly good at the time, because it was very hard to build great tooling for a dynamic programming language. And what TypeScript did is it created a way that we could create fantastic tooling. Whether you were writing code in TypeScript or in JavaScript, it didn't matter. We could create great tooling based on the TypeScript language service in VS Code.
+
+And so the hypothesis was that if we created a great developer experience for TypeScript and JavaScript in VS Code, that every developer that's a web developer, it doesn't matter what you do on the backend, everyone has to do a little bit of JavaScript. And so if we created a great developer experience for JavaScript, that that would open doors that would ultimately allow us to pitch a larger tent that brought more developers into the fold, and help them to consider Azure, or any of our other services.
+
+**Jerod Santo:** In retrospect, it's kind of a master stroke. I'm not sure if you masterminded this, or Satya did, or somebody else... Or if it just happened kind of organically over time, as things tend to do... Where it's like these dominoes just lined up, and really did change the brand and the developer relationship to Microsoft over time.
+
+**Amanda Silver:** Well, I mean, I will say, certainly while I was present and helping and involved in shaping the strategy -- I definitely cannot take credit for the overall direction, and whether it was Anders Heilsberg helping shepherd TypeScript to come to the fore, or Erich Gamma and team kind of building VS Code... You know, just incredible people that I've gotten to work with over the years.
+
+**Break**: \[18:36\]
+
+**Jerod Santo:** So now you have VS Code, and TypeScript, and GitHub...
+
+**Amanda Silver:** Yeah. Well, that was a little bit later.
+
+**Jerod Santo:** Yeah, a little bit later. Maybe I'm jumping ahead. But I'm trying to get to present day, because here we are, at Build 2025...
+
+**Amanda Silver:** That's right.
+
+**Jerod Santo:** I was counting the mentions of "agentic" in the keynote this morning, because I'm a nerd like that, and I got to 187.
+
+**Adam Stacoviak:** Okay, yeah.
+
+**Jerod Santo:** Yeah, I even left off things like "model", or "MCP", or... I left Copilot off, which is like...
+
+**Amanda Silver:** Wow.
+
+**Jerod Santo:** Because I feel like you're just going to get to a thousand. I can't count that fast.
+
+**Amanda Silver:** Well, you could always take a transcript, give it to an AI, and it can count it for you.
+
+**Jerod Santo:** That's true, but I wanted to do it live.
+
+**Adam Stacoviak:** I thought about that, too \[unintelligible 00:22:20.09\] I was like "Yeah, you know... It'll be easier later."
+
+**Jerod Santo:** I thought about it as well, but you know --
+
+**Adam Stacoviak:** "...on a word calculator."
+
+**Jerod Santo:** ...we're going to keep the mind going as well... VS Code - I don't to call it a Trojan horse, because it has negative connotations, but it's kind of this thing that you've gotten out there now, as an open source project, and as a product that... I mean, how many millions of people use it, right? You probably know rough numbers...
+
+**Amanda Silver:** We have 50 million monthly active users.
+
+**Jerod Santo:** 50 million.
+
+**Amanda Silver:** Five zero. Across VS Code and Visual Studio together.
+
+**Jerod Santo:** Okay, so that's two of them.
+
+**Amanda Silver:** Yeah.
+
+**Jerod Santo:** And which one's bigger?
+
+**Amanda Silver:** VS Code.
+
+**Jerod Santo:** VS Code's bigger.
+
+**Amanda Silver:** Significantly.
+
+**Jerod Santo:** So, you have both arms of that. You have the IDE people in Visual Studio, and then you have the text editor people in VS Code. And through those platforms you can launch all this other stuff, right? Like, all this Copilot stuff. Is that how you look at it, or is that just how I look at it?
+
+**Amanda Silver:** I think that we can kind of bootstrap a lot of developers to get more familiar with Copilot, for sure... And I think that in a lot of senses the code editing experience -- the table stakes have changed. You have to have AI as part of the code editing experience. So I don't know if it's as much as us going and forcing it on everyone, as much as it is "This is what is now expected of a modern code editor."
+
+**Jerod Santo:** Right. And I don't mean forcing it. I just think that you have this platform in which you can launch other stuff. And Copilot really has had a great opportunity there, to just be like "Bam. Right there. You're already using VS Code..."
+
+**Amanda Silver:** That's true.
+
+**Jerod Santo:** "It works great in VS Code. Click the button, bam."
+
+**Amanda Silver:** I think that in a lot of senses -- you know, Microsoft has always been a developer-first company, since the MS-DOS BASIC days. That's been where Bill Gates' heart was at, and I think that as it moved through Balmer to Satya, we've consistently had a great sponsorship for our developer tools and platforms throughout history. And I think that the reason for that is because at the core, the CEOs and Microsoft always thinks about its reason to exist as a platform company. We are building a platform that other people build incredible things on top of, and to do that, you have to have developers as your focus. And so I think of the work that we do in the developer division as - it is a platform to bootstrap new things, new platforms, new adoption of new tools, new workflows... No question. But at the same time, there's lots of things that we've tried to launch in that way, and it didn't take off.
+
+**Jerod Santo:** So it helps, but it's not like a cure-all.
+
+**Amanda Silver:** Yeah. Exactly.
+
+**Jerod Santo:** It's still got to be good.
+
+**Amanda Silver:** Yeah.
+
+**Adam Stacoviak:** Well, I think it's telling though that it happened with VS Code, though. It's 50 million developers across two different editors, combining them I guess that way... And that's a lot of developers you have a captive audience. I think that's what he's alluding to, is you have a captive audience to say "Okay, as you launch new things", or even breakthrough, like Copilot, for example, that you have a lot of developers that already have attention... And it's not that much harder to launch. It's distribution for an idea.
+
+**Amanda Silver:** I would say, actually, that developers are one of the most empowered audiences across all of the audiences that we target. That they vote with their feet more than any other audience that exists, whether it's consumer, enterprise etc. Enterprise - totally different way that they drive decision-making. Developers - it's an end-user consumer audience that basically chooses things based on what's working for them.
+
+\[26:13\] And we see all the time developers picking up new coding editors, picking up new frameworks... They're technology enthusiasts. That's actually one of the things that makes the job so rewarding, is I can launch something at the beginning of the day, and by the end of the day I know if it was a hit or a dud, because I have so many early adopters that are kicking the tires. And I think that we don't have a captive audience at all with our developer tools. I think that developers have a tremendous amount of agency.
+
+**Jerod Santo:** Yeah.
+
+**Amanda Silver:** The way that we think about it is we have to win their loyalty every day with actual great product experiences.
+
+**Jerod Santo:** So this reminds me of a post I actually put in Changelog News today - I think Avdi Grimm wrote this - called "Developer Tooling is a Lousy Business." And he actually enumerated some of the points that you're making. You're saying it's great for us as developers, because we have the agency you speak of, and we're really, I guess, adept at handling our tools, and changing our tools... And that makes us somewhat of a fickle audience, because "What have you done for me lately?"
+
+**Amanda Silver:** That's right.
+
+**Jerod Santo:** And so in that sense, I guess, how is Copilot changing? Because I remember when it first came out and it was autocomplete, and it kind of was game-changing in that way, but it was really a non-deterministic autocomplete. That was good; it had its problems. And it has grown since then, and you guys continue to iterate and make it more and more awesome, and this year you're announcing a lot of it being a coding agent. Can you tell us all about it?
+
+**Amanda Silver:** Yeah. Well, so in a lot of senses Copilot has gone through the same epics that AI itself has gone through. Over the past couple of years the AI basically got to the point where it was good at doing token prediction, and things like that, but then over the last couple of years we introduced chat, and that allowed you to have a conversation with a knowledge base based on retrieval-augmented generation... And then just over this past year, it started to get to the phase where it could actually start to take actions, because the models themselves got to the state where they could actually reason over what they were working on.
+
+So for Copilot, we started with completions, and the completions at first were just a single line of code, and then they got to full function bodies, and then they got to longer, maybe a whole file... And then last year we introduced multi-file edits, so that you could actually make multiple changes to your codebase at once based on a basic prompt. We introduced chat capabilities, which could be based in the context of the project or the source code that you're already working in in the repo, and that kind of started to change the game... But I would say nothing has accelerated the capabilities as quickly as what we introduced with agent mode this past February, and then it's kind of rolled out over the past few months.
+
+But agent mode - it is shocking what you can accomplish in just a really short amount of time, just letting it go. Basically, what you do with that -- and it's also had its own acceleration. But what you do with agent mode is you go into the prompt, into the Copilot chat experience, and you switch it to agent mode, pick a model, whether you want Sonnet 3.7, or you want GPT 4.0, or whatever you want to use... And then you give it a prompt, like "Add tests for this particular project." And it will then iterate and self-evaluate as it's iterating.
+
+\[29:53\] So for example, in my demo earlier today, all I did was I gave it a source code repo for a website, and I said "Test this solution and write some tests, write some integration tests for me." And it basically started to do all of the automation to bring up the Playwright Automation Framework, bring up the website, start to traverse all of the different paths that it could go through in terms of the various different customer journeys on the website, and then it started to generate the tests themselves. And so what used to take me hours, maybe a half day, I can now get done in just a couple of minutes using agent mode. And so I think that's really changing the equation in terms of what these tools are capable of.
+
+**Jerod Santo:** Yeah, it's really ramping up the potential gains. One thing that I fear with that is the more you change, the harder it is to find the thing that went wrong. It's like the needle in the haystack problem, where if it's an auto-complete of a single line - cool, I can just see what it says. If it's a function, I can kind of read through that real quick, or maybe go in and make changes. If it's a file - okay, now it's getting big. But as it's like multi-file changes, with all of this stuff going on, and it's been thinking for four minutes, and I'm not sure what it was thinking... It'll tell me if I want to look at it, but you know, here's this big change we're going to make...
+
+**Amanda Silver:** Change log?
+
+**Jerod Santo:** Exactly.
+
+**Amanda Silver:** \[laughs\]
+
+**Jerod Santo:** And now I've got my code review step, basically, and there's so many things there that I just get a little bit apprehensive of like "Am I going to be able to find potentially where that one needle is in the haystack that maybe make the whole thing go haywire?" Are there guards, are there helps? Are there concerns that that might be an issue?
+
+**Amanda Silver:** I think that's something, first of all, that we need to think about in the design of the tool itself. We need to make sure that we are not introducing something that creates so much cognitive load for you to ingest in terms of change management that it's beyond your ability to reason over. So we are very intentional in terms of how we guide the prompts...
+
+**Jerod Santo:** Break it down?
+
+**Amanda Silver:** ...to kind of control just how much code is actually going to be generated. Other reasons that we do that as well. But I think the other thing is we really want to make sure that we are working with the workflow that you're used to. So another thing that we introduced today is the coding agent, where if you think about agent mode in VS Code as the experience that allows you to give it a prompt, and then you can do synchronous supervision over how it's completing all the tasks in that prompt... The coding agent is asynchronous. And so you can almost think about it as agent mode is your peer programmer, it's looking over your shoulder, it's accelerating your capabilities, but the coding agent is your peer programmer, where you can assign tasks to it as though it was another member of the team. And so it's just like in GitHub; if you're assigning an issue to your colleague, you would instead assign it to Copilot, and it can asynchronously go and execute that task and figure it out, come up with a plan and go and create a pull request and complete that task for you.
+
+**Jerod Santo:** Right.
+
+**Amanda Silver:** So what that I think enables is for you to then still think about the tasks that you're assigning to the agent in the same granularity that you would assigning it to another developer on your team.
+
+**Jerod Santo:** Right.
+
+**Adam Stacoviak:** So you could theoretically do a multi-file edit, build something, prompt it for that, and then run an agent against that and say "Debug this. Make sure it's sound code."
+
+**Amanda Silver:** Correct, yeah.
+
+**Adam Stacoviak:** So versus -- your question was "The fear of the code sucking", basically.
+
+**Jerod Santo:** Right. I've got my coding agent, and my code review agent.
+
+**Adam Stacoviak:** Well, don't worry, because at the end you can just unleash your agent, and just say "Agent, check that."
+
+**Amanda Silver:** \[33:47\] That's right. I mean, you could certainly have parallel adversarial agents kind of working against each other. You could set up that kind of system where you basically start to build your code, you then want to say "Okay, I'm going to have one agent that's going to be focused on the readability and maintainability of this code, and another agent that's going to be focused on maybe the performance and optimization of the code." And you could have both of those things kind of going at the same time based on -- really, at the end of the day, in a sense, you kind of can create these things by virtue of just providing the prompt. You just say "Oh, Copilot, I have a new issue. I want to optimize the performance for this particular page."
+
+**Jerod Santo:** Right.
+
+**Amanda Silver:** It's actually going to do it. \[laughter\]
+
+**Jerod Santo:** "It's actually going to do it."
+
+**Amanda Silver:** It's going to do it.
+
+**Adam Stacoviak:** I just think these little functions you can call... You know, "Check this, check that, write tests here, confirm..." That's really wild. And they do it.
+
+**Amanda Silver:** Yeah. And we also have, obviously, pull request reviews as a part of Copilot as well. So we also apply it at that kind of large scale for -- basically, every code change that we make inside of Microsoft, or open source repos as well, we do kind of AI-based code reviews.
+
+**Jerod Santo:** Right. I know you're using the classical data science form of adversarial there, but I kind of had a moment where I was like "It would be fun to just pit these agents against each other."
+
+**Adam Stacoviak:** I was, too.
+
+**Jerod Santo:** Like, "Listen, okay? This agent is not very good at their job, and they always get things wrong. Now, your job is to watch them, keep an eye on them..."
+
+**Adam Stacoviak:** Right. He's known to hallucinate...
+
+**Jerod Santo:** "You don't like them very much, you know..."
+
+**Adam Stacoviak:** I was thinking that, too. Could they either -- when will they get upset, and will they compete? Can you compete them? "You both have the same job. The one who does it better gets the task."
+
+**Jerod Santo:** Oh... \[unintelligible 00:35:38.27\]
+
+**Amanda Silver:** \[unintelligible 00:35:39.01\]
+
+**Adam Stacoviak:** Their job is to do the task and be excited about getting it right, and so they compete with the person or the other.
+
+**Jerod Santo:** They wanna get promoted.
+
+**Amanda Silver:** Well, I mean, I think that actually there's something to that. And I think one of the things for folks who are working on different machine learning models for the application to coding, we have benchmarks. Just like in the old days, when I was working on the JavaScript engine, we had performance benchmarks that we had to work on.
+
+**Jerod Santo:** Sure.
+
+**Amanda Silver:** Nowadays we have benchmarks for these SWE-agent kinds of models that are coming out. And so part of what we -- there's different kinds of techniques that you go through in terms of kind of getting the performance of the benchmark to be better, because you can optimize for different things. You can optimize for token consumption. So like "What's the cheapest way to accomplish it?" You could optimize for performance. "How can I complete the job more quickly?" You could optimize for accuracy. So in a lot of senses, what you're saying is actually not wrong. And I do think that over time, when we think about different competing agents that could actually go and fulfill your job, you could have ones that are experts in different types of tasks.
+
+**Jerod Santo:** Fascinating potential world. You mentioned every line of code, or a lot of the codes that we do here at Microsoft... That got me thinking about your own products.
+
+**Amanda Silver:** Yeah.
+
+**Jerod Santo:** And Satya mentioned code modernization. Of course, a huge opportunity, for these agents to do. I was just thinking about Anders and the team's port over to Go...
+
+**Amanda Silver:** Yeah.
+
+**Jerod Santo:** ...and how tedious that could be unless you have some agents doing the work for you. I'm curious about your old software projects. I'm thinking Visual Studio is pretty old at this point...
+
+**Amanda Silver:** Yeah, we have 20-year codebases, 25-year codebases...
+
+**Jerod Santo:** Are you harnessing these things to modernize things like Visual Studio?
+
+**Amanda Silver:** Can I say I'm glad you asked? \[laughter\]
+
+**Jerod Santo:** You can say that, sure.
+
+**Amanda Silver:** Well, actually, one of the things that we introduced today is allowing GitHub Copilot to help you modernize your .NET and Java code. So if you have a dependency on .NET 6, and you need to move to .NET 9, or you have a dependency on Java 8 and you need to move to Java 21, then you can actually use GitHub Copilot to help you do that. And this is a big deal, because it used to be that those kinds of jobs...
+
+**Jerod Santo:** The worst.
+
+**Amanda Silver:** \[38:10\] That's the thankless job that, as a developer, you hate that --
+
+**Adam Stacoviak:** Didn't somebody spend 10 years doing that stuff?
+
+**Amanda Silver:** Yeah.
+
+**Jerod Santo:** Of course.
+
+**Adam Stacoviak:** We just read the story...
+
+**Jerod Santo:** Yeah.
+
+**Amanda Silver:** That moment when your boss comes in and is like "We need to modernize the codebase, and I need you to work for six months on doing this port." Oh my God, that's crushing, because you don't get to do anything that's exciting.
+
+**Jerod Santo:** Yeah. It's just tedious.
+
+**Amanda Silver:** It's kind of tedious. And from our customers who are using it, they're telling us it takes care of 70% of the code migration. At least. So that's pretty incredible. We have success rates in the 90s of just upgrading your codebase. So I think that's one dimension of technical debt that this kind of stuff can take care of. Another is security vulnerabilities. And we're applying this internally at Microsoft very, very broadly. You can basically go look for CVEs, and once the CVEs are found, then we can actually go and file a PR to get it automatically fixed. And then all you have to do is review the code change.
+
+So my hope -- I think that the industry is still saddled with incredible amounts of technical debt. And if we can actually go and erase a bunch of that technical debt... Like, just imagine how much more innovation could happen.
+
+**Adam Stacoviak:** I got to thinking about this idea of the state of human velocity. We are now doing things that we wouldn't normally do. Not because we can't do them, but because they're just so time-consuming. Like porting a codebase to something else. We were spending a decade to do something like that. And now I think we have access to a tool that lets us think bigger. Not because it does the work, but because we can get past the hard things easier and sooner. So the speed of humanity, essentially, is... It's kind of like maybe the inflection point of speed for humanity... Because we've been going pretty slow. Since the 1900s you've got cars, and industry change, and at the end of the 1900s you have the explosion of computers and the internet... And in the 2000s, it's social media and all the things... And now it's like, AI is here to help us all go to a new plateau, faster.
+
+**Amanda Silver:** Yeah. I think in a lot of senses, we've been struggling -- I wouldn't say struggling. I would say limited by the available developer talent in the world. We still have a shortage of developers, in my opinion. A significant shortage of developers across the world. And especially developers who have higher level systems thinking and reasoning capabilities. And so I think that, in a lot of senses, what this all represents is an opportunity to spread that knowledge a bit more broadly. And think about all of the apps that your organizations wanted you to write, but you never got to, because your backlog was so long.
+
+I think there's huge amounts of demand and need in the industry overall that is just not getting addressed today, because developers are settled with technical debt, and what they already have on their plate is significant enough to occupy them. And so if we can actually erase a bunch of that, I think that's going to make a huge difference overall.
+
+But I think the other thing that's super-important in all of this is there are aspects of the developer job that are not awesome, technical debt being one of them. Another thankless task is being on call, and responding to live site incidents in the middle of the night. I know that a lot of our developers do not relish that aspect of their job.
+
+**Jerod Santo:** Sure.
+
+**Amanda Silver:** \[41:47\] And I think that a lot of the new capabilities in AI allow us to offload, at least the less complex cases of site reliability engineering response to agents. So one of the things that we talked about today is we introduced a new site reliability engineer SRE Agent that can actually deal with - if my app suddenly starts to be unhealthy, it can actually go and do a profile, and understand if it's a memory issue, and even start to auto scale your infrastructure to be able to respond and mitigate the issue. It may not be the permanent right repair item or fix - that might come a little bit later - but it can do that first line of response. And for us inside Microsoft, we've actually been applying it internally significantly and have dealt with a ton of -- like, hundreds; I think we're probably at thousands now, of incidents that have been managed in this way... Such that developers never needed to get involved. And then they just review the repair items from the recommendations from this autonomous SRE agent.
+
+So I think that's pretty cool, because that -- there's aspects of the developer job that's all about creation, and you love that moment where you when you get to write new code, you get to scaffold out new... I always loved scaffolding out new class libraries, and frameworks, and just doing the first rough-in of the application. That was always my favorite part. But this means that developers just -- they don't have to get woken up in the middle of the night, and their job doesn't have to be awful.
+
+**Adam Stacoviak:** It's like self-healing, and I imagine that - you said the fix may not be a permanent one; it's more focused on uptime.
+
+**Amanda Silver:** Yeah, exactly.
+
+**Adam Stacoviak:** "My goal as this autonomous agent is not the best long term fix. It's keeping the application up."
+
+**Amanda Silver:** Yeah. I mean, SRE agents - SREs in general - are always focused on application uptime, and mean time to mitigation. So if there is an incident, how long does it take for them to actually respond to it. And then they also are concerned about things like costs and operations over time... And I think they can much more meaningfully contribute to how to build healthy, large-scale systems that operate well.
+
+**Jerod Santo:** So are these agents actually applying the fixes, and the person doesn't have to come in and hit that button that says "Yeah, let's go ahead and do this"? Did I hear that right?
+
+**Amanda Silver:** Correct. Yeah. I mean, it's all within policy. You can decide like what limits you want it to have. But if you need to scale your infrastructure up to be able to handle more memory, for example, it could deal with that automatically in the middle of the night, without having to ping you and wake you up. And then you come in in the morning and it says "Hey, we had this incident. We had this out of memory exception that happened here", and then you can go investigate it in the morning and go figure out what's the long term fix.
+
+**Jerod Santo:** Yeah.
+
+**Adam Stacoviak:** You could almost have it do the fix and another agent to check the fix.
+
+**Amanda Silver:** That's right.
+
+**Adam Stacoviak:** I mean your policy, you alluded to the policy. The policy essentially is an augmentation of potentially an agent that has different parameters.
+
+**Jerod Santo:** You're trying to take us out of the job, aren't you?
+
+**Adam Stacoviak:** I mean, I'm just saying, that's just the -- that's where it's going.
+
+**Jerod Santo:** \[laughs\]
+
+**Adam Stacoviak:** Not necessarily, but that's what you can do, though. You can have \[unintelligible 00:45:28.20\] all the time. You can have an agent that just checks it to confirm based on policy.
+
+**Amanda Silver:** That's right.
+
+**Adam Stacoviak:** Within these bounds, you have agency to do this thing.
+
+**Amanda Silver:** Correct. Yeah.
+
+**Adam Stacoviak:** Without all those bounds, it is a no.
+
+**Amanda Silver:** Yeah. When we talk about like what are the skill sets that developers are going to need to have for the future, it's still the same complex systems reasoning. When you think about "If I have multiple policies that I am applying to how to manage infrastructure during a live site incident" - that in and of itself, that set of policy rules, that is a complex system. And you have to think through "Well, what happens when this rule conflicts with that rule? How is the system going to respond?" So I think that's where a lot of our brain time is going to start going, is thinking about how these different kinds of systems and agents that are somewhat autonomous are going to interact.
+
+**Break**: \[46:29\]
+
+**Adam Stacoviak:** Are you thinking about how these agents manifest as a visible layer to this agentic internet and web that was alluded to? Because I'm thinking like --
+
+**Amanda Silver:** I love that.
+
+**Adam Stacoviak:** There's this idea of agents available, and so why recreate the wheel...? I was really just thinking about the idea of a secret agent. \[laughter\]
+
+**Jerod Santo:** Why were you doing that, Adam?
+
+**Adam Stacoviak:** I was just like "That's a really cool name for an agent that doles out secrets", maybe. Or just deals with authentication and authorization kind of thing. If there was an agent, if there was a secret agent...
+
+**Amanda Silver:** That's cute.
+
+**Adam Stacoviak:** And I would want to discover that secret agent.
+
+**Amanda Silver:** That's cute. Okay.
+
+**Adam Stacoviak:** Anyways...
+
+**Amanda Silver:** Well, a couple things I would say about that. First of all, yes, I think in terms of thinking about the common way that you can go interact with all of these different agents - yes, I do think that ultimately, our goal is that GitHub Copilot can become that common substrate. It's almost like the omnipresent agentic command center that allows you to interface, no matter if you're talking about your infrastructure, or your code, or your test, or even tasks that I have to do to go work through the bureaucratic layers of our ops team, or whatever it is... I think a lot of that can start to become interfaced through working with GitHub Copilot. So yes on that point. I think what you're bringing up around secrets is a great question, right?
+
+**Adam Stacoviak:** Sure. I was just wanting to call it a secret agent, though. \[laughter\]
+
+**Amanda Silver:** No, but here's the thing, though. Here's the thing. I have kind of two jobs at Microsoft. One is I'm head of product for our developer division, and the other job is I'm basically the GM for our platform engineering team. And so that means that basically we build all of the tools and all of the policies for all of the internal engineering teams at Microsoft. So we kind of take all of our third-party products and we host them and administer them and extend them and incubate in them for our first-party engineers. One of the big things that we've been trying to focus on is expunging secrets from our codebases, because secrets are dangerous.
+
+**Adam Stacoviak:** Very dangerous.
+
+**Amanda Silver:** And if you have them in your codebases, then if you have malicious actors, they can go in and try to exfiltrate your code and get your secrets, and then get access to your infrastructure. So generally, we are trying to move towards a system where we do not have secrets checked into our codebases. That said, that's a great application of the kind of policy that can be applied at your organizational level, to say "Look, if I have any code that looks like a secret, I want you to flag it, I want you to file an issue, because I want that to be manually checked." And we also have in GitHub Advanced Security detection of those kinds of secrets as well, so that we can actually make sure that you never push it into your codebase.
+
+**Adam Stacoviak:** Well, then I propose that you make that a product. \[laughter\]
+
+**Amanda Silver:** Done, sir.
+
+**Adam Stacoviak:** And you call it Secret Agent, and you credit Adam Stacoviak as the idea for the \[unintelligible 00:50:46.28\]
+
+**Jerod Santo:** Just in the fine print at the bottom...
+
+**Adam Stacoviak:** I just like that. It's so cool to have that. Make that a thing.
+
+**Jerod Santo:** Make that a thing... Now you have three jobs at Microsoft. The third one is to get that thing named Secret Agent.
+
+**Amanda Silver:** I'm working on a Secret Agent.
+
+**Adam Stacoviak:** It's catchy.
+
+**Amanda Silver:** I think it's called GitHub Advanced Security.
+
+**Adam Stacoviak:** Well, there you go. It's a cool name, but...
+
+**Jerod Santo:** A.k.a...
+
+**Adam Stacoviak:** ...it's not a brand. Secret Agent...!
+
+**Jerod Santo:** Yeah, if you whisper it, it sounds even cooler.
+
+**Adam Stacoviak:** That's right.
+
+**Jerod Santo:** Oh, my. So do you think about cascading effects? Especially - I'm just back on the SRE side of things, and thinking about turning over so much to...
+
+**Amanda Silver:** Control.
+
+**Jerod Santo:** Yeah... To software agents... Which, again, are - what did Nathan Sobo call it?
+
+**Adam Stacoviak:** A genius golden retriever on acid.
+
+**Jerod Santo:** On acid, yes. Which - maybe those are the models he's working with. But you know, a thing that you don't ultimately know what it's going to do. Now, you can train, and fine-tune, and guard, and watch agents watching agents all you want... However, we've seen at scale the internet operating, distributed systems at scale - things go wrong in ways that are sometimes very quick, very catastrophic, and compounding and cascading.
+
+\[52:09\] One thing that I think about is some of these stock market trades, when you have quick corrections or crashes is because you have software making margin calls and trading with software programs, and eventually what happens is the New York Stock Exchange actually just stops everything and is like "Let's chill out here."
+
+**Amanda Silver:** Yeah.
+
+**Jerod Santo:** And I think there's a potential of that kind of thing with agent SREs, changing the memory on your VM, and then this happens...
+
+**Adam Stacoviak:** Like a race condition amongst agents, or something like that.
+
+**Jerod Santo:** And I'm wondering -- I know you all think about security a lot, and... What kind of stuff is out there for just making sure that maybe there's a pull-the-plug moment, or a way that you can get back in the loop and say "Okay, let's just chill out here, guys."
+
+**Amanda Silver:** That's a fantastic question.
+
+**Adam Stacoviak:** Guys... You called them guys.
+
+**Jerod Santo:** Well, just a team, you know...
+
+**Amanda Silver:** So I think generally what our approach is is that we want to make sure that every agentic workflow is completely auditable, so that we can see what the agent is actually executing, look over it in history; that we do have controls over it to be able to say "These are the resources that you have access to." And we also have to think about things, like ways of testing the models themselves, whether it's a model that we build, or it's a model that you build for your software... And that's part of what we build with the AI Foundry, that allows you to actually evaluate the models against all different kinds of checks. Whether that's safety and security checks, whether that's responsible AI kinds of, harassment kinds of scenarios, or language that's inappropriate... A lot of that is really what has to get built as a part of kind of building and evaluating models themselves.
+
+And then we also want to have this common agentic control layer across all of the software that uses agents, so that we can see what those agents are actually actively doing, what resources they have access to, and restrict what they have access to if they start to go astray.
+
+**Jerod Santo:** Right.
+
+**Amanda Silver:** And we'll be showing a demo of that tomorrow.
+
+**Jerod Santo:** Oh, cool.
+
+**Amanda Silver:** Yeah.
+
+**Jerod Santo:** Yeah, it's really just a big orchestration problem at the end of the day, and then underneath it you have your -- what do you guys call them? Not the frontier, but the Foundry.
+
+**Adam Stacoviak:** Foundry, yeah.
+
+**Jerod Santo:** \[unintelligible 00:54:24.19\] the models themselves. And I think it's really cool how much choice is available...
+
+**Adam Stacoviak:** So much.
+
+**Jerod Santo:** ...at the model level. Because I've even, in my personal use and in my coding use, have appreciated the ability just to swap these different ones, especially as they leapfrog each other in capabilities...
+
+**Amanda Silver:** Yeah.
+
+**Jerod Santo:** ...and I think it's really cool how many different models are available.
+
+**Amanda Silver:** Yeah. I mean, that's one of the things that I think has been really great about kind of bringing GitHub models into -- or bringing the AI Foundry model catalog into GitHub models, is it allows developers to be able to go kick the tires on all the different models that are out there. And they all have different kind of strengths and weaknesses, and in some senses I think about it as a search space of different model characteristics, that have a certain price and a certain performance. And you just need to kind of go find what's the right one for your particular use case.
+
+And I think that the fact that we've integrated GitHub models -- sorry, an AI Foundry model catalog into GitHub models really does allow developers right in GitHub to go test these different models in a playground. And then further, because we have in VS Code the ability to select models as a part of your chat experience, we also find a lot of developers using that chat experience as a way to go test which model is meeting their particular needs for their use case, that they then go write into the application that they're building.
+
+**Jerod Santo:** That's awesome. No further questions... \[laughter\] I don't have any questions about that. I just think it's cool. And you're not the only ones who's doing that. A lot of people are, and I think it's great. I love just giving developers choice, versus saying "No, you're going to use this. It's the best one. Trust us."
+
+**Amanda Silver:** \[56:10\] Always. Always.
+
+**Jerod Santo:** Don't do that to me. That being said, which one's the best one? \[laughter\]
+
+**Adam Stacoviak:** Well, I'm waiting for .agent to become a TLD.
+
+**Amanda Silver:** Oh, yeah.
+
+**Jerod Santo:** That'd be a good one.
+
+**Adam Stacoviak:** I mean, .ai is cool, but...
+
+**Jerod Santo:** .agent.
+
+**Adam Stacoviak:** .agent. Because then you'd have the secret.agent.
+
+**Amanda Silver:** Yeah.
+
+**Jerod Santo:** It's all for this one thing.
+
+**Amanda Silver:** Domain registered. Yeah.
+
+**Jerod Santo:** Yeah. That would be a good TLD, because I think there's going to be -- they're going to be selling agents like you're selling apps at this point. Don't you think?
+
+**Adam Stacoviak:** Well, if there's a company well-positioned to sponsor in some way, shape, or form... Because we just learned about DNS with Anthony Eden, and how TLDs come into play...
+
+**Jerod Santo:** Right. You've gotta have lots of money to host a TLD.
+
+**Amanda Silver:** Yeah, there is a whole namespace question - how do you find all these different agents? How do you figure out which one you want to deploy for this particular problem? And I think that's one of the big next things. Both agents and tools, there's going to be -- just like there's a catalog for models, there's going to be catalogs for agents and for tools.
+
+**Adam Stacoviak:** And the best way to catalog them is .agent.
+
+**Jerod Santo:** \[laughs\]
+
+**Adam Stacoviak:** It really is. I mean, if you're building an agentic web, this open agent web that's happening... .agent.
+
+**Amanda Silver:** Okay. I'll take it to the top.
+
+**Adam Stacoviak:** Two more missions... I hope you don't mind.
+
+**Amanda Silver:** Alright.
+
+**Adam Stacoviak:** And just come back to me for any final sign-off on these ideas, and stuff.
+
+**Amanda Silver:** Love it.
+
+**Adam Stacoviak:** I can give you more.
+
+**Amanda Silver:** Love it.
+
+**Jerod Santo:** So the advancements every year are interesting... It's moving very fast. Here we are, '25... If it were up to you, '26, the three of us sit down... You don't have to unveil any secret roadmaps or anything, but what would be going on next year this time if you were excited about it? What would be the next step? We're at agents... Where are we next year?
+
+**Amanda Silver:** Well, first of all, I think the agents -- we've seen with agents just how powerful they are, and so we've seen a lot of promise. I think there's also a lot of peril in there as well.
+
+**Jerod Santo:** Yes.
+
+**Amanda Silver:** And what's going to start happening is that -- basically, a lot of folks are using them without real security controls. And so I think we're going to need to see more ways, to your point earlier, around "How do I audit and control all of these agents working throughout my enterprise or my team?" So I think that's one thing.
+
+**Jerod Santo:** Sure.
+
+**Amanda Silver:** We're going to start seeing a lot more controls in that sense. The other thing is, part of what we're seeing - and this has really only accelerated over the last three, four months as we've started to have capabilities like agent mode in VS Code, is the capabilities of the software development team are also changing. Developers now can do better designs. They can make things not just prettier, but more easily usable without having to have designers involved. Designers can code. Product managers can now code. So what does that mean for how we think about the evolution of the software development team, and what canvases do we think we need to do collaboration?
+
+**Adam Stacoviak:** I think of it like giving somebody who -- the whole team, or the whole group of people you just mentioned, they all speak a language. Let's just say it's English, for just lack of better terms for this analogy. I think of it like these folks have a limit of vocabulary, and these folks have a limit of vocabulary. And they're all specialized. And now they have a more shared language spectrum.
+
+**Amanda Silver:** Exactly.
+
+**Adam Stacoviak:** Because you have the words, we have the words... We can share the words. That's called speaking, of course, as you know... But I feel like that's what it does. You now give them -- they all spoke English already, because they all understood some of the code, some of the design, some of these different features... But now they can all speak a certain language.
+
+**Amanda Silver:** \[59:53\] I completely agree with you. But I think that part of what's going to happen is that everyone is going to start contributing to the codebase more easily.
+
+**Adam Stacoviak:** Which is better for the product. And the user, obviously...
+
+**Amanda Silver:** I think so, too... But I think we're already starting to see designers contribute code. In terms of like rather than handing a design over the wall, and say "Engineer go implement it", now the designer can actually go and, if they want rounded corners, they can get rounded corners to happen. Or if a PM has an idea around a new feature that they want to experiment with, maybe they can go build an initial prototype with that, without having to go bother the engineers to go get that done. And I think that starts to kind of raise interesting questions around "How do you think about architecting different kinds of systems?"
+
+I think maybe one of the things that might happen is that there's a difference between solutions that are more sitting at that SaaS, level versus services that are more at that systems level. I think that that SaaS level is going to start to be something that is more easily extended. Today, when we think about something like a Microsoft Office, or M365, or we think about something like GitHub itself, it's essentially an end user experience, a SaaS that is being provided and has extensibility points. And those extensibility points are painstakingly crafted, because every time that you expose a new API that allows you to customize the environment, it both empowers your users to go build other things, but it also represents kind of a boat anchor that you're stuck with the back compatibility for that contract.
+
+But I think that what this is now kind of enabling is for you to start to build other systems, other automation on top of other systems much more easily... Even without the API, to be frank, because now it can just traverse the DOM, or whatever it needs to do... And I think that means that a lot more software becomes a lot more extensible. And I think that that means also that the way that you build software itself is going to end up changing over time.
+
+The other thing that I think is super-interesting is we also spend a huge amount of time right now on the view, creating the view.
+
+**Jerod Santo:** Sure.
+
+**Amanda Silver:** Like an MVC kind of model...
+
+**Jerod Santo:** Yeah, totally.
+
+**Amanda Silver:** I think in the future, it might just be that you focus on the model, and the view you actually go and specify with a design system.
+
+**Adam Stacoviak:** Well, I saw a T-shirt downstairs, it had a bunch of things crossed off. I think developer was one of them, and something else, and then it just said "Builder".
+
+**Amanda Silver:** Yeah.
+
+**Adam Stacoviak:** And so I feel like y'all have been ahead of the times, basically, with I think people becoming builders, rather than just developer, or just designer, and the word 'just' not pejoratively, if that's even a way to say it... But more so builder. Everybody's a builder, and everybody can contribute.
+
+**Amanda Silver:** Yeah. Love it.
+
+**Jerod Santo:** I have a question that I think maybe would have been way better back in the VS Code part, but as we wrap up --
+
+**Amanda Silver:** Spicy?
+
+**Jerod Santo:** I don't know if it's spicy or not, but I'm curious... We were kind of debating VS Code's foothold, and is it a strategic advantage, and this and that... And the open sourcing of that, and that it's been a good thing for Microsoft... What we've seen recently in the vibe coding space is VS Code forks.
+
+**Amanda Silver:** Uh-huh.
+
+**Jerod Santo:** And these forks come up fast and furious, and they're getting huge valuations...
+
+**Adam Stacoviak:** They sell for billions.
+
+**Jerod Santo:** Right, they sell for billions. They're getting large user bases very quickly, or at least it seems like they are... And in a sense, that's a little bit of a backfire, because now you're providing VS Code, this platform, which is just forkable, and now you're giving competitors an opportunity to just catch up real quick and compete with Copilot in weeks. I mean they can vibe code their way to a $3 billion valuation. Does that bother you? Is that -- do you even see it? Are these just like the the cockroaches that you just say "Get out of here, cockroach"? Or how do you guys view it? How do you view it?
+
+**Amanda Silver:** \[01:04:01.02\] No, I mean I --
+
+**Adam Stacoviak:** You Kelsey Hightowered that one. \[laughter\] I love you, Kelsey...
+
+**Amanda Silver:** I think that there's real real innovation that's happening in the industry, and a level of competition that we haven't seen for a long time. So I have a tremendous amount of respect for you know the competition that we're seeing right now in the code editing space.
+
+I will say, like - yes, I think that there is a lot that has been built on top of the open source codebase of VS Code, and I think that is creating a foothold that allows others to kind of go and create additional, you know, features or differentiation on top of it... But that's I think one of the reasons why we've decided to actually open source the GitHub Copilot extension for VS Code, and build it directly into the VS Code codebase. We really do think that the AI experience is now table stakes for any code editor. And in the same way that VS Code has been open from the get-go, we think that now the AI capabilities in VS Code also need to be part of the open source codebase.
+
+And we certainly believe in, you know, whether it's VS Code, or TypeScript, or anything that we do in like the Azure SDKs, or .NET - all of that's open source. Half of what my team works on is done in the open and open source. You know, we certainly see that especially when you're trying to build a community around a technology, an ecosystem around the technology, building in the open actually creates better products. You know, it allows more people to contribute, whether they're contributing pull requests, and code contributions, or if they're just logging issues and they care enough to actually follow through with really great issue descriptions. That's an important way to contribute to the codebase, and we see that across all of our open source codebases.
+
+And I think what we hope to see now is - there's a tremendous amount of innovation that's been happening in AI-based coding. What we hope to see is more of the community contributing back to the VS Code codebase to really advance the state of the art for everybody.
+
+**Jerod Santo:** Yeah.
+
+**Adam Stacoviak:** I'm not sure I would call them competitors, though...
+
+**Jerod Santo:** You don't think Windsurf is a competitor of Copilot?
+
+**Adam Stacoviak:** I would say --
+
+**Jerod Santo:** I'm going to use one or the other, aren't I?
+
+**Adam Stacoviak:** I guess so... But is that your customer, for a lack of better terms?
+
+**Jerod Santo:** Oh... They want all the customers? Don't you?
+
+**Adam Stacoviak:** I don't think so. I think --
+
+**Jerod Santo:** Let her answer... \[laughs\]
+
+**Adam Stacoviak:** Well, this is what I assumed you were thinking. And you said a lot of things, but you didn't say this. But if I read between the lines and I think - if I were you, I wouldn't see it as competitors, because you are focused on developers using VS Code, and I don't think they're not not developers, but if you're vibe-coding, it's not a developer action; it's a different way to get the end result.
+
+**Amanda Silver:** Yeah. Right --
+
+**Adam Stacoviak:** A developer writes code and cares about the code, whereas the other way is -- it's not so much less about the code, it's a different way to get there. That's how I look at it.
+
+**Amanda Silver:** So I think I think there's kind of two pivots to that. First of all, I think that code editors, generally, if your primary task is to write code, then I would say any of the popular code editors are competitors, in a sense. But to the point earlier that we were talking about, we believe that we live or die by day to day product truth, and we have to basically win every developer based on their usage and experience with our products. And we strive to make the best products we possibly can.
+
+I think to your point around vibe coding and how does that relate - I see vibe coding as a really interesting evolution. It's not quite --
+
+**Adam Stacoviak:** I'm throwing zero stones at that. I'm a agnostic when it comes to all innovation. Whatever gets us to the next place, we're gonna love it. That's what I'm for.
+
+**Amanda Silver:** Totally.
+
+**Adam Stacoviak:** And if vibe coding is one of the ways we get there, or it invites more people in to build software - cool.
+
+**Amanda Silver:** \[01:07:59.15\] I think vibe coding invites a lot of people in to go build more software. I think that it's not your typical pro dev software developer, but I think that what vibe coding has kind of started to create is more of this pattern of what I would call like natural language-driven development... Which is like starts with maybe a vibe initial prompt, but then it evolves into --
+
+**Adam Stacoviak:** Something \[unintelligible 01:08:24.07\]
+
+**Amanda Silver:** ...a full spec. And the spec is still written in natural language, it's not written in C\#, or TypeScript, or JavaScript, or anything like that. It's written in a natural language; English in my case, but it could be written in whatever. But then you take that spec, and then you use that spec as the prompt. And that allows you to then iterate to this level that's like, you can get to a much more sophisticated first implementation of the code that you're aiming to implement. And then you continue to iterate, and you may even modify the spec, as opposed to modifying the code. So I think that's starting to change things.
+
+From there, then it's like, okay, well, now that PM can contribute even a spec for a feature, or they can contribute a spec for the initial product... The active testing could also be, in some senses, a large prompt. The designer could contribute a design system into the codebase... And so I think what we start to see is that over time the codebase is not just everything that builds; it's all of the prompts...
+
+**Jerod Santo:** It's all the metadata.
+
+**Amanda Silver:** ...for all of the systems, and all of the different, you know, phases that you go through of the software development lifecycle.
+
+**Jerod Santo:** Yeah. I think that's really insightful, and one of the things that I've noticed as an experienced developer trying to adapt and adopt the tools is that software isn't built in like a chat scenario. Like, you don't chat your way to a software system, because there's just like so much chatting that goes on -- you know, you design specifications... And yeah, you may have conversations that lead to that design decision that you make... But once you make that, you don't want that to be like one moment in a conversation that was way up here and you've gotta tell your coding agent to scroll back and "Remember what I said back then?" You want to actually have tangible outputs.
+
+A spec that gets created through - whether it's a vibe session, or just a peer programming session... Now I have this written document that evolves. And it would be so cool to be able to take that spec and be like "Alright, here's a different model." You know, start fresh. We don't need to use the code. We have the spec. And you can take the same spec, six different models, write the program. Maybe take the best one on each, or whatever it is... Have something that you could like start from, so you're actually building out an architecture. Is that formalizable?
+
+**Amanda Silver:** Yeah, it is. I mean, we now are starting to do more spec-driven development. We have .prd files that, you know, would be the description of the spec. And those kinds of things are starting to get checked into the codebases. But I think also there was a nugget in what you were talking about in terms of like design decisions that are made throughout the process. Sometimes it's not just the spec that you actually use...
+
+**Jerod Santo:** It's the why.
+
+**Amanda Silver:** Well, it's the conversation. Like, if you think about, you were having a conversation with your designer, or another engineer on the team, in terms of how something should work... Maybe over time the history of that conversation should actually be something that's persisted into the codebase, in some senses.
+
+**Jerod Santo:** Yeah. And hopefully in a summarized fashion, so it's actually grokkable...
+
+**Amanda Silver:** Yeah, exactly.
+
+**Jerod Santo:** I know there's some people that keep actual -- I can't remember if it's called a why document... But it's basically around their decisions. Not the decision we made, but why we made the decision. And so you can go back to that and be like "No, here's why there's this fence here."
+
+**Amanda Silver:** Exactly.
+
+**Jerod Santo:** You see a fence, you're like "Why is the fence there? It doesn't need to be there." It's like, "Well, there was a reason... And none of us know why. It's lost to history." But now you can go back to that chat or that context, and at least link it somehow, whether it's summarized, or linkable, or whatever it is, to be like "Here's our spec. And then this part of the spec here - why is it like that? Well, here's the why."
+
+**Amanda Silver:** Yeah. Maybe in that world developers don't have to write documentation.
+
+**Jerod Santo:** Now you're selling me.
+
+**Adam Stacoviak:** Oh, my gosh... \[laughter\] Or they write one. They write one, the initial spec.
+
+**Jerod Santo:** That's right.
+
+**Adam Stacoviak:** Yeah, I was thinking about that, too.
+
+**Jerod Santo:** Alright, awesome. This has been a blast.
+
+**Amanda Silver:** Yeah, thanks for having me, guys.
+
+**Jerod Santo:** Thanks so much for sitting down with us, Amanda.
+
+**Adam Stacoviak:** Very fun. Thank you.

Werner Vogels predicts the future (Interview)_transcript.txt

@@ -0,0 +1,315 @@
+**Adam Stacoviak:** Well, everyone, we have a treat. We're here with the CTO of Amazon. My gosh, talking about predictions... Every year you have five; they mostly come somewhat true in a couple of years, so you're pretty accurate, Werner... Welcome to the show again. The last time you were here was for the NoSQL Smackdown back about 12 years ago, Jerod, 10 years ago was it?
+
+**Jerod Santo:** Forever ago. Episode 18.
+
+**Adam Stacoviak:** Back when NoSQL was cool, fresh and burgeoning... Now it's ubiquitous and everywhere, and I guess still cool... Coolish. Coolish.
+
+**Werner Vogels:** Engineering has come a long way since. I think the first NoSQL tools were really the first tools. Now if you look at DynamoDB, if you look at Mongo, these are super-robust, highly scalable tools, where everybody can build a business on.
+
+**Adam Stacoviak:** Come a long way. I think it was burgeoning. DynamoDB was coming out then, it was fresh and brand new... I guess as CTO, you probably architected a lot of that stuff. What do you do as part of your role? What is your role maybe then, 15 years ago, and then how is it now?
+
+**Werner Vogels:** So first of all, it's 21 years now, so that's a bit... And there is a predecessor to DynamoDB. That's Dynamo. And one of the things - when I joined Amazon, I joined as what was called a director of systems research. The idea was to bring... Let me go a little bit back. Think about 1994, when Jeff Bezos starts thinking about sort of the internet, and he starts a bookshop. He doesn't really want to start a bookshop. He's just fascinated by the internet. What are the things that you can do online, that you will never be able to do in real life? And he just picks a bookshop. A good bookshop has 40,000 titles in stock, yet there's millions of books out there. So he thinks you could do that online. But there is no book. There is no software that you can buy. There is no book that says "And here is an e-commerce operation." The word e-commerce doesn't exist yet. So everything that the Amazon engineers had to do to build Amazon was to invent everything themselves. Because the kind of technology that they could buy couldn't operate at that scale. We've had a number of bloody noses because of that.
+
+So when I joined Amazon, engineers at Amazon were brilliant at scaling, but from a, let's say, practical, lots of scars kind of approach. And Jeff hoped that by bringing a former academic, you get some more robustness. You get a better fundamental approach to scale, and reliability, and things like that. And we did lots of large projects, around removing all single points of failures, or how to best measure -- how do you measure, what does it mean to measure? If 50% latency on your webpage means nothing... well, it means that 50% of your customers are getting a worse experience. You need to know how much worse.
+
+If I think about CTOs, there's sort of four or five different types. And there's no real well-developed, well-described one. First of all, there's the data center manager. He reports up to the CIO. Then you have the CTO, that is the second person in the startup, often the co-founder, the first coder... And then you have the big thinker, sort of the role that I got when I came into Amazon; sort of like look at the whole picture, what other kinds of things we need to do. But then, when we started doing AWS, your world changes. You go from an internal-focused CTO to an external-focused CTO. And I think Scott Dietzen, who was at BA, I think it was that at that time, he called it really an external technologist. The ability to talk to your customers, look at "How are they using my products?" and "What are the problems that I see with 10, 20, 100 of my customers, that they may not see as a single problem themselves, but if I can build a tool for that, I can really help my customers?"
+
+\[08:15\] So your world changes from being purely internal to being external, and sort of bringing back things into -- and then over time, I've become more and more interested in those organizations, both profit and non-profit, that try to solve hard problems. And with hard problems, I mean hard human problems.
+
+The United Nations expects that by 2050, we have 2 billion more people. I mean, how are we going to feed them? How are we going to make sure they have an economic future? How are we going to make sure they have healthcare? Those kinds of problems are the ones that I mostly focus on today. On one hand - you know, by the way, we build technology at Amazon, but also, by looking for those often young businesses, and how can we support them? Take, for example, the ocean cleanup project. It's a massive problem that the Grand Ocean garbage patch is full of fishnets, and plastics, and there's about 30 rivers that sort of contribute mostly to that. So these guys have built plastic with GPS in them, threw them in the river and see them where they end up. Or they have these AI cameras on boats, to first see where do these boats go, but secondly, also what are the things that they're seeing? And so working with those kind of customers is extremely satisfying, because we're solving actually real human problems. We're not building spam filters.
+
+**Jerod Santo:** Well, 21 years is a long time... And there's that old quote, "The best way to predict the future is to create it." You've spent 21 years with your teams, creating in many ways the future that we're currently living in, but you're also out there predicting it. This is the fifth year perhaps you're writing these predictions... You have five of them, and they're fascinating. I agree with most of them. The companionship one I'm skeptical of, but I'd love to hear your full case for us here. But why write these? Why predict the future in these ways, and continue to do so?
+
+**Werner Vogels:** There's a long history by large technology companies, and if you think about IBMs and the locals and whatever, to sort of -- or McKinsey, or whatever, the consulting companies... About kind of predicting the future. And always when I read them, I always felt that they were a bit self-serving. And the things that I saw in the real world were different. I see significant human problems, and some of them are our own nature, but quite a few are caused by technology, or the way that we use technology. There's a leadership principle at Amazon, "With success and scale comes broad responsibility." That means you don't only have a responsibility towards your shareholders to make money, you also need to serve your customers... But at AWS, of course, that is builders. How can we help builders best?
+
+\[12:02\] And then, you know, I write these predictions about things I see around me. I have the fortune to travel the world. I was just in - or just... In spring I was in Sub-Saharan Africa. So Nigeria, Rwanda, Kenya. Now in Kenya, we were in Nairobi, and of course, you stay in the center of Nairobi; beautiful buildings, whatever. You drive 15 minutes and you see how the real Kenyans live. And if you think how they work, most of them are day laborers. That means that they go to the bank, they get $2, they buy something, they try to sell it, they bring the $2 back, and hope to have 40 cents left to buy food. They probably have that, but then they don't have money left to cook it. Why? Because these big canisters, these butane canisters cost $10, and they don't have $10.
+
+And so there's a young company called Coco Networks, that built a sort of ATM where you can go with a canister, you plop the canister in and you say "Give me 15 cents of gas." And then you can take that home, you can cook your food. Now, this is not a world shocking problem, but it is solving really hard problems that people have. And those kinds of companies are really the ones that -- and those kinds of problems are often kind of the things that I like to service through the predictions. And there are things every year... I mean, I could have done 10 predictions.
+
+**Adam Stacoviak:** You could have, yeah.
+
+**Werner Vogels:** but that doesn't work that fairly well.
+
+**Jerod Santo:** \[laughs\] Five's the magic number...
+
+**Adam Stacoviak:** Three's the magic number, but five's better.
+
+**Werner Vogels:** yeah, but... Was it two years ago? An example, a topic that has become more and more openly discussable is menopause for women. It's a major problem for them. But even mothers don't talk to their daughters about it. However, that is slowly changing. What kind of things do we see in the startup community that are trying to build - whether it's technology products, or whether it's medicine, or whatever, to solve those kinds problems. So I also try to talk a little bit about things that people normally find a bit harder to talk about.
+
+**Jerod Santo:** Well, let's talk about one. Companionship, loneliness... Robots. I do not have any disagreement whatsoever with your casting of the problem, which is that loneliness is on the rise, especially amongst elderly, but even across all demographics we see an increase in loneliness, which has all kinds of health and mental wellness problems... It's just overall bad. But you seem optimistic, because we have a new swath of robotic companions that are coming out, and...
+
+**Werner Vogels:** Well, it's one of these problems... So I did this documentary series called "Now Go Build", a TV series. Probably two, three years ago, I was in Japan for that, and that's where I first really dove deep in this particular topic. So Japan really was a society where kids take care of the elderly. The grandparents live with the grandkids, and the kids. And that was normal. But the younger kids want to start making career, and there is a clear shift in Japanese society happening, where the elderly are no longer taken care of by the younger people. And as such - you know, that's a shocker; not for the young kids. It's a shocker for the old people. They're really on their own... And indeed, I think that there are a lot of technology solutions possible to help these people.
+
+\[16:34\] One of the things -- and loneliness is something we'll get to, but anything we can do as technologists to keep people longer out of hospitals, or care homes, or whatever, the better it is. If that means that -- a company in Japan that I met called Z-Works - they try to do all these innovations around these elderly people, to help them. Make in your mattress a little bit of a sensitivity pad, and maybe in the middle of the night you go to the toilet, and you don't come back within 15 minutes - the alarm goes off. And this doesn't seem like a world-shocking kind of revolutions, but it does mean that people can stay independent longer. And then they don't feel that old... Let's put it like that.
+
+Well, I think more and more people -- I mean, becoming 100 years old is actually not that strange anymore, is it? And getting to that particular age. Many of these people may have lost their partner if they had a partner. Or live far away from their other family. I mean, especially in the U.S. -- I mean, after high school you probably move away from where you lived, where you grew up. You go to a college somewhere in upstate New York, you go to work in California, and people are not close anymore. That truly is a loneliness epidemic. So what can we do?
+
+And I love the work that is happening around sort of companion robots. Why? Because I always thought that people would treat a robot like a piece of metal mechanics, you know? It turns out it's not at all the case. They treat them like pets. Really. Kate Darling, MIT - she's done all the research around here. I can tell you, it's absolutely amazing stories about how people get attached to their devices. The story, I think, is also in the predictions... 80% of the people that have a Roomba, the cleaner in the house, have given it a name. And one of the stories that Roomba tells, like "Well, one sent an old Roomba back", because it needed to be repaired. And they look at it, and they tell the customer, "You know what, we'll just give you a new one." And they say "No, no, no, no, no, no. We want \[unintelligible 00:19:42.26\] back."
+
+**Jerod Santo:** \[laughs\] Really? They want the old one, because they're attached to it.
+
+**Werner Vogels:** \[19:51\] Or Kate tells this story where at some moment she divides two groups of people, and each get a dinosaur, a sort of baby dinosaur. And they really fall in love with it, they give it a name, they play with it, blah, blah, blah... And then she gives them a hammer, and she tells them "Kill the dinosaur." Nobody in the group will take the hammer and kill the dinosaur. They have no problem though going to the other group and kill their dinosaur.
+
+So why do I tell these stories? It's that while these things are technology, we treat them as living beings. And especially, I've always -- of course, because before I was in computer science I was in healthcare, the application of these kind of tools, toys in hospitals are incredible. If you're a young kid, a doctor is someone that is tall, and I don't know, dangerous, or whatever you want to think. But for example, there's this Huggable robot - it's a small, green thing, with a nice snout, and it talks to the kid, and whatever... Kids with a Huggable have no problem talking to a doctor, because they have their mate with them.
+
+And so those kind of things, I think, are crucial for us as technologists to think about. We see a problem, not necessarily a problem that is going to make us a ton of money, but we see a problem that is really important for people's psyche, for people's feelings, for loneliness... Who wants to be lonely? Nobody. And if we can't do this through pure social interaction, what's the next best thing? We have - or Amazon has - Astro, but there's more robots like Astro. And they'll just run, they'll go search for you through the house, and ask you if you've taken your medicine. Families or people that have these kind of robots - their medicine taking goes up 80%. Why? Because they're being continuously remembered.
+
+And actually, what way predates this is early on when we launched Alexa, I got a story from a man who was suffering from early dementia. And he knew what was coming.
+
+**Jerod Santo:** Speaking of Alexa...
+
+**Werner Vogels:** That's her.
+
+**Jerod Santo:** And there she is. \[laughs\]
+
+**Adam Stacoviak:** That's her.
+
+**Werner Vogels:** You can't help me with that. No, no, no... \[laughter\] So the only thing he had to do was put a post-it note on the devices that's called Alexa. Because one of the things that he'd noticed was that during the day his caregivers were getting more and more irritated with him. Why? Because he was asking the same thing over and over and over again. "What day is it today? What day is it today?" Because he had completely forgotten. And the thing with that, let's say, voice assistant was that you could ask it the same thing, over and over again. And it would never get angry. It would always be happy, and it would always give the same answer.
+
+So there are technologies that we can develop to help people have a good life. And I think if you can do that... Spam filters - fine.
+
+**Adam Stacoviak:** \[24:13\] Sure. Yeah, right? What I find interesting with this experiment is - and something I don't think I heard you say here or in your post - we're actually, to some degree, programmed. I mean, I know I had a lovey of sorts when I grew up as a child. I know that I've got two boys that are growing up, and they both have a thing that they were given at a very young age, that is not technology, but it's about the best abstraction you can get. And a lot of people have had it in the last 50 years, even 200 years, or even more, is this idea of a companion that you draw an affinity to. Now, this thing does not talk about medicine, obviously, it does not remind them about things, but it's this object that they attach to and cling to.
+
+**Werner Vogels:** I mean, how many kids don't you see walk around with a little rabbit after 10 years?
+
+**Adam Stacoviak:** And you mentioned the dinosaur... I mean, my kids would no way ever destroy that thing. It's their lovey, it's their prized friend/companion that has no humanity whatsoever, no ability to speak, no true reciprocal affection even, but it's theirs and it has been theirs, and it's their affection point, that is uniquely theirs and only theirs. So even from young age they've had this companion... So it doesn't surprise me that your prediction is maybe that as an antidote to loneliness.
+
+**Werner Vogels:** Well, I don't think it's an \[unintelligible 00:25:47.02\] to loneliness. It is an assistant.
+
+**Adam Stacoviak:** I said, antidote.
+
+**Werner Vogels:** But people are still lonely.
+
+**Jerod Santo:** Yeah.
+
+**Werner Vogels:** I have seen elderly ladies with a fake cat next to them, stroking the cat, and the cat meows... And the elderly lady is extremely happy about that. And so why not? And the fact that we have had our phase in our existence where we think "Oh no, all relationships need to be human to human..." Well, maybe not. Maybe there is more to that... Especially because we are getting older. I mean, look at all the baby boomers. Everybody over 65 or over 70 - there's a very large group of people that is growing older and older, and come with a new set of problems. Can we solve those as technologies? Can we make a contribution to that? That would be great. Because I think building technology is cool, but helping people is much better than that.
+
+**Break**: \[27:13\]
+
+**Jerod Santo:** I think we leave it right there and turn to Adam's favorite topic, quantum. Quantum. Quantum. I'll say it three times.
+
+**Adam Stacoviak:** Thank you, Jerod.
+
+**Jerod Santo:** The headline for this particular prediction is "Quantum safe becomes the only safe." Now, we've covered quantum computing here in the past, and it's always been right around the corner... It seems like it's always around the corner... But your contention, with some data to back it up, is that corner is getting closer and closer, and it's maybe not five years away, like it always has been, or 18 months away... It's actually -- there's things happening, and it's like, we're getting there. We're getting there. Can you tell us more?
+
+**Werner Vogels:** The most important thing is that I don't necessarily want to put a timeline on it.
+
+**Jerod Santo:** Sure.
+
+**Werner Vogels:** \[unintelligible 00:29:38.00\] two years ago, three years ago I think I said 5 to 10... I think five is now more realistic... But I've seen a number of problems arising that I think we need to be aware of and take action on before that. I mean, we were \[unintelligible 00:30:00.21\] with Caltech, which is an institute around quantum... And I see different institutions making massive progress in terms of error correction, networks, things like that, that actually make me believe that it's no longer - how should I say - a vision. It's an execution now. Now it becomes an iterative process in improving the technology that we have created step by step, until it becomes workable.
+
+Now, there is a ton of other problems to solve. How do you program these beasts? I mean, how do you do that? How do you do with something that is not necessarily 100% reliable? So how do you do that? So there's a ton of things that still have to be solved before we have production quantum machinery. But one of the things that these things -- and they won't solve necessarily new problems; they'll solve existing problems much faster. Things that would now take a thousand years in compute, will then take a year, or six months, or two months. That's the premise of quantum. Well, the one thing that we'll then be able to do really, really, really, really fast is decrypt. And so anything that is, let's say, elliptic, or the usual RSA-encrypted kind of things, that are perfectly fine now, because it will take way too long to decrypt, then will be decrypted with a snap of your fingers.
+
+\[32:06\] So the reason for the prediction, actually, is to put in people's minds that we need to start taking action now, if we want to be protected five years from now, from, let's say, the ability to decrypt anything that we have now. Now, all the hyperscalers - Google, Microsoft, Amazon, whatever - we all have been building as a post-quantum cryptography, we've been inserting that into our frontends, and we're doing all of that. Unfortunately, we're not the only players in this world. There's lots of people with their own data centers still, with their colos, with their whatever, that will have a hard time protecting themselves. That's one thing. So I'm really getting it into people's minds that they have to start taking action now, if you want to be safe a number of years from now, when these things become reality.
+
+The other thing is that you can't just lean back right now, because a lot of the break-ins, or data stealing, or whatever you call it, is basically data harvesting. Nobody is interested. Basically, encrypted data is being retrieved, and people don't try to decrypt it. They'll wait till the machinery is there to be able to decrypt. Your medical data now will be decryptable in a few years from now.
+
+**Jerod Santo:** Well, that's not cool... \[laughs\]
+
+**Werner Vogels:** And with that, I don't necessarily mean script kiddies, or whatever. It's really state actors, it might be commercial actors, who are interested in that kind of data, even if it's five years old. It might have all the relevant data for them to take action on.
+
+**Adam Stacoviak:** That's the scary part, really... The famous TV show - you may have heard of this, Werner; it's called Silicon Valley. And in season six - I'm going to spoil it for some folks, just a little tiny bit, let's just say... And you may know this, because you may be a fan, or maybe you're not, because it's too close to home. Who knows...? Is that the AI they build, or accidentally build, to some degree - they give it a task, and it essentially goes out of their bounds that they planned for, and breaks encryption. So this is predicted in a TV show that's now ended, but very much prophetic in a lot of its satire. And that's been my major concern with quantum, is that in the future we have a machine that is - as you said, it's not deterministic, or unpredictable, in a lot of ways. It's also unpredictably predictable in the fact that it's unpredictable. But it has so much possibility that that's when encryption will essentially go away, and the world we live in that is safe because of encryption, is no longer safe. That's been my major thought around that, really, for quantum.
+
+**Werner Vogels:** Well, there's of course different types of encryption. You have lattice-based encryption, and you have that... So there are types of encryption that you will be safe from sort of these type of decryption. But you need to start installing that now.
+
+\[35:56\] LibSSS was 500,000, 600,000 lines of code, or something like that. I mean, you can imagine that it had to have all sorts of different types of encryption in it. At some moment, we started realizing at Amazon that running all of our frontends with LibSSS was a major vulnerability, and so we reimplemented TLS, with a limited set of encryption technologies, and just put that in front of S3, for example. First of all, much cheaper, but we made it open source, and we made sure that it is fast and inspectable... And more important, written in a way that we can use automatic reasoning to see whether or not we're protecting ourselves.
+
+So the most important part -- I mean, now you hear more and more about automatic reasoning, but people like Byron Cook, who have been working at Amazon for the past 15 years on these kind of problems, proving to ourselves that we are protecting our customers. That's not a press release that is going out, or something like that... But there are technologies that you can use to sort of prove to yourself.
+
+Now, \[unintelligible 00:37:23.05\] Signal to Noise, was the original name of the open source project - it was purely also intended to get post-quantum encryption in the hands of everyone. Now, I told you -- now you got scared about the data harvesting, let me scare you a little bit more... Probably every of your home devices hasn't been updated for a number of years. They run Linux, whatever version, no automatic updates or anything like that... Your automatic garage door might have a little bit of encryption in it between the key and you... Or think about hotel rooms. Your key is an encryption encoder. So there are so many small parts... I mean, first of all, coming back to these house devices - I recently looked at one of the house that we have, how many devices were on the network, and that was something like 56. There's your Sonos, and there's your garage, and there's your light opener, and there's the thing that keeps track of this... And all these kind of things. Are all these things up to date? Are all these things protected? Are all these things up to new standards? We have a number of years to get this done, but we do need to start now.
+
+**Adam Stacoviak:** So the task today is what? What is the task we could do today as developers, as maybe potentially inventors? What should we do today to plan for quantum?
+
+**Werner Vogels:** That really is the question in there. First of all -- and there's a number of areas... I mean, there's just using newer encryption technologies... But you can imagine that some piece of data that you encrypted 15 years ago, you're not going to go through all of your data and re-encrypt it again. But are there gateway possibilities that protect you from that? Are there other ways? Let's use our inventive brains and find solutions to this, because I believe we can. But only if we make it a focus. And the focus is not that quantum is five years away, the most important part there is that we need to keep ourselves and our businesses safe. And if that happens to be quantum, or if it happens to be an alien spaceship that lands from whoever, and suddenly turns out to be able to decrypt all our data... That doesn't really matter. We need to start protecting ourselves.
+
+**Jerod Santo:** \[40:35\] How sure can we be that quantum safe technologies, as we define them today, actually will be quantum safe? Do we know?
+
+**Werner Vogels:** Yeah, well... Do the work. Do the work.
+
+**Adam Stacoviak:** Do the work. \[laughs\]
+
+**Jerod Santo:** I was hoping you already did it, Werner. I was hoping you did it.
+
+**Werner Vogels:** Yeah, no, no. It's just -- do the work. No, that's it. The same goes for everything now. Let me take a whole different example.
+
+**Jerod Santo:** Okay.
+
+**Werner Vogels:** When at Amazon S3, we went from eventual consistency to strong consistency. That was a major change that impacted every piece of technology. Now, how can you make sure that you touched every edge of this particular problem? Because it is either 100%, or not. So automatic reasoning plays a role in that. Annotations, TLA+, other technologies that you have to prove to yourself that what you've built actually does exactly what you want it to do. So definitely, I think automatic reasoning continues to play more and more a role in giving us a secure feeling that we are on the right path. The same is -- for example, we recently launched Cairo. That is the spec-driven IDE with an AI system.
+
+**Jerod Santo:** Yeah, we did a show on it. Yeah.
+
+**Werner Vogels:** We use automatic reasoning there to limit the number of hallucinations that can happen.
+
+**Jerod Santo:** Yeah.
+
+**Werner Vogels:** Not that automatic reasoning knows that terribly much more, but it knows A squared plus B squared and C squared is probably a triangle. And as such, it knows all the triangles in the world. If then this LLM says "This is a triangle", automatic reasoning will say "No, that is not a triangle."
+
+**Jerod Santo:** Right.
+
+**Werner Vogels:** And that's just a simple example. But we start building more and more technologies to prove to ourselves that we are doing the right thing.
+
+**Jerod Santo:** Well, let me just take a quick moment to say... We talked Wikipedia last week, Werner, and I called it perhaps the eighth wonder of the world... And if it is the eighth, I would say Amazon S3 is probably the ninth wonder of the world... So just a quick moment of props to you, and thank you for what is an amazing, amazing piece of technology, that's unlocked so much for so many businesses and people over the years. S3 is wild.
+
+**Werner Vogels:** Yeah... You know what my first programming languages were at school?
+
+**Jerod Santo:** No.
+
+**Werner Vogels:** \[43:47\] COBOL, 68000 Assembler, and Pascal. None of these programming languages anybody writes anything anymore... Although maybe I could make money as a COBOL programmer. \[laughter\] I'm not too sure I'm doing well in 68000 anymore... But things change over time. And we, as humans, adapt. We're often ahead of the game. We went from things like Pascal, which probably was sort of the first version of a little bit of structured programming, to C++, and Java, and Python, and the whole world. And none of us panicked there. We just learned a new programming language. It's not like that we're suddenly frozen in our chair... And now we have AI tools to help us building the systems that we do. Why should we start to be scared suddenly? Why should we suddenly become -- we just learn to use new tools. Because remember, it's you that builds, not the tools. It's also you that own it, not the tools. If the tools have built something - and you're in a financial company - that suddenly no longer makes you compliant, it's pretty hard to say to the regulator "Oh, that was the AI. That was the tool." No, no, no, no, no. It's still your responsibility. And as such, I think that no matter how good tools we build, we still need to educate ourselves. We need to continue to keep ourselves on a path of education, learning, continuing to learn. And looking at you two, you're not the youngest...
+
+**Adam Stacoviak:** Well, thank you.
+
+**Werner Vogels:** So I'm pretty sure that before you wrote something in Python, you wrote something in C++, or in ABC.
+
+**Jerod Santo:** Oh, yeah.
+
+**Werner Vogels:** \[unintelligible 00:46:11.04\] learning Python, or going to the next programming language, my favorite being Rust at the moment... But there will be something else after Rust.
+
+**Jerod Santo:** Sure.
+
+**Werner Vogels:** We won't panic suddenly, and thinking that the world will collapse and all our jobs are going to go away. No. No, we make sure that we educate ourselves. And crucial in all of that is curiosity. If you just want to lean back and just write some code, and things like that, and don't want to learn the next language, or whatever, then probably you've always been in trouble. That has nothing to do with AI.
+
+**Jerod Santo:** Yeah.
+
+**Werner Vogels:** You have to -- because curiosity leads to learning. And those two things are -- remain curious. Have you guys seen Ted Lasso?
+
+**Jerod Santo:** Yeah.
+
+**Adam Stacoviak:** A hundred percent.
+
+**Werner Vogels:** The TV series. There is this brilliant darts scene, where he throws a hundred and seventy. And he tells this story that...
+
+\[47:38\]
+
+*"I saw this quote by Walt Whitman. It was painted on the wall there. It said, "Be curious, not judgmental." I like that.*
+
+**Werner Vogels:** And he said "I suddenly realized that nobody had been curious in who I was. They were just bullying me just because they could." And curiosity is such an important part. Being curious about AI, being curious about quantum. Be curious about how defense technology actually transforms into civil technology. If you're curious, you'll find out. And I think if you look at young kids, they're massively curious... Until we send them to school. And then we make them conform to one standard. They all need to become the same. And regardless what their mentality is, regardless what their level of interest is, I think everybody has the ability to become special.
+
+I think one of the talents - and given that we have the Renaissance developer as one of the points in there... In the Renaissance, the core of Renaissance humanism was the belief that humans are limitless in the capability of invention, as long as you keep their curiosity alive. The moment they're no longer curious, they'll lean back and it's over. So I believe that that's really -- we're unlimited in the capabilities of inventing new things... If we're just being allowed to.
+
+**Jerod Santo:** I don't disagree. You hit on a couple of your points, the Renaissance developer and the personalized learning, which I think is hugely opportunistic, or a huge opportunity for so many... Because like you said, sometimes by necessity we push all of our children into a square hole, regardless of their educational shape, and they must come out conformed to whatever standard. And I do think that technology is allowing us to break out of that conformity to a certain extent, and provide personalized, customized education, leaning into what the kid's interested in, and playing to their strengths...
+
+**Werner Vogels:** Yeah, but even also just learning maths. I have two daughters, they're now in close to their forties; don't tell them that...
+
+**Jerod Santo:** They've got to know that, right...?
+
+**Werner Vogels:** One of the things -- so I'm not American, I'm not from the US. I come from the Netherlands, where high school has 10 different forms, depending on what you're -- you go to school to become a welder, or you go to a school where you go to university, and there's a whole range in between. And then they come to the US, and the kids have to go to high school, and all kids have to go to the same high school. They have to take the same classes. They have to score good in the same classes. And when one of my daughters, the first time that she got into math class, there were four guys in the back of the room that were four years older than she was, still had failed maths every time, but had to still sit in that class. Disruptive, and -- well, you know the stories. And as such, I've always been a real fan of Ken Robinson, a professor from \[unintelligible 00:52:04.06\] from the UK, that moved to Los Angeles at some moment. And I've been a real fan of him, because he really looked at that - instead that we let our kids bloom and explore and become really what they can become, we turn them into factory workers, all doing the same thing. And that kills all the possibility in what these kids have in them. So...
+
+**Jerod Santo:** \[52:47\] What do you see coming out or happening in the education space right now, that leads you towards this prediction? Because you're predicting infinite personalization, and these things... Are there solutions out there today? Are there companies working on it?
+
+**Werner Vogels:** Well, I think there's enough companies working on them, but what I see more and more is that these kids are doing it themselves. Gen alpha, the current kids that are in high school, or going to high school, they really know how to use AI, or AI assistance. And they know how to make a curriculum for themselves. And they know how to make something from which they can learn. That is a really a tool for them to build something that they just want to do. It helps them build things without having to have to go through a complete textbook first. Now, I still believe that you need to be able to do maths before you should be allowed to use a calculator. That seems like a reasonable kind of thing. But providing -- if you're really interested in, let's say, how in the 1600s in Italy, from small states, this became one country... For some reason, I don't know; kids have immense curiosity in anything. You can build yourself a curriculum that you become -- not that you become an expert, but that you know everything about it. And this is something -- suddenly, you have control over that.
+
+Now, there's always two sides to this. The other side to this is that -- and this is actually something from Ken Robinson, again; he says "There are no good schools. There are only good teachers." And he says "And what do most teachers spend their time on?" Grading homework, doing administration... All these kinds of things that actually have nothing to do with education. So that seems to me like a standard technology solution problem. And whether you solve that with AI, or whether you solve that with just some simple SQL, fine. That doesn't really matter. I believe that there should be a lot more support for teachers such that they can focus on things that is important. Individual interaction between a teacher and a student has so much more impact than the teacher just standing in front of the class.
+
+\[55:57\] \[unintelligible 00:55:53.24\] tells this wonderful story... He's a teacher, and he tells this story that -- he's giving drawing lessons. And there's always this one girl in the back of the class that is never interested in anything. But now they're drawing lessons, and she's fanatic. So he goes to the back of the class and looks at what she's doing, and he says what's she doing? And she says "I'm drawing God." And he says, "But nobody knows how God looks like." And she looks at him and she says, "Within a minute they will..."
+
+**Jerod Santo:** \[laughs\]
+
+**Adam Stacoviak:** That's funny. Yeah.
+
+**Werner Vogels:** I mean, would we think "Here's a blank paper... Why don't you draw God?" No. But this is a kid. There are no boundaries to your curiosity, and to what you can. All these things that these kids feel as constraints are put upon them.
+
+**Adam Stacoviak:** Yeah, it's a double-edged sword with this curiosity point with children, I think, especially with AI... Unfettered curiosity at a young age with such a powerful thing is not that smart, but I've personally experienced my ability to be curious and to explore at my older age, and I like to translate some of that to the younger generation... Because I agree, I think there's this idea of just-in-time learning, when you learn about a new thing and there's a curiosity there to accomplish a goal, but you can't really accomplish that goal until you've gone a little deeper here, here, and here, for context. I think that's the new term or the new word of maybe potentially 2025, is context. Sure, AI and maybe even agentic, but I think a subversion of that is context. This just-in-time learning is really, really fascinating.
+
+**Werner Vogels:** Yeah, And you must have been to a conference or whatever, to some tech conference at some moment, where someone describes something about a new cache algorithm...
+
+**Adam Stacoviak:** Yeah.
+
+**Werner Vogels:** And they look at that and they think "That shouldn't work." Or -- actually, at Amazon we have something called the Builder's Library that's available for everyone to look at. These are articles written by Amazon's most senior engineers. And there is this one article by \[unintelligible 00:58:28.23\] that I couldn't get my head around... Because I've grown up in a time that you minimize the number of bytes on the wire. One more byte means more time, more delay, whatever. That's no longer the case. He describes this system in which if the receiver always has to do exactly the same number of things, because there are 10 things in this packet, but eight of them may be empty... But everything is just standardized, you don't need to think about it, and the resilience goes up. So it took me a long time to get my head around that. But the first time I read the intro to the story, it immediately piqued my curiosity. And it took me a long time to learn what he really tried to achieve there. And so I think no matter how old we are, there are moments every time when you get sort of "Oh, how would that be?"
+
+\[59:48\] I live here in a place where there are a lot of electric scooters. There's Lime, that Uber bought, and some others, and whatever... And I've always seen them, but I always was interested/curious in "How would it be to have one?" Well, now it's just a click on your phone and you jump on it and you go to the supermarket to buy your groceries, and then you come back, and it's just fun. But curiosity is really one of the most amazing things that is uplifting in our life.
+
+**Break:** \[01:00:29.17\]
+
+**Adam Stacoviak:** It makes me think about this next notion, which I'm not sure I agree with, but it's been potentially the phrase, or at least the sentiment of the last year or so, mainly around fear, uncertainty and doubt, which is "Developers are dead." That we're a dying breed, AI is taking our jobs, it's going to replace us... And you seem to push back on that idea through being a renaissance developer.
+
+**Werner Vogels:** I do think if your job was only being - and that sounds a bit negative - a code monkey...
+
+**Adam Stacoviak:** All you do is code. That's all you do. You don't think, you don't plan...
+
+**Jerod Santo:** Right?
+
+**Werner Vogels:** If that would be -- but I think most of us have had to think about some of the stuff that we built. We had to think about algorithms, we had to think about trade-offs, we had to think about sort of -- you went over to your colleague on the other side and you said "This is what I need to solve here... But I think there's a cache overrun if that happens, and that will brown out. Do you think that, too? Shall we set up a test environment and see whether that works or not?" There are so many parts to development that will require our brains. And that's why I picked the word renaissance, is because before the renaissance, there were the dark ages. A thousand years, basically... The black death, the plague, all these kinds of things... And everything was ordered by basically religious institutions. And that changed at some moment where people became interested again in all the amazing things that have been done a thousand years before by the Arabs, by the Romans, by the Greek... To where they suddenly completely started to change. And there is this word that is sort of associated with the renaissance, that is polymath. And many people think that that is someone who can do mathematics in many different forms. No, no, no, no. Poly means \[unintelligible 01:06:02.22\] means learning. Much to learn. And if you look at the people in those days, definitely, they suddenly started to become interested in many more things than just only the one thing that they were specialized in. Da Vinci was not just a sculptor. I mean, he built airplanes, although they didn't fly... But he did build them. And just trying. I mean, think about all the people in those days that suddenly allowed their curiosity to come out.
+
+And I think as developers, there are a number of things that we will need to re-encounter. First of all, I do think we've been learning for 50 years now - or however long we've been doing computer science; probably 1960s, so... Yeah, 60 years. So we've been learning programming languages for 60 years. We've been going from mainframes, to minis, to PCs, to cloud, to whatever... Everything required a new approach to building, or enabled this to do new things. Before cloud, you may have only had one colo, and never will be able to build a replicated database, but then cloud came, and you certainly could replicate your database. You needed to learn how to do that, but still, it is. And definitely, I believe that it's not just sufficient to -- and now I need to be careful, because I don't want to give away too much of my keynote...
+
+**Jerod Santo:** Oh, go ahead...
+
+**Adam Stacoviak:** Spill the beans... I'm just kidding.
+
+**Werner Vogels:** \[01:08:17.17\] I believe to be successful, not only in the future, but also now, is to not only have a deep specialism, but start to become interested in the broadness as well. It's called a T developer. It is someone that is not just a database expert; not just a database expert, but he knows about the other types of application, knows about other programming languages. He may even be able to help his colleagues on the other side, or at least can have a conversation about it; understands the business principles of the business that he's actually running to.
+
+Another area that I think the newer type of developers - the renaissance developer needs to be an excellent communicator. And with that, for example, I mean, if you look at the Amazon website - just imagine that for you - there are some parts of the Amazon website that always need to work, because otherwise we can't sell anything. Browse, search, shopping cart, checkout, and reviews. Because if there are no reviews, people don't buy things. Five things that absolutely need to work. And then we have all the other things that actually are really important as well. Then we have recommendations, personalization, whatever. And then there is stuff that is just nice to have.
+
+Now, take for example bestseller lists. Now, as a technologist, you need to be able to have a conversation with the business, and say "How many nines do you want behind that? Oh, you want four nines behind your bestseller list? Well, that's going to cost you so much." And then it becomes a business decision how I am going to implement this. But I need to be able to explain to the business what the risks are, what the costs are, how resilient can I make this... But without that kind of communication, you're not building the things that you want for your customers.
+
+So a very important part of the renaissance developer is to be able to communicate in exactly the right way. Not only to the business, but also with your customers. "What are you building? How does this interact with this?" If I show you a picture of the thousands of microservices that builds out amazon.com... If I do something here, what's happening there? Is there a relationship between the two?
+
+So system thinking becomes crucial as well, not just about your little piece that you're building, but what impact does that have on other things? Or, you know, I'm happy going away, my database is running, my code is running, everything's happening, transactions are playing through... And suddenly, the number of transactions doubles. But there is the doubling of the number of the customers on the website. Who the hell is doing this to me?
+
+\[01:12:08.27\] And as such, being a specialist in one particular area and just staying there is safe, it's fine, but you live in a bigger system. And that really comes out of '60s, '70s, \[unintelligible 01:12:33.17\] Meadows began something called system thinking. In nature, if you remove one piece out of that equation, of the complete nature equation, you ruin the overall thing.
+
+At some moment -- what was it? Yellowstone Park. There were too many wolves. They removed the wolves from the park. Oh, they were killing all the elks. So they removed all the wolves. And then the number of elks grows like mad.
+
+**Jerod Santo:** Right. Backfire.
+
+**Werner Vogels:** Which starts eating all sorts of Yellowstone. And so not until they reintroduce the wolves, the balance gets back. So on systems thinking is something that we as the future renaissance engineers need to be part of as well. We need to understand that the one thing that we're working on at this moment has impact on all the other pieces around it. We need to be able to communicate with the others. We need to be able to see how we are part of this bigger problem. And you know what? It will be fun. It will be fun. I absolutely think.
+
+And so one of the things I think that I -- I won't say hate... Let's say one of the less interesting parts in the beginning of my life as a developer were code reviews. It's a bit like standing in front of the class, and the whole class gets a chance to say "You're an idiot."
+
+**Jerod Santo:** Right.
+
+**Werner Vogels:** No, but code reviews are fun. Sometimes you think "Ah, there's mistakes there", or sometimes you go like "Wow, that's an elegant solution. \[unintelligible 01:14:40.16\] And code reviews are super-important for junior engineers. Because the senior engineers are in the room, and discussing the code on the screen there, and you're learning on the spot from these senior engineers, from all their years of experience... Where someone will say "I tried that when we were building product X, Y or Z, and it didn't work then. Or the problem at that time was that." Now, whether the code on the screen is generated by you or by some AI agent doesn't really matter. We still need to do code reviews, because we have ownership. And whether we used a tool to create the code that we're looking at or whether we actually wrote it ourselves - we're still responsible for it. We still own it. And so those things don't go away.
+
+And as such, if your system has been created by one of the AI tools - and as I said earlier, you are responsible for it, especially if you're in life sciences, if you're in healthcare, if you're in financial services, in whatever; you cannot only rely on the fact that "Oh, but AI built it."
+
+\[01:16:18.17\] You are still responsible for what you deliver. And that means you need to be able to look at it, and as such, generation of code may go faster, reviews will probably go slower. Because this is not code that I've written. If it's my code, I can go stand up in front of the class and say "Oh, and this is why I did this. And this is why I did that." But if a complete unknown, unhuman entity created this code for you, it takes longer to get used to it.
+
+And that's not bad. I mean, none of this -- don't take this as that I'm sort of cracking on AI as being... No, there's parts that will never leave us. We need to learn the next steps, we continue to keep ownership of what we do, we continue to need to learn the next language, the next tool, the next system that we built... And we need to know more than just that one piece. We need to be more than that.
+
+And my favorite - and he comes back in almost every presentation that I give - was Jim Gray. He won the Turing Award. He's the guy that actually built System R at IBM, the inventor of transactions. Jim was brilliant. Absolutely brilliant. There's an article, if you want to look it up, it's called "20 questions to Jim Gray". Jim said "Give me 20 business questions or questions that you as a user of a database would want to have the database answer, and I will build a database for you." And by the way, one of the most interesting questions is - he walks into the room where the big machinery is, and he hears the disks rattling, and he looks sideways and listens to the disks... He says "Your database layout is wrong." Now, that was something he could hear.
+
+Now, he was not just a database expert. He had a lot of other skills in other areas. He was what we would call a T developer. An I developer is someone that knows only one thing. A T developer has a number of knowledges in other areas, whether it's computer science, whether it's from the world, or whatever. It's a broader person than just only the database developer. And I think in the future, this is something that we will continue to harness. You need to know more, you need to be able -- because how can you communicate with your colleagues that build other parts of the system if you actually don't know anything about what and how they're doing it?
+
+**Adam Stacoviak:** This idea of code review, one thing you're making me think about is - at least the sentiment it seems today is getting through code review. It seems like you're suggesting, even though it's painful, not so much pausing, but not getting through it just to deliver the code to production and correctness, but to understand why the code is what it is... And especially if it's AI generated, and the human in the loop is less informed, specifically about the code itself, that it's more of an artifact of a plan, not an artifact of actual implementation themselves.
+
+**Werner Vogels:** \[01:20:17.27\] Yeah. And I'm getting old... I just had a great answer to what you wanted to \[unintelligible 01:20:22.28\] and I completely lost it. \[laughter\] I do think that we've changed over time, as developers. I mean, I've written VB. I don't think there's anybody that writes Visual Basic anymore, is there?
+
+**Adam Stacoviak:** Not on purpose... Fix a bug maybe, in legacy code, if it's still there...
+
+**Werner Vogels:** Yeah, something like that. But it was fun when we were doing it. And I wasn't thinking about five databases replicated over five different data centers throughout the world at that particular moment. I was trying to fix my Excel code. And so we've all grown as developers over the past 10, 15, 20 years, with all sorts of new tools. We have a new tool. There is a particular point, a small problem I have with it... Is that if you look at all the other tools that we've developed, we took some time to introduce them. We had learned people, we told people why we were building the tools, how we were building the tools, how they could best use them... The typical early adopters, and you had to cross the chasm, and then you had the early majority, and things like that. And it would take time, but you would learn, you would teach your users why and how that tool was there. With some of the current, the new generative AI tools, they were just dumped in people's lap, without telling them what it could do, or what it couldn't do. And there isn't almost a CIO that I meet today when I'm traveling that asks me, he says, "What should I be doing with AI?" And I go "Well, my excuses... It's very inappropriate to answer a question with a question, but why are you asking me this?" And he's like "Yeah, but those guys next door, they will be ahead of us." And I say, "Are you really certain of that?" And then maybe because you're a little bit older, you start to drive down into what is actually the problem they're trying to solve with this technology, and "Is this the right technology for this?"
+
+Every week, we see five new models, or ten new models. Suddenly, we went from regular LLMs to reasoning LLMs. I think as a business there is no shame in hitting the Pause button for a moment, and say "Why don't we get ourselves educated about all of this?" And not just us as technologists, but also the business. Because at this moment, quite a few of our architectures are being determined by the media, not by us. That shouldn't be the case.
+
+\[01:23:58.06\] We, together with our business, should determine how our architectures should look like, not because the newest thing... And of course, that's the task of the media. And the task of the media is also to write negative headlines, even if things are looking up... "That company is massively behind that company." Yeah, you know what? AWS was never a consumer company. So we don't build consumer products. We build tools such that you can build your chatbots.
+
+**Jerod Santo:** Right.
+
+**Werner Vogels:** If AWS doesn't have a chatbot, it means that we're out of business in AI. It just means it's not our business. And I think that's really crucial in all of this, that we take time to learn the technologies, the capabilities of the technology, and where it can help us in our businesses. And if we also give it to a whole bunch of gen alpha kids that can go do wild with it and build their own educational curriculum with them - great. Let them have it. It's a bit like us 50 years ago with hammer and nails, and building your own thing up in the tree, you know... And now kids do different things. But we need to give them the tools for it.
+
+But definitely, in business I think there is no shame in hitting the Pause button and spend some time on education and making sure that you and your engineers are making the right decisions, with the right tools. There is no hurry. There is no hurry. No company will go out of business in the coming two months because they weren't using AI.
+
+**Jerod Santo:** Well, you heard it here first, because very few people are saying that out loud... And while I agree with you and I think that it is a problem, I wonder if we have an appetite as an industry to address said problem. I think we're just barreling forward, and we'll see what happens. But I also agree when you say give the kids the tools. Give the kids the tools...
+
+**Werner Vogels:** Well, one of the other things... It didn't make it this time into the predictions... Did it do it last year? One of the things that -- I think last year one of the things I talked about, one of the things that worries me, is that parents driving in their car, the kids are in the backseat, they're getting "Dad, when are we getting home? When are we getting home?" And that was how we used to drive home. Now, the kids get an iPad, and become YouTube experts, and they're quiet in the back, and the parents love it. But you set kids up from four or five years old to get dopamine reactions. They know how to manipulate YouTube continuously. And it's not only that they are getting this dopamine high, there's something else called dopamine too and that actually removes your willpower in these kind of cases. Whenever you stand at the bus stop next time, look at how many people pull their cell phone out, even if they know that the bus is three minutes away.
+
+Nobody wants to be bored anymore. Nobody wants to just stare out in the distance, and use their brains... And we can't help ourselves, because we've ruined ourselves. And if we do this with our young kids, we will have an epidemic on our hands 10-15 years from now in terms of addiction. Anyways... Sorry, guys, I keep yapping... \[laughs\]
+
+**Jerod Santo:** No, we appreciate it. We know you've got to go... Didn't quite make it to your fifth, but great, great predictions. Thank you for sharing your extended thoughts on them, and giving us some time today.
+
+**Werner Vogels:** Okay. Well, I really enjoyed it, guys. Come and watch my keynote, the last keynote at re:Invent. I'm doing it at 3.30 on Thursday.
+
+**Jerod Santo:** Nice.
+
+**Werner Vogels:** After me, everybody can do whatever they want.
+
+**Jerod Santo:** You're the closer.
+
+**Werner Vogels:** I am the closer. Awesome, Werner. Well, you're welcome back anytime.
+
+**Werner Vogels:** Coming close.
+
+**Adam Stacoviak:** Yes.
+
+**Werner Vogels:** Okay. Bye-bye.
+
+**Adam Stacoviak:** Thank you, Werner.

When vibe coding goes viral (Interview)_transcript.txt

@@ -0,0 +1,633 @@
+**Chris Anderson:** *Yo, other noSQLs, your tech is a wreck like an oil spill. You call that web scale? More like toy scale. Your memmap file is a database fail, and your user's going to set sail for the one they can trust. Because if you're serious, then you must be running Couchbase*.
+
+**Jerod Santo:** Wow...
+
+**Adam Stacoviak:** Wow.
+
+**Chris Anderson:** That turned out to be true.
+
+**Jerod Santo:** It turned out to be true. \[laughs\] It's nice when you put together a wrap and it turns out to be true as well. It's like a bonus.
+
+**Adam Stacoviak:** Right?! Yeah, I feel like it's been literally 15 years though, since that moment. Almost 15 years, right?
+
+**Chris Anderson:** 15 years of focusing on the same thing in a lot of ways... And what we're doing with Vibes is an outgrowth of that.
+
+**Jerod Santo:** What is that focus, if you could just put a name on it? What are you focusing on?
+
+**Chris Anderson:** Programming is way too hard. It shouldn't be a thing -- like, if you're me, code is never the hard part. And I want it to be that way for everybody.
+
+**Jerod Santo:** And how were you doing that 15 years ago with CouchDB and Couchbase?
+
+**Chris Anderson:** Well, my role over there was the CouchApp side of the project, which was to make CouchDB into this serverless application runtime you could just drop some JavaScript into and go.
+
+**Jerod Santo:** Did you get that done?
+
+**Chris Anderson:** It was - it kind of still is - responsible for a lot of people who otherwise wouldn't have had infrastructure to support them being able to do things like rural healthcare, or fun apps for their club. A lot of the same things that we're seeing people do with Vibes DIY.
+
+**Jerod Santo:** So for our listener catching up, Chris was first on the show episode 54, back in 2011. CouchDB, which I assume is still extant... You'll have to forgive me, Chris; I'm not out there using it, but... It was a massively popular open source database, businesses formed around it... Damien Katz was the original creator, or was it a group of people? I mean, there's a lot of people on the Wikipedia page, but help us understand the culture there.
+
+**Chris Anderson:** Damian's the guy who CouchDB came to in a dream, essentially. He had been working at Lotus Notes and getting frustrated with the fact that it had to pause the world to do its synchronization... So he invented the data structure that allows the streaming synchronization with fairness.
+
+**Jerod Santo:** And how'd you get involved?
+
+**Chris Anderson:** I started using it to spider the web. This was like early, early on, and I had to teach myself Erlang essentially so I could drop a debug log function in there for myself. Next thing I know, I'm through the whole stack, and never looked back.
+
+**Jerod Santo:** Nice. So what'd you do, reach out to Damian say "Hey, I'd love to help out", or "Let's start a business", how'd that whole thing come together?
+
+**Chris Anderson:** All good things start on IRC.
+
+**Jerod Santo:** \[laughs\] That's a great answer.
+
+**Chris Anderson:** We were hanging out on IRC... I think it was \[unintelligible 00:06:12.00\] who came out to Portland for OSCON, like we were talking about... We were already buddies from online. And then after that, Damian invited the Brain Trust out to his place in Asheville, North Carolina, where I said "Hey, let's start a company." So my role in that, besides not being scared of the code, was not being scared of the lawyers.
+
+**Jerod Santo:** Tell the story of that company from there. Tell the story. As long or as short as you want it to be.
+
+**Chris Anderson:** So there's a lot that went on, but the thing that really sticks with me is the fact that we kind of accidentally incubated Node.js. We had people like Isaac Schlueter and Ryan Dahl hanging around the office, mostly because of Mikeal Rogers, rest in peace... And it was that crowd that was just informally writing npm at the desk next to me... So really fun to be in the presence of greatness for so long, and to be able to incubate things like PouchDB etc, kind of taking JavaScript to the next level.
+
+**Jerod Santo:** My first exposure to CouchDB was via - now we're name-dropping - a Geoffrey Grosenbach PeepCode on CouchDB. And he made it look so cool. I was just a fledgling Ruby on Rails developer, of course, dipping my toes into the JavaScript ecosystem, and kind of exploring databases at the time, and like wondering... Because Rails made it so easy to kind of switch your database adapter, but nobody's really doing that... But I was interested in the different databases, and CouchDB was so radically different than what I'd been using. And yeah, Geoffrey's voice, which is epic to this day...
+
+**Chris Anderson:** Legend.
+
+**Jerod Santo:** Yeah, a legend. It made CouchDB look like this is where the cool people are coding... And it turns out it was true; I mean, from your story, it's like - yeah, a lot of very talented people are involved in the project.
+
+**Chris Anderson:** \[08:10\] You have to remember 2008-2009, that was the very beginning of the API revolution. The idea that you could put your database behind a REST API. Almost nothing was even JSON yet. So we were all frustrated with XML, we were all writing kind of monolithic servers... And by the time that project kind of matured, even as early as 2011, at least the early adopters were already doing microservices. So really, that was a big sea change right then.
+
+**Jerod Santo:** Mm-hm. Was it good business, too? I mean, you started a business around it. Was the business good? Was it hard? How'd that go?
+
+**Chris Anderson:** It was good and hard, and in some ways, we brought like a cultural cachet that allowed us to, as we discussed on episode 54, merge with Membase, who had a whole bunch of really smart people who understood the market and how to drive a startup to success.
+
+So Steve and me, who I learned revenue from over there, is now on the board of Vibes DIY, and helping me do the right things here. It took a while, but they IPO'ed. They IPO'ed about five years after I left...
+
+**Jerod Santo:** Nice.
+
+**Chris Anderson:** ...and I'm always watching it real close.
+
+**Jerod Santo:** So what's your journey from leaving there to Vibes, Vibes.DIY, which is your current thing?
+
+**Chris Anderson:** Well, I bounced around a little bit, checking out other database technologies, worked on FaunaDB, which I still maintain; it has some of the strongest integrity that you can get. Then I went to McKinsey and Company... Kind of a fun story there. Like, you wouldn't think that -- my whole family was like "Really?" But Jason Smith is another legend. He's till there, but he brought me in, and his claim to fame is - well, for Couch, he hosted npm. So npm used to have a one man host, and it was him. And yes. So... Great to have him show me what the real world is like. And what I learned about the real world is you don't get infrastructure, right? Us here in the Silicon Valley culture, where we can boot a virtual machine, or request an S3 bucket and get it - like, that's just an unbelievable luxury. What I learned from this constrained environment of "The client doesn't know what cloud you're going to use until it's too late" is I need to build a database technology that can run anywhere to allow permissionless innovation. So that's the gem of what became Vibes DIY.
+
+I built a deep tech thing, along with Mikeal Rogers' team at Protocol Labs, who did a whole bunch of research on immutable data structures with content addressing. I joined that team for a little while, and saw the potential of it, and built an embedded database that runs in the browser, with end-to-end encryption for multi-user synchronization. And that's what powers Vibes DIY. The database is called Fireproof. Basically, everything I'll talk about today is open source, but if you want to get into Fireproof, there's plenty about it. The thing that it does uniquely is give every operation cryptographic data provenance. It's almost like a mini blockchain in the browser, super-lightweight; no network dependency, just runs right there... But it means that when you do a replication, you know what you've got. There's not questions about what kind of data you're sitting on.
+
+**Jerod Santo:** Let's dive into that. When you say cryptographic data provenance, what do you mean exactly?
+
+**Chris Anderson:** So it goes back to Rich Hickey Clojure and the immutable data structures that they were pioneering around the same time we were doing Couch. And what you see in those data structures - as soon as you have an invariant that you're working with immutable backing storage, you have to do everything different, everything new. With Fireproof and the IPLD data structures that it uses, that constraint is enforced because the only pointers you're allowed to use are hashes of content. And so everything is addressed by its own hash.
+
+\[12:27\] That means you can only talk about stuff that's already been written. And that means that every time you're doing these B-tree like updates, you're either going to churn data in a real heavy way, with lots of kind of write amplification that's unnecessary, or you're going to solve these hard, hard problems that the research team over there, people like Alan Shaw and \[unintelligible 00:12:51.17\] worked on, where you have trees that have really interesting properties... So what you need, which is a lot of work to get, is something that is a search tree, like a B-tree, but regardless of if you put all those records in, in one order, or a random order, you still get the same root hash at the end, so your snapshots have stable identifiers. And that means your replication doesn't have to be mechanically the same everywhere. It means you can replicate essentially in an optimized, efficient way, but still end up at the same endpoint.
+
+**Jerod Santo:** So... Fireproof? Firestore? What's it called?
+
+**Chris Anderson:** Fireproof.
+
+**Jerod Santo:** Fireproof - is this your implementation of these concepts, or...?
+
+**Chris Anderson:** It's my essentially making it so easy to use. We talk about technology, especially databases, and one of the phrases people like to say is "a footgun", right? A footgun is when if you don't do everything exactly right, it blows up on you.
+
+And the requirement I had for this technology is no footguns. It's literally designed for the code school dropout.
+
+**Jerod Santo:** Okay. So you can't actually shoot yourself in the foot.
+
+**Chris Anderson:** Right. And that turned out to be the right place to be, because --
+
+**Jerod Santo:** Because we've got a lot of people vibe coding.
+
+**Chris Anderson:** Well, even more than that, I'd say with LLMs and GPTs and all that, we have an infinite supply of code school dropouts.
+
+**Jerod Santo:** Oh, true. Yeah, exactly. They are infinitely supplied... Depending on how much money you have, right? There's another one there; assuming you can afford those tokens. So you're working on Fireproof, you're building this, making it simple for people to build things with it. And then where does Vibes DIY come into the story then?
+
+**Chris Anderson:** So we were going to market with Fireproof after the technology got mature, and there's use cases like black box recorder for self-driving car. Things that are very serious, that are served well by this content-addressed immutable data structure. But there's not very many people who need that, and they don't care how easy it is. That's a different set of problems for them.
+
+When we started to see in our Discord people coming in and saying "I haven't programmed in 15 years, but I built this drum machine along with ChatGPT and your API, and it worked right out of the box", we realized we've got vibe coders.
+
+And so I brought in an old friend of mine, this guy Marcus Estes, who I've known for like 20 years, and he's just a hustler. He makes it happen. He understands where the opportunity is and how to go after it. I've been looking at this vibe coding stuff, thinking "That's fun. That's cute." He realized "No, let's just do this", and we didn't look back. So it's been so much fun.
+
+The big difference, the real difference that matters is my kids, who were so sick of hearing me talk about Merkle CRDTs and all this stuff, now they're grabbing my phone out of my hand.
+
+**Jerod Santo:** You're cool again. "Dad's cool again", you know?
+
+**Chris Anderson:** They just want to make apps. They're like "Dad, can I make an app?"
+
+**Jerod Santo:** "Dad, can I make another Merkle tree?" You know, no one ever says that.
+
+**Chris Anderson:** \[laughs\] Right...?
+
+**Jerod Santo:** \[16:12\] What kind of apps are being built with this thing, or can be?
+
+**Chris Anderson:** Well, I was really inspired -- I was trying to see essentially what's the lowest common denominator that we can address... And the answer to that is I was using ChatGPT Canvas and Claude artifacts and thinking about how for like a normie, somebody who doesn't already know their way around the next JS stack, or deeply understand React and all the tool chain there, those are the best kind of interface and set of metaphors. If you look at those closely, they don't even say Deploy. They say Publish. And so that's the take we've done with Vibes DIY, is when I'm designing a feature, if it's going to require introducing a new concept, don't. Just make it simpler.
+
+And we had this tremendous validation... We just got back from Render ATL. It's a tech conference in Atlanta, where we had a deep line at the booth, crowded up with people who had kind of maybe never even heard of vibe coding, or if they had, they hadn't tried it yet... Or if they had tried it, they thought the existing tools were too complicated... And when we said "Hey, if you make an app, we'll install it on this sticker for you", they would spend the 90 seconds it takes to make an app.
+
+**Adam Stacoviak:** Wow... What's on the sticker? Is it like a QR code, or an install kind of thing, or launch it?
+
+**Chris Anderson:** The way we install it in here is with the near field communications chip.
+
+**Adam Stacoviak:** Okay...
+
+**Chris Anderson:** I actually learned about that at like a Burning Man event, where someone was doing manicures with NFC chips in them.
+
+**Jerod Santo:** Really?
+
+**Chris Anderson:** Yeah, which is super-hot. Like, "Here's my number."
+
+**Jerod Santo:** Right. You can just embed a little bit of information on that chip and then carry it around in your fingernail.
+
+**Chris Anderson:** Yeah.
+
+**Jerod Santo:** That sounds very Burning Man.
+
+**Chris Anderson:** Yeah. You learn a lot. You never know where you're going to learn a lot.
+
+**Jerod Santo:** "Here's your next big idea."
+
+**Chris Anderson:** Right?
+
+**Jerod Santo:** Wow. So if I take the sticker home, how can I use it? Like, I've got this information embedded in it... What do I do?
+
+**Chris Anderson:** You just kind of get it close enough to your phone, and you get one of these little -- and now you can open the URL.
+
+**Jerod Santo:** Okay. So it's similar to a QR code, except for not for the camera, but actually the NFC chip in the phone. Makes sense. That's interesting.
+
+**Chris Anderson:** Right. And for us, it's really this ownership metaphor. Like, code is all in there, over there... Especially if you're not a coder. Like "What am I making? Where is it?" But if I put it in your hand when you walk away from the booth, then you're going to come back with your friend the next day.
+
+**Adam Stacoviak:** Does that give them access to the code they just created, or vibed? Or does that just give them access to the application they just vibed?
+
+**Chris Anderson:** Well, for us, really our --
+
+**Adam Stacoviak:** How do they continue development, how do they keep going?
+
+**Chris Anderson:** You're totally right. So most of these people made it on their phone they walked up to the booth with, because it's a mobile-first experience. But if you didn't, if you made it on my phone, or if you made it on the booth laptop, then you still get the sticker, you can still scan it into your phone... And anyone who goes to a vibe can hit the Remix button in the corner and get into the code, or get into the chat and say -- the default remix we have preloaded in our form is "Make it pink." But really, just whatever you want to do to change it.
+
+There's a great story, and we had a lot of fun... I was walking back from lunch in Atlanta, and I said to my phone, I said "I want to take pictures and have you turn all the faces in the pictures into Georgia peaches." And it worked on the first try; sometimes you get lucky... And people thought that was hilarious. So we were showing it off, and then we were sharing a booth with Netlify. They're great friends of ours... And one of the Netlify leaders, Sean, brought his kid.. And so she saw what we were doing, and she said "I want to be a cat." \[laughs\]
+
+**Jerod Santo:** \[20:23\] Remix it. Remix that sucker.
+
+**Chris Anderson:** So basically, a four-year-old turned a peach app into the cat app, and that was a hit... And then we saw what was going on there and we made this -- we made a caricature app I've got on my laptop here. Here's me as a podcaster.
+
+**Jerod Santo:** Nice...
+
+**Chris Anderson:** And it was also a big hit. People at the conference just all day long building stuff.
+
+**Jerod Santo:** It's just fun. This is the stuff I'm talking about, Adam. Just --
+
+**Adam Stacoviak:** Total vibes, man. This is the internet, man.
+
+**Jerod Santo:** For joy. It's for the joy of it.
+
+**Chris Anderson:** It totally is for the joy.
+
+**Adam Stacoviak:** It is.
+
+**Chris Anderson:** And so many people --
+
+**Jerod Santo:** This is like Geocities for the modern era.
+
+**Jerod Santo:** Seriously.
+
+**Chris Anderson:** Yeah. It's like a writable Geocities, right? View sources back, baby.
+
+**Jerod Santo:** \[laughs\] Yeah. I'm reminded of the old cloud to butt extension. You guys remember that one? Every time the word cloud was on a website, this Chrome extension would replace it with the word butt. I'm not saying this is an idea for a remix, Chris, I'm just saying I'm reminded of it.
+
+**Chris Anderson:** You know, if I wouldn't get distracted, we'd just build it right now. \[laughter\]
+
+**Break**: \[21:26\]
+
+**Jerod Santo:** So is the platform the web then? Or how are they -- how are these things running?
+
+**Chris Anderson:** That's the way I think about it. I wanted to build the experience I had early on in the black and white Mac days, with HyperCard... But I realized that when I look around people trying to build similar ambitions, often there's a "do it our way completely or don't" kind of feel. I call that like "They make you wear the socks." With us, the point is there's already so much platform that's done. So we just build React apps, and we're essentially taking the web runtime, which is what a browser can do, and making it as easy to express yourself there, maybe easier than making a TikTok. So if making an app is easier than making a TikTok, now what? And that's the question Vibes DIY is out to answer.
+
+**Jerod Santo:** I love it. I think Adam and I were just talking about this new possibility, specifically in light of Apple allowing local models on phones, and Xcode, and things are getting localized there, where it's like you could build and deploy your own iPhone app. It's just a one off, just for you; nobody else even has to have it. And there are still platform limitations there with regard to distribution, side loading, and there's lots of stuff. But with the web, it's like "Nah. Just URL, publish that sucker." And that's really awesome.
+
+**Chris Anderson:** Yup.
+
+**Jerod Santo:** So just to nerd out again on the sticker, is the embedded information in the sticker just a URL to where the thing actually runs? Or is there code in there? Or how does that work?
+
+**Chris Anderson:** It's just the URL.
+
+**Jerod Santo:** Yeah, that's what I thought.
+
+**Jerod Santo:** So once you have the URL, we make -- actually, our service layer is so simple. If you look at a lot of the existing vibe coding tools, they're built for professionals, and they're going to mimic a professional output with a full Next.js app. The only thing that the model - which is just a commodity model; we have a switcher. It's actually a great way for AI-interested devs to try out all the different models and get their personalities down. But when you ask Vibes to make something for you, it's only writing app.jsx, and that's it. So our focus is on being fun, done, and alive with AI.
+
+\[26:07\] This last bit means that the apps that you make, get the same API key that we use in your browser to write the code. And you press the Demo data button, and it fills out your to do list with whatever is plausible. Or it makes an app -- I built this because it takes 90 seconds to make an app, it takes longer to figure out how to use an app. So instead, I just mashed the demo data button, and it shows me what the app would look like if you'd spent some time in it.
+
+**Jerod Santo:** So let's say I mash out a 90-second thing, and it's rad, and we're having fun... Can I then expand/extend/build that sucker out, or I have to remix and start something else?
+
+**Chris Anderson:** That's where one of our core values comes in, is that it has to be real. So the 500 lines of JSX that it's going to write is React code -- it's as good as Claude 4 is going to make it, which means it's not that bad. And the standard library we bring to the table with it means that all these apps are fully functional just in the browser, but then they can be remote synced via the cloud to other users for collaboration. And that means copy that -- there's a Copy button; you get your 500 lines of code in your clipboard, and you go drop it into anywhere. Maybe into a \[unintelligible 00:27:21.00\] template, or maybe into another vibe coding tool, and say "Hey, expand this." Maybe you turn it into 10 pages, maybe you take 10 vibes and put them into one app... But the sky's the limit. It's really fully malleable software that pros can take and run with, as well as beginners can just use.
+
+**Jerod Santo:** Adam, what do you think of all these sayings he has? He's got like rhyming words, he's got like stickers...
+
+**Adam Stacoviak:** He's a poet, man.
+
+**Jerod Santo:** This guy's bringing it.
+
+**Adam Stacoviak:** I'm over here, trying to vibe over here... I think it might be a Safari issue.
+
+**Jerod Santo:** \[laughs\] I thought you might be. That's why I pulled you back in. I can tell when Adam's like testing something out...
+
+**Adam Stacoviak:** I was vibing over here, and I think I might have hit an issue with Safari, potentially... Because the thing it made me is on a black screen, and so I'm having an issue.
+
+**Chris Anderson:** Well, that's not Safari, that's Claude. I would say you rolled the dice, and...
+
+**Jerod Santo:** You got snake eyes.
+
+**Adam Stacoviak:** Oh, dang. So it made nothing.
+
+**Chris Anderson:** One in 20 times Claude is like "Oh, I know what you want... You want a black canvas that's like 20 times bigger than your screen." It just does that -- we have work to do, but...
+
+**Adam Stacoviak:** Yeah, I think that's about it. It's kind of like forever scrolling.
+
+**Jerod Santo:** Well, close the tab, and open a new tab and try again, you know?
+
+**Adam Stacoviak:** I can share with you my -- it was just intense, man. This is an intense process. I saw it write a bunch of code, I could see all the code, but there's no application to be viewed.
+
+**Chris Anderson:** So basically copy that prompt you did and just make a new one. That's what we learned from vibe coding. Don't go iterate on something unless you're already stoked about it.
+
+**Jerod Santo:** Right.
+
+**Adam Stacoviak:** Yeah.
+
+**Jerod Santo:** It's just cheaper to throw it all away and start fresh, right?
+
+**Adam Stacoviak:** Yeah, right? I mean, when it's just spoken word, it's like "Let's try again and just tweak a little bit", you know? That's kind of cool, though.
+
+**Jerod Santo:** That's kind of what I do. Remember with Google searches -- remember when you used to google things? Some people will put in a search and then they'll read the results, and if there's not the right answer, they'll go to like page two, page three, page four... You know, like an infidel. I just research, right? Like "Oh, my search wasn't good enough." I just research... Now I have the same thing with these LLMs. I just copy a prompt into like four of them, and I'm like "Let's see who's got the best answer", or the fastest, or whatever. Or I can like mentally munge them into one answer and like pull things from each one... And that's kind of the new research, is this like "Nah, I'm not going to tweak this prompt. I'm just going to try it on more things and see what happens." So yeah, close tab, open new tab, paste prompt and start vibing again.
+
+**Chris Anderson:** Yeah, we're considering seriously doing that for you. We may as well just vibe three of them in parallel and let you pick the best... Although every bit of complexity we add is going to take the simplest user and turn them away.
+
+**Jerod Santo:** \[29:57\] And cost. I mean, that's what Sora does, which is Open AI's video tool. You can pick how many different iterations that you want, and it will make two or four or six. And obviously, you have to have the tokens, and stuff... But especially with creative tools, where it's like "I would like to see four different things", it's just like "Yeah, we'll just crank out four", versus you having to say "Try again, try again, try again."
+
+**Chris Anderson:** The fun part is looking at how much better Claude 4 is than even 3.7. Or if I switch it over to Codex Mini, then I'll get like an app that works great, but isn't showboating at all. Like, GPT models make these lean apps, with no extra stuff... And Claude 4 is going to have your backgrounds, transparency with blur, and \[unintelligible 00:30:41.20\] It's fun to see the personality in there.
+
+**Jerod Santo:** I think Claude 4 - and I don't have any personal reason to say this in terms of like we're not making money by promoting Anthropic... But Claude 4 is just I think good enough now that it's sold me on the entire thing... Whereas up until now I've been very skeptical, because of the results have been so bad. I'm not sure if you guys saw Simon Willison's "The last 12 months of LLM progress as outlined by pelicans riding on bicycles." Did you guys see that post?
+
+**Chris Anderson:** Oh, yeah...
+
+**Jerod Santo:** Because he's such a nerd for every new model that comes out that he tells them all to do the exact same thing, which is to write a code that produces an SVG drawing of a pelican riding on a bicycle. And he documents that, of course, because he documents everything. And he's just showed the progress; he was like "Here's this model, here's that model", all the way down through the last year, and how far they've come and which ones do better. He scores them, and stuff like that. But the progress has been amazing. And Claude 4 - I think we're at a point now where it's like above my threshold of sucking, to where it's like "Let's do it." Like, I want to use them now. I'm not sure what happened between 3.7 and 4.0, but something happened.
+
+**Chris Anderson:** That black screen that you ran into, Adam - the other one that I'm just still trying to teach Claude to never do again... For some reason, it loves to give you buttons with white text on white background.
+
+**Jerod Santo:** Dark patterns. The opposite of dark patterns. Light patterns. Did you get better results this time, Adam, or are you still spinnin'?
+
+**Adam Stacoviak:** It's still working on it. It's a little slow, for some reason. It's taking its time.
+
+**Jerod Santo:** It's bleeding edge.
+
+**Chris Anderson:** How many lines of code does it want to do?
+
+**Adam Stacoviak:** At this point it's trying again... So I got two black screens in a row, so I wasn't sure...
+
+**Jerod Santo:** What are you trying to build?
+
+**Adam Stacoviak:** Okay, so I told this thing "I want to golf with friends. Help me create a tee time and invite folks to it." So "Hey, Jerod, I'm going to golf this Saturday, at this time. You're invited to my tee time." And you can join, kind of thing. So that's -- oh, it finally did something. It's got a golf tee time organizer, I can create a new tee time, I can invite people to it, it seems... It looks like more iteration. It's got some additional features that could be added, which I think is kind of cool... But all in all, quite colorful. It's vibrant. It's got a cool vibe. It's vibing. This is vibing, man.
+
+**Jerod Santo:** This is vibing, man.
+
+**Chris Anderson:** Right?
+
+**Jerod Santo:** So do you guys have a system prompter? I'm not sure if you have some sort of scaffolding in there, that when we write our prompts and put it into vibes, it does more than just that, right?
+
+**Chris Anderson:** The whole thing's open source, so you can go on GitHub on the Vibes DIY org and get into the repo prompts.ts and see what we actually do. But it's fairly simple... I mean, as I was mentioning earlier, I was one of the coauthors on the O'Reilly ChatGPT shortcuts book, and what I learned from that - and the experience continues to pan out... If you talk to the LLM like a person, less technical, more just what you want, you get better answers. So this golf thing is the thing you want to do. There's this great YouTube - we can find it for the show notes - there's two types of vibe coders. And there's the one guy who's all frustrated, with his 4,000-line LLMs text and his feature specification, PRD... And the other guy who's just like "Make GTA. Make GTA six." \[laughter\]
+
+**Jerod Santo:** \[34:13\] And he's having a \[unintelligible 00:34:15.27\] you know?
+
+**Chris Anderson:** Here's my golf tracker... \[laughter\]
+
+**Jerod Santo:** Oh, nice. Did you do that one just now, as Adam was doing one, or is that --
+
+**Chris Anderson:** Yeah, after I heard what he was doing... And I think my prompt was just --
+
+**Adam Stacoviak:** Let's see your prompt. What is it again? Show it to me?
+
+**Chris Anderson:** I said "golf with friends."
+
+**Jerod Santo:** That's all you said?
+
+**Chris Anderson:** Yup.
+
+**Adam Stacoviak:** Oh, my gosh... Okay, so how do you share that with me, so I can see it, or play with it, or remix and revibe?
+
+**Jerod Santo:** Pass the address into the chat. We have a chat here.
+
+**Chris Anderson:** Right?
+
+**Adam Stacoviak:** That's how it works?
+
+**Chris Anderson:** So again, we don't deploy. You press the purple button, the big button, and now you publish it, and... I was getting into this a minute ago, I guess I distracted myself... So when it publishes, all it does is take, like, let's say the 475 lines of JSX and put it in a CloudFlare KV associated with the subdomain name. When you go hit that subdomain, we have a static HTML that's essentially the runtime, that's got all the packages in there... And it loads any additional packages via esm.sh. And now you have your app right there... It costs us almost nothing to host, almost nothing to serve, unless that app goes and does some AI calls, in which case that goes against your token quota. So it makes it super-easy for people to write apps that are alive with AI, whether that's... Our benchmark, compared to Simon Willison - he's got the penguins on bicycles; what we have is, for instance, I'll just tell you something in English, and you'll give me a structured to do list that reflects back what I need to do to accomplish that. So I'm always running that when we do an upgrade, or move to a new model... This morning I said "Take my kid to summer camp", and it gave me everything: sunscreen, lunch, all that.
+
+**Jerod Santo:** Alright, I just loaded up Adam's golf tee time organizer app.
+
+**Adam Stacoviak:** Is it working?
+
+**Jerod Santo:** I mean, it rendered... I don't know if it's actually working. I think it has a white button with white text, as Chris knew it would...
+
+**Adam Stacoviak:** Something's going wrong here, because I can't get it to load for me.
+
+**Jerod Santo:** But I can create a new tee time...
+
+**Adam Stacoviak:** Oh, sweet.
+
+**Jerod Santo:** Oh, I see some. Sunnyvale Golf & Club. I can join that round.
+
+**Adam Stacoviak:** Yeah. Pop your name in there. Boom.
+
+**Jerod Santo:** Okay, I'm joining that round... My name is Jerod, thank you very much... And Confirmed. Jordan Phillips is going, and Jerod is going. So... Where is it storing that information, Chris? Where does the data that I enter in go?
+
+**Chris Anderson:** So all of that stuff goes into Fireproof, which has got that Merkle CRDT. And that means that it just goes in your browser local storage. The way Fireproof works is every database operation is a diff, like in Git. It is end-to-end encrypted and then written to - we use IndexedDB in the browser, because it's real fast with Uint8Arrays... But then asynchronously can be replicated via S3 bucket and WebSocket. So that part works in Fireproof. We're rolling it out in Vibes right now. That means today all Vibes are single player. That's why it's best for these kind of viral AI experiences. But tomorrow, all Vibes have essentially the security model of a Google Doc. So I don't want an end user to have to understand Postgres Row Level Security, or having to do like where clauses to keep from writing bugs... Instead, if you're in the app, you're in the app. Maybe you have write access or read access, but it's as simple as a Google Doc for the sharing. And that means as soon as we activate that feature, then Adam, you could invite Jerod to your app via email, and now y'all are both in there, setting up what are you going to bring for lunch to the course.
+
+**Adam Stacoviak:** It's all on the web platform.
+
+**Chris Anderson:** \[37:58\] Yeah, yeah. None of the people who don't know what code is, who've never deployed anything - they can't depend on any setup. I like to say, if you don't know what an API is, then I'm certainly not going to ask you to do an API key.
+
+**Jerod Santo:** So this thing went Render ATL viral last week, because they already had a fun time making caricatures of themselves... But when's it going to go TikTok viral, or youth viral? Like, it seems like this is set up specifically for 13 to 15 year olds to find out about it. Just like the six, seven thing. I'm not sure if you guys are aware of six, seven, but they're huge on six, seven. And all of a sudden it's like, this thing's screaming.
+
+**Chris Anderson:** Well, I'm pretty sure that my key into that market is you guys. \[laughter\] But really, what we aim to be -- there's a lot of people who use more serious tools for vibe coding, and I'm not going to give a senior developer more control, more options, more details than other incumbents. But what I will do is be exactly the right thing that they're going to turn around and show their friends when they want their less technical friends to start getting into writing code.
+
+**Jerod Santo:** Gotcha.
+
+**Chris Anderson:** So if you've got a buddy who has a dream, or keeps bugging you to build something for them, say "Hey, vibe it up and I'll take it from there."
+
+**Jerod Santo:** Right. Well, it's very much a GeoCities then, in that sense. Or maybe Glitch. Of course, Glitch recently changed the way Glitch works, but... Similar vibes then, right? Glitch was very much the thing where it's like get people started, learn on it.
+
+**Adam Stacoviak:** How does versioning work in that multiplayer world?
+
+**Chris Anderson:** So if you look inside the chat that you've been having, you can refine it. You could say "Oh, I didn't mean regular golf. I have these interesting rules of golf I want to bring to the game." Claude will go rewrite the code for that, and now you have a slightly different version of the app. That could be within a single chat session, in which case we're not remixing yet. We're just iterating. But if you go look in that chat, you can click on the older versions of the code and it brings them up instead. Whichever one is on your screen is the one that gets published when you click Publish. So publish an old version if the new one wasn't as good...
+
+But once it's published and you go to that URL... You know, when you scan the sticker and you go to the URL, then you're going to have a Remix button in the corner of the screen. And when you hit that, you've essentially forked. Now, no insult to all of the people out there who are kind of like me in this regard, but... I've been forking forever, and what we say to ourselves to make sure we stay on target at Vibes DIY is we say "Forks are for dorks."
+
+**Jerod Santo:** \[laughs\] These guy's got a saying for everything. So because you don't like forks, or because you do? I mean, it depends on if you think a dork is a good thing or a bad thing.
+
+**Adam Stacoviak:** That's right.
+
+**Chris Anderson:** Well, when it's time to get the dorks involved and do the forking. But when you just want to try something and remix, then you don't need all that conceptual baggage.
+
+**Jerod Santo:** Gotcha.
+
+**Chris Anderson:** Just hit the Remix button and go.
+
+**Jerod Santo:** So this is akin to the publish button versus the deploy, right? It's like, remix it, don't fork it, because you're not a dork. You're a normie, as you said earlier.
+
+**Chris Anderson:** Yeah. Yeah.
+
+**Jerod Santo:** Well, your audience, and it's not us. It's not us... \[laughs\] Because I'm looking for the fork button over here. I'm like "Excuse me, sir... I would like to fork this."
+
+**Chris Anderson:** Totally. Well, it is us though, because when you go deep on something, it's got that Merkle CRDT, it's made out of standard React components, it's got all the TypeScript annotations you could hope for... And it's ready to go big, if that's what you want to do with it.
+
+**Jerod Santo:** But I imagine sitting on Cloudflare, with an HTML file and some KV backends; big in terms of features and commerce and production big, business big, but not big in terms of scale. I imagine one of these could scale pretty easily in terms of that TikTok viral moment, couldn't it?
+
+**Chris Anderson:** It costs pennies on the dollar compared to anything that requires a VM or any kind of complex state or CPU. When I think of Kubernetes containers, all that, I only want one if I basically want to rent some live RAM.
+
+**Jerod Santo:** Right.
+
+**Chris Anderson:** \[42:13\] And you don't need that. All the live RAM lives in the user agent at the edge.
+
+**Jerod Santo:** Right. So Cloudflare is pretty cool... I mean, just - no further comment. I wonder if you could maybe expand on that. Like, you guys are doing cool stuff because of what Cloudflare is helping you to do, right? It seems like a very interesting platform in that way.
+
+**Chris Anderson:** Our stack is, I guess, native to that way of thinking about things... So not just the publish flow that goes into a KV, but also the fireproof synchronization is powered by Cloudflare workers, which is a great way to rent some RAM on the cheap if all you need is a list of your currently connected clients. So these -- essentially, the flow is that tiny encrypted diff, you write it locally, push it up via our API, which you don't have to know about, but it ends up as another record in an R2 bucket... And then an update to the pointer that corresponds to your database. So everyone who's listening to that pointer pulls down the latest. And because it's a CRDT, the merges are resilient.
+
+I was thinking about recovery mechanisms. Say like you lost the pointer, and all you had was the files... It's a valid exercise to just load them all and merge them, and you'll get the same result as if you'd started from the end one.
+
+**Adam Stacoviak:** Why RAM? Why does it need to be in RAM?
+
+**Chris Anderson:** I mean, it doesn't. We can do it with the S3 bucket only.
+
+**Adam Stacoviak:** It's cool, right? Because it's cool.
+
+**Chris Anderson:** It's a little faster to keep it on \[unintelligible 00:43:43.19\] And what I mean by in RAM is --
+
+**Adam Stacoviak:** Is it necessary, though? Or is it just because it's cool?
+
+**Chris Anderson:** It's just because it's faster. So we need essentially a directory, a link to -- okay, let's get technical. The real reason we need any of that is because it's multi-writer safe. And so if you just had like a SQLite and you were trying to do this kind of architecture - which people do, where they put the SQLite segments into an S3 bucket, and then they load them in the browser and query them... That's all good until you need to start writing. And now you're doing some kind of bespoke, you know, take the write-ahead log, and put it in there, and manage kind of conflicts against it and everything... It's not for the faint of heart. Instead, what we do, because of this Merkle CRDT, the merges are handled for you.
+
+So the only thing we need to add to the recipe to make it multi-writer safe is it could be that pointer that points to the most recent file, the diff from which you've got to start to reconstruct - well, it can point to more than one file. So if you're under concurrent load, people can all -- you have 50 people all write on top of the same starting state, without having coordinated each other yet... And now we just have a pointer that's 50-wide. The next read pulls those in, merges and writes back up the single pointer. So that's what we do the performance thing for.
+
+**Jerod Santo:** So in your mind or in your heart, is Fireproof the thing, and Vibes DIY is a thing to demonstrate the thing? Or is Vibes DIY the thing that Fireproof let you build?
+
+**Chris Anderson:** I feel like I'm in the Millennium Falcon and I just like mashed the gas, you know?
+
+**Jerod Santo:** \[laughs\]
+
+**Chris Anderson:** So it's a little bit of both, but it's it's super-aligned. Fireproof is for the kind of people who like to fork stuff, and Vibes is for the kind of people who --
+
+**Jerod Santo:** And Vibes is for remixers.
+
+**Chris Anderson:** ...don't want to be bothered with all that. It's just done.
+
+**Jerod Santo:** Okay. Do you think you're gonna have to pick a certain point, or like focus, or do you have infinite bandwidth to just like mash that Millennial Falcon - what's it called? Warp speed? No, they don't call it warp speed. What do they call it in Star Wars? I'm blanking.
+
+**Chris Anderson:** The hyperdrive?
+
+**Jerod Santo:** Yeah, the hyperdrive. Go to hyper speed.
+
+**Chris Anderson:** So we are focused 100% on Vibes.
+
+**Jerod Santo:** Okay.
+
+**Chris Anderson:** \[46:04\] And in a lot of ways, it's because Fireproof is basically done. I mean, it never is the case with a database, but it's not like it needs new features. It's just about hardening it, and continuing to release it. So in that world now, Vibes gets cool new features. A fun new feature we get to add - this is lightweight stuff. Like I said, I don't want people to have to know what an API is, much less an API key. So... Bake that YouTube API key in there. It's not in there yet, but this allows you to say "Hey, I want to turn my playlist into a YouTube screen", and it just works already. Same thing for Spotify, or any other of these kind of mass market media APIs. There's no reason that you as a Vibe coder should have to know about any of that to use it.
+
+**Adam Stacoviak:** Have we talked about money yet?
+
+**Jerod Santo:** No.
+
+**Adam Stacoviak:** How do you make money?
+
+**Jerod Santo:** No.
+
+**Adam Stacoviak:** Let's talk about that.
+
+**Jerod Santo:** Let's talk about it.
+
+**Adam Stacoviak:** Does it matter at this point? Or is it just for the vibes?
+
+**Chris Anderson:** It matters that we don't get it wrong... And then the more we think about it, the more excited we get. So if you look at what it would take to build something with a vibe coding tool and a database backend - those are both kind of like maybe just to get started; table stakes, $25 a month, $25 a month - that's $50 a month right there. We do an all-in-one bundle at $5 a month. And that's because all we need is that R2 bucket. It's not costing us a whole Kubernetes container to keep your Postgres instance running. It makes pennies on the dollar for us. And then the real cost becomes those tokens, that either your app consumes when it's generating cool caricature images, or parsing your to-do list, as well as the tokens that you use when you write an app.
+
+$5 a month gets you started. If you want to go over that, there's a meter which is competitively priced. And then where we see this going in the long run is almost like this YouTube-style model, where, say -- first, think about a corporate case.
+
+Think about some big marketing agency that builds like "See our product in your house. Replace the Statue of Liberty with our product", whatever it is. "Take a photo, put us in there." You get 100,000 views on that from your email list... You're probably writing us a check for $20,000 or something on the tokens it consumes. But if you're some kid who figures out a cool way to scan your house and turn it into a video game level, you don't have that kind of money. But your app costs just as much to run, and it goes just as viral. So in that world, there's a little bit of a freemium. Somebody comes along, uses the app - they can generate a simple level or a couple of images. But then they're going to get through their set of free vibes tokens... If they like what they're experiencing, then now they're subscribing for $5 a month. So we're not just selling it to developers, we're selling it to people who want to make cool caricatures of themselves in their dream job, or whatever it is that these end user apps do. So now if you're that kid who made a cool video game and you don't have any money, but you get us 50,000 signups, then my goal is to write you a check, like YouTube.
+
+**Jerod Santo:** Does that "Turn your house into a video game" level thing exist? Because I want that.
+
+**Chris Anderson:** No, somebody's got to make it. I mean, why wouldn't it exist?
+
+**Jerod Santo:** That's such a good idea.
+
+**Chris Anderson:** Right? Gaussian splat is cheap now, running on the phone.
+
+**Adam Stacoviak:** How would you do that though? Would you just walk around your house just like taking a video, and uploading it?
+
+**Jerod Santo:** I don't know.
+
+**Chris Anderson:** We like to say that the best we can hope to do -- you know, we're kind of grizzled, old vets of the industry. The best we can do is invent the electric guitar. It's up for the kids to invent rock and roll.
+
+**Jerod Santo:** That's like the best non-answer ever, right? He's like "You guys figure it out."
+
+**Adam Stacoviak:** Alright...
+
+**Jerod Santo:** Yeah.
+
+**Chris Anderson:** \[49:51\] I mean, we've got roadmap features that expand. So everything I've talked about is really just the browser runtime. It can do a lot. It's underutilized, especially for people who really haven't pushed it to its limits yet. But roadmap features is something we learned from the CouchDB ecosystem. Essentially, Fireproof is the CouchDB model running in the browser. And one of the things that you see a lot of -- the way npm was powered, all those packages stored in CouchDB, and then when someone uploads a new one or makes a change to some metadata or whatever, there's just an event reactor subscribed to the database feed that can then go do some heavy lift, or send an email, or update some analytics package somewhere.
+
+Same thing, just -- if you wanted to today, you could stand up a CloudFlare worker that subscribes to that golf database. And when somebody puts in a new tee time, you could get a push notification. Our roadmap brings that to the masses by making it so there's just a backend.js that we also let the model write.
+
+**Adam Stacoviak:** Yeah. See, I wouldn't have even thought about that, I guess... It'd be helpful i there's a guide. "Here's the next step you can take." They mention different features that can come, but not what you mentioned there, which is like push notifications. That'd be kind of cool.
+
+**Chris Anderson:** Right. And there's no end to that kind of stuff. And we really -- essentially, if I think about the job you're hiring that backend.js to do, it's the stuff that's essentially hiding API secrets. You can't just publish the way that you send your emails, your Postmark key to the world. You can't put it in the browser, but you could put it on the backend, and then the backend's job is just to forward those messages.
+
+**Adam Stacoviak:** What happens when it's successful? What is a year, two years from now -- like, how does this change things? What is vibes.diy a year and a half from now?
+
+**Chris Anderson:** Well, now that it's easier to make an app than a TikTok, we picture people swiping right on the good apps, and long-pressing to launch it, and having essentially their phone blow up in their hand when their app gets popular. So the same thing you see with an Insta, or a TikTok or a YouTube, where it's really about the excitement you get from coding. That's what I saw in Atlanta.
+
+When I was first starting coding, the coolest thing in the world was just to change one line in your HTML and then hit Refresh and there it is. So educating people how that feels, and then letting them share that with their friends... I think what we see is that coding becomes something that is cool.
+
+**Adam Stacoviak:** How do you deal with the potential of scams, or just like security issues in that unfettered world where it's like "Yeah, check out my vibe app", and it's cool, but like what happens when it's a scam?
+
+**Chris Anderson:** There's thankfully a whole bunch of prior art. Essentially, for us dev tool builders it's a little bit of alien land, but if you've done social media before, it's that playbook. So we've done some of the basic things already... For instance, there's no way to publish to a fire hose. It's not like just everything's all showing up. We moderate the stuff that goes out on the homepage. That doesn't mean you can't share something. So if you start building honey pots, then now we have a different moderation challenge. It's not new, it's just new to do it with HTML instead of video, or images.
+
+**Adam Stacoviak:** Yeah. I just wonder... It's one of those things where you don't want to have to deal with it, but you do... And it's early, because trust will be everything to get to the TikTok stage of sharing, and virality, or whatever... It's going to be something you can trust, or has to be. And if you can't trust it, then it's like -- it's a today problem, basically. Like, at launch.
+
+**Chris Anderson:** \[53:49\] Yeah. So if you look at the history of using Mechanical Turk to moderate profile images, and having to do two of three thumbs down, and all that - so much cheaper, so much easier to do that with a model now. Just let Claude tell you whether it's not any good, you know?
+
+**Adam Stacoviak:** There you go.
+
+**Chris Anderson:** And then on top of that, we aren't going to run into these problems to -- it's sort of blunted. We don't give you the ability to go into the editor and type. It wouldn't be hard for us to, but we figure that that's the point you should copy and paste it somewhere else. And Claude's not going to write something out of the box that's going to be extremely concerning. We just use the Open AI image gen. It makes our job easier. Eventually, we'll want to value-engineer all that stuff, but that comes with scale.
+
+**Adam Stacoviak:** Solve it with more AI. That's interesting. Right...? All problems get easier with AI...
+
+**Chris Anderson:** Well, AI makes new problems faster, then it makes the other ones easier, but hopefully we've isolated ourselves from that with this one file app JSX thing. And don't underestimate the power of the single-origin policy. The web browser kind of already worked out a lot of this stuff.
+
+**Jerod Santo:** So if this thing takes off - I mean, now's the time to get in there and really be the vibe DIYer in chief, right? The people that were early on YouTube, the people that were early to TikTok, they had a much better time... Is there actual virality built into it? Can I build an audience? Can I distribute? Is there a distribution channel? I know y'all are just getting started, so the network effects probably aren't there yet, but are the tools to build an audience for my apps in the product?
+
+**Chris Anderson:** So if we're talking 90 seconds from prompt to app, on the cases where it works, probably another 90 seconds to be like "Is this app any good?" and then hit Publish. And now it's on a URL, so we go wherever URLs go. But the built-in features around like a newsfeed, and a For You page, and all that... These are the things I'm super-excited to build.
+
+We're growing the team right now. People who want to do that should probably talk to me about it, but it's definitely... It's a kind of interesting green field, because so much of the design language is already worked out. It's not like we have to decide what an algorithmic feed even is. It's more just about making it for -- basically, we're bringing a new media object to the world.
+
+**Jerod Santo:** Right.
+
+**Chris Anderson:** And the same thing you do with a video, you do with this. And we even have plans -- for instance, the same Jason Smith I mentioned earlier, he used to host npm. I talked to him about it yesterday and he said "Hey, I've been working on some video gen stuff. I think you could --" Lik, we have this demo data button that fills the app up with data for you... You could also do a video tour button. You just click one button and it takes your face, lip syncs, puts it on TikTok, "Here's my app."
+
+**Jerod Santo:** Right. That's cool. So what kind of engineers or people are you looking for?
+
+**Chris Anderson:** Really right now if you go use Vibes, you're totally allowed to roast it, because I am a React developer, I guess, but I'm also a CEO... And it's like 75% CEO code right now. So we've got --
+
+**Jerod Santo:** C-level code. \[laughter\] I like that.
+
+**Chris Anderson:** So if you're just like absolutely cracked, and the person hire is -- you know, when somebody is kind of like new enough to a stack where they're like dangerously enthusiastic about getting everything right... If you're senior enough, you're going to see my flubs and be frustrated, but at the right part of your skills curve, you're going to just like chew up fixing that stuff, and get it into best practices.
+
+**Jerod Santo:** Every level of CEO code is like an opportunity for the right person, you know?
+
+**Chris Anderson:** Yup.
+
+**Adam Stacoviak:** I just got like WordPress vibes, in a way. Like, in a positive way. When I was looking at this golf friends thing I got built out here, it's like Kubrick. Like K2 was to WordPress. Very -- every app has a certain style, kind of thing to it. I kind of feel like it's a good thing.
+
+**Jerod Santo:** Is there a shared -- is that on purpose, Chris? Is there a shared design aesthetic, or something?
+
+**Chris Anderson:** \[57:59\] Well, we made the choice to differentiate ourselves from everybody else. We're not using Shadcn... And that's not because it's not great. We just want to be different. We want to be more lightweight. We want to have a little bit more variance and let Claude or whoever express themselves a little more in there. So the way we prompt the design language is -- actually, my co-founder Marcus I mentioned earlier just did a bunch of research, like art history type research around like, you know, "We don't want synth wave. That's played out" etc.
+
+We've found -- there was a design school in Italy in the eighties that came up with a style they called Memphis. And so we tell it to make Memphis... So we described Memphis a little bit, and it does a pretty good job of that. That's all user editable. So if you go into the settings page right now, you can just type in your own style prompt. You could switch it to synth wave. You can switch it to synth wave and DeepSeek. And then if you run it on DeepSeek now, you probably have to generate the app 10 times before you get a working one, but... \[laughter\]
+
+**Jerod Santo:** Certainly a good playground for testing out these models, and comparing them to each other. Probably better than my copy-paste prompt into like four thing is actually making -- probably better than Simon Willison's pelicans on bikes, is like make the same app against all these different models.
+
+**Adam Stacoviak:** That's cool. I see the style prompt now. It's pretty cool, you can just like swap that out, make it whatever you want. That's so cool.
+
+**Jerod Santo:** It's your own.
+
+**Adam Stacoviak:** Yeah.
+
+**Chris Anderson:** Right. Or if you really want to go deep, just hit up our GitHub, fork the whole thing - you know who you are - and fix up my React.
+
+**Jerod Santo:** Yeah, I was gonna ask you that... Like, how much of this can you run? How much of this is just completely "if I had enough gumption, I could get it running on my own Cloudflare account"?
+
+**Chris Anderson:** So... Kind of yes. What we have set up is even for - if you're going in via the GitHub, it deploys to our end point, and it uses our login system... And that means that you could have projects -- basically, we want you to be able to be a Vibes DIY user, go run a fork, maybe your fork is like internal for your work or something, but not have to do the whole heavy lift of the full stack. The fork will just run against our backend. If you want to get even heavier than that - yeah, now you can start to code up the backend... But that's a senior engineer task.
+
+**Jerod Santo:** How do you keep up with your underlying technologies in terms of what you're generating, or what you're telling Claude to use? Because those things are also moving targets.
+
+**Chris Anderson:** I have more work to do there. There's models that aren't represented yet. I would love to get \[unintelligible 01:00:27.09\] personally when I'm in Windsurf... And so, there's some things we can do to keep up. We essentially need like a cultural ambassador to LLM land to keep us up.
+
+**Jerod Santo:** What about to Next.js land? Is that a fast enough moving target? I guess you're always generating a new one, so it doesn't matter so much.
+
+**Chris Anderson:** Right. And we're just doing that one file. I chose to do app.jsx instead of app.tsx, mostly because it's one less thing the model can hallucinate wrong. Now, if you wanted it to be TSX, I'm pretty sure Windsurf can fix that for you in a single pass.
+
+**Jerod Santo:** Sure. But why? I guess by then you copy-pasted it, so it's yours anyways, right? Do what you want.
+
+**Chris Anderson:** Exactly. Yeah, the JSX is all about the simplicity, and then the -- it really allows anybody to get in there, edit these apps, and bring them to any backend. When you do eject, it runs through those same APIs. Again, you can move them over. I don't know if there'd be a good reason, but there's a world in which, especially when we spin up the user pays mode for like image gen AI usage by the app, you might just pull our npm modules directly into whatever environment and use those, because that lets you ship apps without having to worry about the cost of the LLM.
+
+**Jerod Santo:** Say that again. I don't think I tracked it.
+
+**Chris Anderson:** \[01:01:52.12\] So built into these npm modules that's part of our standard library is that when you run out of your free tokens, it pops up the login window. And after you log in and you get another batch of free tokens, you run out of those, it asks for your credit card. If all you care about is going viral and the thing that's holding you back is that going viral is going to cost you $50,000 in ChatGPT fees when people start using your prompts in production - well, do it on Vibes or do it with the Vibes npm modules and you'll get the same benefit that we'll handle it for you. And if you get enough usage, then you can be looking at creator payouts.
+
+**Jerod Santo:** Yeah, that actually scales really well. I guess that's what you're referring back to the YouTube model, where basically as an app creator, I'm bringing you token use by way of people using my app, and you being Vibes DIY is making money off of every one of those people who decides to put down some cash in order to use the app some more. And so you're now talking payouts to me for bringing you guys additional customers. That's an interesting model.
+
+**Chris Anderson:** A real kind of eye on the prize... You know, Couch did well, but I also learned firsthand what it's like to get whooped by MongoDB. \[laughter\]
+
+**Jerod Santo:** How did that feel?
+
+**Chris Anderson:** At the time, I wrote that rap about other NoSQLs, right? You could tell how it felt. But now I appreciate what they did; they understood that the real answer to taking the market is to be on the ground floor when people are writing apps without permission in the enterprise. And so we're probably a decent choice for say you're running a veterinarian clinic, and your customer management system doesn't keep track of what the dog's favorite treats are... Vibe up a treat tracker... And on down the line the cryptographic causal consistency that's enforced by Fireproof means that you could use this for regulated supply chain. Maybe your cannabis dispensary needs to keep track of some stuff, because someone with a badge is going to come knocking... And Vibes has got the underlying infrastructure to make that safe. So after we make coding cool, we have a lot of bandwidth to go and make this system scale for serious use cases.
+
+**Jerod Santo:** Do you think your brand scales that way?
+
+**Chris Anderson:** We'll find out, but I kind of think so.
+
+**Jerod Santo:** Here's a sub-brand: Serious Vibes.
+
+**Chris Anderson:** Totally. Yeah, we talk about gray mode... \[laughs\]
+
+**Jerod Santo:** Gray mode? There you go. Yeah.
+
+**Chris Anderson:** There's a lot we can do there. There's a whole --
+
+**Jerod Santo:** It's cool stuff.
+
+**Chris Anderson:** ...like, if you wanted to fork the core thing and make it work with, say, not React... Like, I've got nothing against Vue. I've got nothing against Solid.js. All these things are excellent technologies. We chose React because it's the one that's going to give you the deepest training set. But being able to take the wrapper and deploy a Vue app or a React app to our Cloudflare endpoint - it's not going to change things. From an infrastructure standpoint, it's ready to go.
+
+**Jerod Santo:** Alright, dorks, get out there and fork away... For the normal people, get out there and remix somebody else's golf tee time tracker...
+
+**Adam Stacoviak:** Hm.
+
+**Jerod Santo:** Adams has already published it.
+
+**Adam Stacoviak:** And improve my vibes. My vibes were not so viby.
+
+**Chris Anderson:** Get your friends into coding. Get some -- make an image generator that no one would have expected, and make it go viral on your Facebook.
+
+**Adam Stacoviak:** I've got a phrase for you. Ready for this?
+
+**Chris Anderson:** I love phrases.
+
+**Adam Stacoviak:** Vibe with friends.
+
+**Chris Anderson:** Vibe with friends. That's right.
+
+**Jerod Santo:** There you go.
+
+**Chris Anderson:** \[01:05:54.24\] And that's coming -- I mean, that's just an easy on switch... Some of the folks we've been talking to are most excited about that this chat interface where you're talking to the LLM and it's writing a code for you - that's multi-user. As soon as we flip on the same sharing mechanism that allows Adam to invite Jerod to the golf app, you can also invite Jerod to be the coder on the golf app, and y'all can be in the chat together.
+
+**Jerod Santo:** Now you really are vibing with friends.
+
+**Adam Stacoviak:** There you go.
+
+**Jerod Santo:** Awesome, Chris. Anything else that you expected us to ask you that we haven't asked you, or that you can't wait to tell us about before we let you go?
+
+**Chris Anderson:** Y'all did a great job. I just can't wait to see what you vibe up.
+
+**Jerod Santo:** I'm going to vibe some stuff. I'm going to get my kids vibing... This is going to be fun. Vibes.diy, if we haven't said it enough times on the pod. That's your URL.
+
+**Adam Stacoviak:** Vibes.diy.
+
+**Jerod Santo:** There you go. Tell your friends. Get your kids, get your nieces and nephews - if you're of age similar to ours - out there and do it yourself. I mean, I'm going to throw a few things at this. I think the image gen, like the caricature stuff - do you have a gallery of where... Because sometimes one idea just strikes. You've gotta have something to start with and you're like "Oh, I can just change that", like the remix thing. Do you have a gallery out there?
+
+**Chris Anderson:** We moderate the stuff that shows up on the homepage. So we essentially have a carousel. Every time you hit refresh on the homepage, you'll get some new choices.
+
+**Jerod Santo:** Gotcha.
+
+**Chris Anderson:** There's some ones that illustrate how you do real things. If you have in your mind, as many of us dorks have, a mental list of APIs that have cores and don't require an API key, I've got for instance a vibe on there that is using Algolia's hacker news search to just bring you the vibe code news. So go remix that. Any API that's got that kind of access control - super-easy to use. And then, as I said, with this backend support coming soon, you'll be able to drop your own API keys in.
+
+**Jerod Santo:** And what's the best way to get in touch with the creators of this, yourself, your team, your community? Is there ways that you can say "Hey, I'd love support for this, that, or the other thing" and you guys don't have it yet? Feedback, or a Discord, or a Zulip, or anything.
+
+**Chris Anderson:** It's in the Discord, and the Discord is linked from our About page.
+
+**Jerod Santo:** Alright. Go forth and vibe code with friends... Especially once that multiplayer feature gets toggled on. Let us know, Chris, when that's out there, because we'll definitely give it another shot, and probably another shout-out, as Adam and I could vibe code together. I could be changing his design as he's trying to add features. That'd be fun.
+
+**Chris Anderson:** I'm looking forward to the day when your guests are on here with things they made in Vibes DIY.
+
+**Jerod Santo:** That'll be amazing. That will be full circle, won't it?
+
+**Adam Stacoviak:** That's what I'm trying to talk about. That's cool stuff. That's the way it's got to be done. I can't wait! I'm vibing. I'm vibing.
+
+**Chris Anderson:** Thanks, Adam. Thanks, Jerod. Super-excited to be here. And yeah, remember, everything we do is open source, so you're invited.

XO Ruby is hitting the road (Interview)_transcript.txt

@@ -0,0 +1,927 @@
+**Adam Stacoviak:** Well, we're here with a good friend of ours... A good friend of mine, at least, from back, back in the day, back in the Hashrocket days. This is OG, early days Ruby, Obie Fernandez, leading the charge, building apps on weekends... And that's where I met you, Jim, but you've been in the Ruby world for so long, and now you've got this roadshow you're doing, you've done conferences before, Madison Ruby... Welcome to the show.
+
+**Jim Remsik:** Thanks so much for having me. It's a pleasure to be here.
+
+**Adam Stacoviak:** What do you think about the past? What do you think about Ruby, where it's going, where it's been? Let's catch up a bit. What are you thinking about Ruby?
+
+**Jim Remsik:** We're still building our business on Ruby, and definitely there is a bit of a renaissance... In fact -- so we'll get to the conferences and whatnot, but one of the bits of feedback that we've been getting from people who haven't been focused on the community is "Oh yeah, that was really cool in 2010." And it was cool because the promise of being able to deliver a web app and a mobile app, and all in one codebase. And with Hotwire Native, that's coming back. Speaking with --
+
+**Adam Stacoviak:** Can you give like maybe a deeper detail? Like, it was 3, 2, 1... It was revolutionary. This was groundbreaking, and I think a claim to fame for Hashrocket, to be -- one, you had extreme talent. You had a great team to execute, but you had --
+
+**Jerod Santo:** What's Hashrocket?
+
+**Adam Stacoviak:** Okay, there you go. Let's go back to Hashrocket. There you go. Tell the whole story, Jim.
+
+**Jim Remsik:** Yeah. So Hashrocket was the first boutique Rails consultancy. And so we started in -- I actually got hired in 2007, before Hashrocket was a thing. There was a client involved, and then they decided to build a consultancy out of the team that they had built to develop that software. So we became Hashrocket... And sort of early startup days, lots of late nights, and lots of alcohol, and things like that. I've grown up, but -- yeah, it was an incredible time. We had people from various backgrounds all over the US that sort of assembled in Jacksonville beach, Florida, and we were just very excited about the technology. It was also right around the time that GitHub started, and Twitter... So it was early, early and pretty fantastic days.
+
+**Jerod Santo:** Yeah. Heroku... Adam, how did you know Hashrocket? How did you know Jim?
+
+**Adam Stacoviak:** Well, I was an up and coming Rubyist. I wanted to work there so desperately. I was like "Y'all are so cool..." The Hashrocket logo was the coolest, still is the coolest. It literally was the Ruby Hashrocket. Just all the right vibes were there. I actually met Obie Fernandez, who is I think a DJ, and now a dad, and growing up like the rest of us, and stuff... And we've actually wanted to get him on the pod a couple of times, because I think he's kind of bullish on like Ruby and AI... And I've been meaning to catch up with him, personally, in podcasts, professionally... But I knew of Hashrocket mainly through friends. I met Alan Branch, and Steven Bristol, and others... And I really don't know how I connected to Hashrocket. I just know that when I went to Future Web Apps, which was Ryan Carson's conference for web apps, which was a big, big deal then... This is, I think, in 2008, 2009. It was in Florida. And as a podcaster, you get invited places, and so Ryan invited us out there. I think we actually had to pay our own flight and hotel, Jerod, so we've grown up a little bit since then... But I was happy to go anywhere I was invited at that time, because I had nobody; I had no network. I was at the ground floor. And even that era was, I think... A lot of the networks and connections I built on - Jim, you're included in that - since then, really.
+
+And it was kind of funny, because when we were trying to get to the conference, we had a -- I think we had a rental car, or Alan had a rental car. And we were kids. When I say "We were kids", we were just young, and we were stupid. We did whatever we wanted. It was just a whole different world then. And Obie was in the car, and of course, I'm like fanboying, because Obie's cool. And we were podcasting in the car, and it's just like -- it was nothing like it is now, but it was so raw and so different. It was like "What is a podcast?" It was so weird. We were walking around with like short microphones, wired to this thing, and just like... It was epic.
+
+\[08:09\] But that's when I met Obie. We walked the beach in Miami beach, at night, and just talked, we got drunk, we did things we probably shouldn't do, like when you're young... And I really wanted to work at Hashrocket, because I was like "You guys are the coolest, you're doing this cool stuff... And if I come there for a month, or do like an engagement", they'd call them engagements, "I can get 25 grand for an engagement", which would seem like a million dollars to me... And I'd get to work with some of the most highest caliber talent out there. I'm like "Even if I'm not as good as they are, I will learn something through this engagement." I never got hired. It was so close... Obie did want me on the project, but it just didn't work out... And I think I got another gig and just moved on, or whatever. But that's how I knew Obie. I think Obie was kind of famous, to some degree, in the early Ruby days...
+
+**Jim Remsik:** For sure.
+
+**Adam Stacoviak:** He's written books, he's given talks, he's led the way in a lot of cases... And I think at some point Obie personally kind of like stepped back from the light. I don't know, really; I'm not sure of his story. You may know, Jim, because you work with him more closely... But that's how I was connected and how I knew Obie, and how I knew Hashrocket... And it was just such a -- these guys were rock stars, Jerod; you know this. They were rock stars, man, back in those days.
+
+**Jim Remsik:** Yeah, certainly. Steven Bristol, if you were tied to him, he and I were good friends.
+
+**Adam Stacoviak:** Yeah.
+
+**Jim Remsik:** And he was local to Jacksonville. He was on the other side of the city, probably 45 minutes away, but...
+
+**Adam Stacoviak:** And a Rubyist. Everything was built on Ruby.
+
+**Jim Remsik:** Famous in his own right, a Rubyist...
+
+**Adam Stacoviak:** Yes.
+
+**Jim Remsik:** And he and Alan were just delightful people. I still go down every time I'm in Florida, I try and get over to Panama city and visit Alan, who is now a mayor of Panama City Beach.
+
+**Jerod Santo:** I heard that.
+
+**Jim Remsik:** Yeah.
+
+**Adam Stacoviak:** So cool. So cool.
+
+**Jim Remsik:** Yeah. So the way that I got to Hashrocket was I worked at the state of Wisconsin department of financial institutions, I was in the .NET stack, and I saw this video about building a blog in 15 minutes... And then me and my coworker started diving head deep into Rails on the side. He was a little bit more fluid than I, and so he saw a post from Obie saying that they were looking to hire people, and he literally just flew down and showed up on his doorstep. That was John Larkowski.
+
+**Adam Stacoviak:** That thing rings a bell. Was that like Obie's right-hand person, and like just in every detail of Hashrocket?
+
+**Jim Remsik:** Yup. Yup.
+
+**Adam Stacoviak:** Yeah, okay.
+
+**Jim Remsik:** Absolutely. And so he came down there, they were pairing all the time, and they had four people, and then one of them wound up stepping away, and so they needed a fourth. And John was like "I know a guy." And so I went down and interviewed... I spent a week in Florida interviewing... Again, made some mistakes... But I came back home, told my wife that we're moving cross country, and went back shortly thereafter. We packed up the house and six months later we were living in Florida.
+
+**Adam Stacoviak:** Wow.
+
+**Jerod Santo:** How long were you there? Or are you still there?
+
+**Jim Remsik:** I'm not still there. We lived there for four years or so, and then the call of coming back home was a little bit too loud. So now we're back in Wisconsin. When I came back, I took what I had learned down there and I took part ownership of a consultancy in Madison, and have been basically -- I worked with them for two and a half years, and then started my own consultancy soon thereafter.
+
+**Adam Stacoviak:** \[11:48\] Ruby's been through some really interesting eras, I think. You know this, Jerod. You're a Rubyist at your core. I mean, I'm not speaking to the choir. I mean, I am speaking to the choir here, but... I feel like there was this -- the video you alluded to was DHH explaining how you could build a blog with his "Whoops" in 15 minutes, which was just groundbreaking... But I feel like that broke the mold for what was expected to get an application built. Now, deploying Ruby in those earliest days was super-challenging, and that's what open source is about, right?
+
+**Jerod Santo:** For sure.
+
+**Adam Stacoviak:** These things all came out of how to actually deploy and run an application in production. And it got easier over time with Heroku, and things like that, and Engine Yard... But this was such an interesting era for, I think, software. It's either interesting because it's interesting, or it's interesting because my personal history is tied to it... Like, the nostalgia I have as a 40-something looking back on my life is just littered with good days of Ruby, and this new framework called Rails, Ruby on Rails, to take us into this new thing... Which was building apps quickly, with CRUD, and scaffolding, and these unheard of ways to do things, easily, more quickly, and with more joy.
+
+What does that bring up to you, Jim, when you think about these olden days? Would you call them the good, old days? Is that how you'd frame them? The good, old days?
+
+**Jim Remsik:** I feel like when you say "the good, old days", that suggests that these days are not good.
+
+**Jerod Santo:** Yeah.
+
+**Jim Remsik:** Or like it elevates those old days. And not to say that those weren't good days. These are good days, too.
+
+**Adam Stacoviak:** For sure.
+
+**Jim Remsik:** Technology-wise. We've got our challenges in front of us, but whatever. I think -- yeah, that period of time was a lot of fun... But I think we also kind of put on those rose-colored glasses a little bit. There were things like Merb, and there was a whole "We're going to fork the codebase, and people are going to go in a different direction for a while", and then things got resolved... And things got better for everybody. So yeah, there certainly are a lot of moments in there, both at the community at large, through my experiences and whatnot... And most of them were really good.
+
+**Jerod Santo:** And you've never left. You're still doing Ruby, still running Ruby-related meetups and events... Has it been consistent for you? Like, from 2010 to 2025?
+
+**Jim Remsik:** Yes, it has. I think the thing that changed was events; the pandemic got rid of events, for a period of time there. And so we did Madison Plus Ruby from 2011 to 2018. We actually stopped just before the pandemic. But we had done a Madison Plus Ruby event in Chicago, we had done a Madison Plus Ruby event in Los Angeles... So it was traveling with those back then as well. And then the pandemic just sort of shut everything down.
+
+We moved -- I think like a number of people, we moved out of a city center, and we're sort of out in the middle of nowhere right now... And I don't have a local community, so I've had to sort of seek it out. And that's a little bit the genesis of the events that we're looking forward to over the next two months, which is XO Ruby.
+
+The feedback that I constantly get from meetup organizers is that it's hard to get people to show up every month. And so if they get a good speaker, or if there's some coincidence with something that's happening in town, they'll get spikes... But then attendance is up and down. It's easier in a place like New York or Chicago or wherever, but for the smaller towns, it's a little bit more difficult. And it's so easy to not put on your shoes, it's so easy to not leave the house. I can go on YouTube and find a thousand talks to watch, or what have you.
+
+**Jerod Santo:** \[16:08\] Right.
+
+**Jim Remsik:** And I think the big thing - and Adam, you alluded to this earlier - for me is that in-person connecting with people. It is the ability to sit down... One of my favorite stories from a conference was talking to an attendee from Madison Ruby, and they were so excited to sit down with Tenderlove... And they talked about America's next top model for like an hour. And it's like, you would never think, looking at Aaron's public persona, that that's the thing that you're going to wind up in a conversation about.
+
+So I think there's so many surprises that happen when you talk to people face to face. Or I really want to talk to Adam, but he's talking to somebody else over there, so I'll talk to Jerod in the meantime. It turns out Jerod and I have things in common that we didn't know. And so there's this serendipity that's possible in in-person events that just -- you cannot replicate it in online events. Or if you can, you have to be extremely intentional about it.
+
+**Jerod Santo:** You can't replicate it, and you also can't guarantee it either. And so there's a risk involved, and like you said, it's so easy just to keep your shoes off... I think that it's hard for folks to see that value when it's variable, and when there's so many barriers to getting there, to sticking your neck out, to shaking a hand that you don't know... To all the things, when it's just easier to stay home and watch videos on YouTube, or watch America's Top Model.
+
+What does a small group look like, and what does a big group look like? What's the difference? We're talking like 10 to 50 people, or like 2 to 20? What numbers are you guys talking about?
+
+**Jim Remsik:** For meetups today, you mean?
+
+**Jerod Santo:** Yeah, you were talking about people struggling maybe with Madison, or...
+
+**Jim Remsik:** Yeah, I think on the smaller end you're talking 5 to 10 people. On the larger end... You know, San Francisco's an outlier in a lot of ways, but I think they regularly get a hundred people or so. I could be wrong about that number. Chicago's routinely 40 to 60 people, depending on the speaker... And as I've been doing research for XO Ruby, Columbus Ruby Brigade, which was always very strong... You know, it appears to have made it out of the pandemic, and they're doing pretty well to this day... But I would say that that's not super-common.
+
+**Jerod Santo:** Are speakers always involved? It seems like you could do meetups just for meetups' sake, like "Let's just hang." And that might be less -- maybe it's more intimidating, maybe it's less intimidating... But does that work or no?
+
+**Jim Remsik:** I'm a consultant, and so the answer that I give most often is "It depends."
+
+**Jerod Santo:** Okay.
+
+**Jim Remsik:** But you're right, it does sound easier. The real answer is "What do you do instead?" There are lots of groups that just do social nights, and that's less intimidating. Not everybody's -- especially men of a certain age, whatever, might not want to go out drinking... So you can do social nights, you're going to get some people in, you're going to get some people out... You can do an activity night, where everybody's downloading software, and running it, or what have you. There's perhaps more work for their organizer in those rounds.
+
+And in general, it seems like, while speakers can sometimes be hard to find, there is usually somebody that wants to promote something that they're doing... And so lots of meetups will engage remote speakers at times, because that way they don't have to pay somebody to travel if they're from out of the area. But the good ones are mixing it up, and having a local person speak, and then have a remote speaker as well.
+
+**Jerod Santo:** \[20:12\] And so now you're deciding to do this regional - is it conferences? They're actually conferences? Or these are meetups? What are these things that you're going to do this fall?
+
+**Jim Remsik:** They're single day, single track conferences. We're aiming for about 60 people. And that's the sort of -- that's the size where you can actually meet everybody in the room. And we're also able to -- we're not going to be in beige boxes, we're not going to be in hotel ballrooms... We're going to be in Chicago, we've got a loft, like a warehouse space, with soaring ceilings... In Atlanta it's a small black box theater, and in new Orleans we're in a -- I should really look up how old this place is, but it's called the Marigny Opera House. It used to be the Holy Trinity Church, and it is just every bit a new Orleans church, which is... It's a white brick building, but it looks really gritty and dirty on the outside... You get inside and it's just fantastic, and I'm super-excited about that venue in particular. In Portland we've got an industrial chic warehouse space. And we're bringing projector PA stage backdrops... Pretty much the entire conference is going to go into the back of the vehicle and drive from city to city. So I've got 7,500 miles of driving ahead of me in the next two months, and with the goal of -- we could go into any space. If you've got power, seats and bathrooms, we'll be able to put on an event. That was the thought.
+
+**Adam Stacoviak:** What's the what's the venue for Austin? That's your last leg, is Austin. That's my stomping grounds, you know that, right? I'm in Austin now.
+
+**Jim Remsik:** I do know that. And my theater -- somebody came in and sniped my venue out from underneath me.
+
+**Adam Stacoviak:** You're homeless...?
+
+**Jim Remsik:** I mean, I am --
+
+**Adam Stacoviak:** Well, you know what I mean.
+
+**Jim Remsik:** ...I'm venueless in Austin at the moment.
+
+**Adam Stacoviak:** Venueless, yeah. Better word.
+
+**Jim Remsik:** I'm experiencing --
+
+**Adam Stacoviak:** Maybe we can help you.
+
+**Jim Remsik:** I'm experiencing venuelessness... So I've been working with the folks at Alamo Drafthouse, so there's possibility there... I don't want to shade anybody, but there's a well-known incubator space there that I haven't had any response from; I reached out there... But yeah, I'm unfazed. Our demands on space is not that large that --
+
+**Adam Stacoviak:** And guess who else is going to be there, Jerod? Landon Gray. Does that name ring a bell to you?
+
+**Jerod Santo:** No.
+
+**Adam Stacoviak:** Think back to "An aberrant generation of programmers."
+
+**Jerod Santo:** Oh, Landon.
+
+**Adam Stacoviak:** Yeah, Landon. He was the young buck... I don't know what was the --
+
+**Jerod Santo:** Our Gen Z representative.
+
+**Adam Stacoviak:** Yeah.
+
+**Jerod Santo:** Yes, with us...
+
+**Adam Stacoviak:** Landon is speaking there, according to the site.
+
+**Jerod Santo:** That's cool.
+
+**Jim Remsik:** Sure is. And you know where I'm --
+
+**Adam Stacoviak:** How do you know Landon?
+
+**Jim Remsik:** Did you ask how I know him?
+
+**Adam Stacoviak:** Well, I assume he was in Austin, but yeah - how do you know him?
+
+**Jim Remsik:** I know him from a conference. I went to Blue Ridge Ruby two years ago in Asheville, and met Landon. He was working at Test Double at the time. And then - I forget if it was... It must've been last year. I traveled, and we did a traveling meetup in New York city, and he lives in upstate New York. And so he came down and spoke at our event in New York city. Yeah, he and I have been bouncing off with one another ever since.
+
+**Adam Stacoviak:** Is he an Austinite, or...?
+
+**Jim Remsik:** No, he's upstate New York.
+
+**Adam Stacoviak:** Okay, so he's coming down then for that. Cool.
+
+**Jim Remsik:** \[24:00\] And the point of doing these conferences is to put them in people's backyards, so that there's not the expensive -- we've been working with Ruby Central to do the designs, both digital and in physical space for RailsConf and RubyConf, and the feedback that we constantly get is "This is great, but I've got to take an entire week off work. I've got to spend a week in a hotel in a major metro, and all those things are expensive." And what is also true is that you can't really get that experience without doing it that way, where you bring all the people to -- like, you bring a bunch of the community together, you get 800 people together in the same space. That's hard to do in any other alignment.
+
+But there was a gap that we saw, which was "What if you didn't have to take a plane, what if you didn't have to drive eight hours, or whatever? What if there was no hotel involved? And what if the ticket, instead of being 300 per day, was a hundred dollars?" And that was what we were trying to solve for. So we spoke with the folks at Cisco, and they saw what we were trying to do, and so we teamed up with them. They're our foundational sponsor. They're sponsoring the entire series of events we're doing this year. And honestly, I couldn't have done that without them.
+
+**Adam Stacoviak:** Yeah, for sure. When I first heard about this, if I'm being honest with you, I was like "Gosh, man... Events are so hard. Smaller events have to be even harder. How do the economics work, and how do you get motivated by it?" You don't have to share the numbers behind the scenes necessarily, but what is it that makes you want to do this? Like, what makes this hard -- it seems like a hard journey to me. What makes this hard journey worth it to you?
+
+**Jim Remsik:** Events have always been a -- I have always had events, going back all the way to 2008. I submitted a talk to Scotland on Rails, and I pitched a talk on geospatial searching in Rails, which was using Thinking Sphinx at the time... And got accepted. I'd never done a talk before... And it bombed. It bombed so hard. Of a 30-minute talk, the first 15 minutes was mostly me on stage, sweating, because we were trying to figure out how to get my computer to work with a WiFi signal... A classic thing that would only happen in an event. You live and learn, and I've never gotten caught by that again. But what also happened was I was waiting in the registration line, and I'm watching Ellen Francis, Grant Paul just completely swamped by this rush of people who were trying to get through the registration process. And so I jumped behind the table, and I'm like "I can handle T-shirts. Tell me your size, I'll get you the shirt size that you need after you get registered, and whatnot." And so that was -- I've shared the story a number of times, and also with them, but I was like "Man, if these guys can do it, I could do this as well."
+
+We've done Madison Ruby for about 10 instances, including last year, and at the last one I was just like "This is really risky. Our expenses are significant, and what would make them easier to pull off would be if we can lower the risk profile." And so through the sponsorship that we got from Cisco, we were able to cover all of our venues, and cover some of our expenses... And then working with our local sponsors, if we are able to reach our sponsorship goals, we have everything covered, and then any of the ticket sales is on top of that.
+
+\[28:17\] We haven't hit all of our sponsorship goals yet, but I'm positive that we'll get pretty close. And if people don't have to plan a trip, they can buy a ticket the night before, and still make the show. And so we're going to cities where there are significant Ruby user groups, there is a significant population... And we've tried to lower the risk at every stage, so that we're not left holding the bag.
+
+When we ran our first Madison Ruby, I personally lost $9,000. I'm going to adjust that to say "My wife and I, who's sitting in the other room right now, lost $9,000."
+
+**Jerod Santo:** Does she appreciate that?
+
+**Jim Remsik:** Yeah. We eventually figured that out and got it back... That was also the only time I ever wound up on the front page of Hacker News... Because we cut the ticket price from 495 down to 250, and said "We're refunding everybody who bought at that 495 price. We're refunding you down to 250", and got a bunch of press out of it. And we still lost money, but I think it would have been even worse having not done that.
+
+**Adam Stacoviak:** Was it you reduced the price because you wanted to get more -- or because it was selling less, to make it more attractive? And then you felt bad charging some people higher, and then some people lower, kind of thing?
+
+**Jim Remsik:** That's right. It felt like the right thing to do.
+
+**Adam Stacoviak:** Yeah. Because usually it's early bird, you buy early, you get the better price, not you get the worst price.
+
+**Jerod Santo:** Right.
+
+**Jim Remsik:** Yeah. You don't want to take the people who supported you early and say "Sorry, this shaft is for you."
+
+**Adam Stacoviak:** Yes. Yes. That's not what Budweiser's tagline is...
+
+**Jim Remsik:** It's close, but it's not --
+
+**Jerod Santo:** "The shaft's for you"?
+
+**Adam Stacoviak:** Yeah, the shaft is not for you, okay? Don't, take it. Well, maybe in an attempt to shed some light on your sponsorship opportunities for XO Ruby, where in the world can people go and find out how to sponsor this thing and learn about it? I know xoruby.com, but is there a sponsor page? I couldn't find one.
+
+**Jim Remsik:** That is an excellent point, Adam...
+
+**Adam Stacoviak:** I'm so sorry...
+
+**Jerod Santo:** \[laughs\]
+
+**Jim Remsik:** No, no, no, by the time this airs, there will be -- if you go to the footer, there's a promo kit. If you go to that link, there are some sponsorship information --
+
+**Adam Stacoviak:** Some stuff in there. Okay, I can see that.
+
+**Jim Remsik:** And also, if anybody's just interested, you can go to /sponsor and it will take you there. You know what? I I'm feeling better about myself. If you -- in the top nav there is a link to the Firestarter, which is a campaign to support this. It includes sponsorship at the event and city level, but also if you wanted to --
+
+**Adam Stacoviak:** Okay. Oh, I see this now. Yes, yes...
+
+**Jim Remsik:** ...if you want to buy a ticket, or buy us a tank of gas, or tacos, or however you want to support... Like, if we're not going to be coming where you are and you still want to support, just go to the Firestarter page on xoruby.com and all the details are there.
+
+**Adam Stacoviak:** That actually is a proper page to send people to. If you scroll a couple of times, listeners, if you're going there, scroll a little bit past some of the things that are important, but... If you want to go right to the sponsor details, this is actually really cool. You've got like one leg sponsor, where somebody can sponsor a leg... Or if you want, like you said, tacos, or whatever... Or one of each. I thought that was kind of cool, where you spend 600 bucks and you get one ticket for each event. Maybe you go on the road with you, or whatever. I don't know. This is kind of cool, I like this.
+
+**Jim Remsik:** Yeah, I think the truth of our work days and the events that we put on is that we can kind of make them look however we want. Like, there's going to be lines, but you don't always have to stay within the lines.
+
+**Adam Stacoviak:** Yeah.
+
+**Jim Remsik:** \[32:11\] This is quite literally me putting -- somebody suggested I was playing Johnny Appleseed, going around planting a bunch of trees... And if we go to Atlanta, and we're working with the local organizers in Atlanta, and if they say "Hey, this turned out really well." "Would you mind if we do this next year?", they're like "No, here's a playbook. You can do this next year."
+
+So if somebody takes that playbook and runs with their own conference in Atlanta next year, we'll go somewhere else. We've already spoken with folks in Boston, Denver and Kansas City, that are asking "Would you bring an XO Ruby conference here?" And the answer is let's get the first three out of the way first, and then let's talk. But there's nothing stopping us from doing that, and we have a pretty good understanding of how to put these on.
+
+**Adam Stacoviak:** Yeah. Does that mean that -- I'm on the Madison Ruby website now... Sorry, if I'm pulling up some more things here, but... I'm noticing that it doesn't say 2025 on this webpage I'm on now, MadisonRuby.com.
+
+**Jim Remsik:** Adam, I'm already putting on six conferences this year. Do you want me to put one out in Madison to you?
+
+**Adam Stacoviak:** I'm not suggesting that you are. What I'm asking is does that mean this is paused, so that you can do this?
+
+**Jim Remsik:** Absolutely.
+
+**Adam Stacoviak:** Okay.
+
+**Jim Remsik:** This is a heavy lift, and the first year of a conference is always the hardest. Anybody who's been to a Madison Ruby will recognize bits and pieces of what we've done in Madison in each of these, including some of the speakers, including some of the activities and whatnot. So I want to do something in Madison again, but honestly, it will probably be XO Ruby, rather than Madison Plus.
+
+**Adam Stacoviak:** Does the website get a lot of traffic to Madison Ruby? Or is it just kind of like fickle whenever there's no event really happening?
+
+**Jim Remsik:** It seems like it's fairly fickle, and on Monday, for some reason, we got 400 hits, and I have no idea why.
+
+**Adam Stacoviak:** Yeah. Well, those 400 people did not know about XO Ruby. That was gonna be my next thing, was like "Okay, let's tell--" I mean, because you're famous for doing this for ten seasons, as you had said. Was it nine or ten?
+
+**Jim Remsik:** Ten.
+
+**Adam Stacoviak:** Ten. I mean, ten's a lot, right? That's a decade of producing a conference... So there's gotta be some sort of brand equity there, where people are coming back to it, or there's content that people are coming to... Leverage that. Promote XO Ruby on Madison Ruby. It's all your company, Flagrant, so I would happily like say "We're paused, but that doesn't mean we're not here. We're elsewhere in the world", kind of thing.
+
+**Jim Remsik:** "We'll be back..."
+
+**Adam Stacoviak:** Yeah, "We'll be back." Thumbs up.
+
+**Break**: \[34:55\]
+
+**Jerod Santo:** What are some of the challenges that have come up with organizing six in a row, I guess, or... You have to organize them in parallel, but they operate in a row. I assume call for papers is probably harder, and lots of stuff's probably harder when you have six things going on.
+
+**Jim Remsik:** So yes, that is true, which is why we didn't do a traditional CFP for this.
+
+**Jerod Santo:** Okay.
+
+**Jim Remsik:** I largely curated the talks for the first couple of events. If you go to the schedule, if there's a slot where there's not a speaker, there is a link there to sign up and say that you'd like to join... Because again, the hope is that either we or somebody else will do this again next year. And so we really want to find out where people are... So that's part of it.
+
+The failures cascade... Thankfully, in the first round, we actually lost our venue in Chicago once, and had to go to a backup. And while I'm extremely happy with the backup - in fact, that's what happened with Madison Plus Ruby the first year, and we wound up with this amazing venue that we really loved... But that meant that we had to go back through and reschedule both of our events in Atlanta and New Orleans after that venue fell through.
+
+So that is a risk in doing these sort of in series. Typically, you get into a workflow of -- you know, there's a \[unintelligible 00:39:00.19\] conference, and then you're done and you get your rest. And it's just been building for the last couple of months, and in two weeks we'll get some relief.
+
+**Jerod Santo:** So you picked this every Saturday cadence for a group of three, and then it seems like there's some short break, and then another group of three. Thoughts behind that? Like, what -- versus, you could have done it like every other week, and maybe just spread them out and relax a little bit in between... But here you have Saturday, Saturday, Saturday, and then a little break.
+
+**Jim Remsik:** I should have done Sundays... Sunday, Sunday, Sunday, you know?
+
+**Jerod Santo:** \[laughs\] Hey, you could have sold more tickets that way. Sounds like a deal.
+
+**Jim Remsik:** You buy the entire seat, but you'll only use the edge. Sorry...
+
+**Adam Stacoviak:** \[laughs\] That was a good one.
+
+**Jim Remsik:** I grew up going to racetracks. My dad drove stock cars... But there is a gap, you're absolutely right. In fact, I drive from Chicago to Atlanta to New Orleans... And then Austin's right here, but instead of going to Austin, I drive back to Wisconsin. My wife's birthday is in that gap, and so I wanted to come home, spend some time at home. And then something that's not even on the XO Ruby map - I'm actually going to Rocky Mountain Ruby before I start the Western leg. So I'm going to drive out to Denver. Flagrant is sponsoring Rocky Mountain Ruby. And then we'll start the rest of the second leg after that.
+
+**Jerod Santo:** Gotcha. You're going to take a -- do you make a documentary or something, like a road trip... You know, road rules, but RubyConf style, Ruby conferences, XO Ruby road rules.
+
+**Jim Remsik:** I do have a little mobile camera rig, like a cage to put around my mobile camera... And I've been talking with some filmmakers about getting -- like, even just behind the scenes stuff. If I'm driving past the largest ball of twine in the country, we'll stop and take a picture. We're also -- DNSimple is sponsoring, and one of their mascots is a... Like, in the write-up that they have about their mascot it says that he's always up for a road trip... And so I'm going to try and get them to send me a plushie to ride shotgun with me. But yeah, we're going to try and do something.
+
+We're also working with Confreaks, trying to get them to record all these events... And if nothing else, maybe we'll do something a little bit DIY to get them recorded if we need to.
+
+**Jerod Santo:** Man, I haven't heard the name Confreaks in a long time.
+
+**Adam Stacoviak:** It has been a long time. So long, I was going to ask you to remind me of the person's name. And as soon as you say it's going to be -- who runs that...?
+
+**Jim Remsik:** You are probably thinking of Coby.
+
+**Adam Stacoviak:** Yes...
+
+**Jim Remsik:** \[41:56\] Coby Randquist. He's at Heroku. He's got bigger fish to fry. He worked with Cindy Backman, who has been running Confreaks for the last six, seven years, and always brings a really... Like, she sees so much. She goes to so many conferences and so many Ruby events and whatnot... She's got an interesting perspective on the history of the Ruby community. And actually has found herself on stage at RubyConf, and... Yeah, and then not only does she record the talks, but she goes through and edits all the talks as well.
+
+**Adam Stacoviak:** It's a lot of work. So much work.
+
+**Jim Remsik:** Yeah.
+
+**Adam Stacoviak:** Wow... Actually, the last time I think we saw -- is it Coby? Is that right?
+
+**Jim Remsik:** Mm-hm.
+
+**Adam Stacoviak:** Last time we saw Coby was at Keep Ruby Weird, first edition here in Austin, Jerod. Remember that? 2014.
+
+**Jerod Santo:** I remember the event, I don't remember that being the last time --
+
+**Adam Stacoviak:** Coby was in the rafters with us, recording.
+
+**Jerod Santo:** I remember that part, but didn't we see Coby at like OSCON as well later?
+
+**Adam Stacoviak:** Maybe...
+
+**Jerod Santo:** Maybe. Who knows?
+
+**Adam Stacoviak:** It may be.
+
+**Jerod Santo:** It's fuzzy back there in the history books...
+
+**Adam Stacoviak:** It's fuzzy back there... \[laughs\]
+
+**Jerod Santo:** But I do remember Keep Ruby Weird at Austin, because that was the first time that we met, wasn't it, Adam?
+
+**Adam Stacoviak:** I think so, man. Yeah.
+
+**Jerod Santo:** And that was at the Alamo Draft House, Jim. So...
+
+**Jim Remsik:** Right on.
+
+**Jerod Santo:** That was a good venue. Good venue.
+
+**Jim Remsik:** Yeah. I'm talking with -- I wanted Caleb to come speak at the Austin event. He's up in Denver. He's not going to make it down... He has a conflict. But I've also got a call in to Terrence as well, so...
+
+**Adam Stacoviak:** Oh, yeah. Terrence Lee could help out, for sure.
+
+**Jim Remsik:** Yeah.
+
+**Adam Stacoviak:** If you mean Terrence Lee. I assume you mean Terrence Lee.
+
+**Jim Remsik:** Blue hat, of course.
+
+**Adam Stacoviak:** Yes, the blue hat. Famous for the Blue Hat. Was it Tar Heels? North Carolina Tar Heels? Isn't that what it is? Or is it just a blue hat?
+
+**Jim Remsik:** I think it's just a blue hat.
+
+**Adam Stacoviak:** Okay. I wasn't sure. I thought it was the Tar Heels, but... That's cool.
+
+**Jim Remsik:** It could be.
+
+**Adam Stacoviak:** I actually spoke to him earlier this year. We had a sponsorship with Heroku and I got a chance to catch up with him again... And then he gave me this number, and I've been too busy to reach out to him since then... So Terrence, you're on my list, man. I'm going to reach out. Because he's like "Hey, you're in Austin, dude. We need to hang out."
+
+**Jim Remsik:** You do. And perhaps you could do that October 25th.
+
+**Adam Stacoviak:** You know, I'm going to come.
+
+**Jerod Santo:** You're going to make it?
+
+**Adam Stacoviak:** I don't see any reason why I can't make it. Let's see, what day of the week is it? We already did -- Saturday, right?
+
+**Jerod Santo:** Yeah, man. Saturday.
+
+**Adam Stacoviak:** We established the joke already?
+
+**Jim Remsik:** Yes, it's not Sunday.
+
+**Jerod Santo:** It's well established.
+
+**Adam Stacoviak:** There's nothing --
+
+**Jerod Santo:** Saturday. Saturday, Saturday.
+
+**Adam Stacoviak:** Yeah. There's nothing on my calendar for that day.
+
+**Jim Remsik:** That's not true...
+
+**Jerod Santo:** So you have zero excuses.
+
+**Adam Stacoviak:** Let me enable my wife's calendar now.
+
+**Jerod Santo:** Oh, he's trying to find one...
+
+**Jim Remsik:** He's trying to find any excuse not to --
+
+**Jerod Santo:** He is. He's trying to back out, live on the air here.
+
+**Adam Stacoviak:** We do have a YMCA parents night out on the calendar that day...
+
+**Jerod Santo:** Well, you could still go to that later.
+
+**Jim Remsik:** Yeah. This is nine to five. That's later.
+
+**Jerod Santo:** Yeah. That's your evening.
+
+**Adam Stacoviak:** And I think that means that we would have no kids, which is great, because the kids go to YMCA and have fun, and then we can go somewhere else.
+
+**Jerod Santo:** Oh, there you go.
+
+**Adam Stacoviak:** So maybe she's trying to take me on a date.
+
+**Jerod Santo:** Go to the after party. Is there an after party or anything?
+
+**Jim Remsik:** The thing about after parties is that they cost money. \[laughter\]
+
+**Jerod Santo:** Okay. Nine to five. I see what kind of Conf it is.
+
+**Jim Remsik:** Yeah. Add a hundred-dollar ticket... Like, what has happened in the past is -- once you get people together, plans form, and people wind up in places together. But we're not planning an after party, but we do suspect that people are going to gather afterwards.
+
+**Adam Stacoviak:** Yeah, I've got some ideas for you. We should talk. At least here in Austin. That's about it. Some venues, and some thoughts... Just whatever. If I can be helpful, I'll be helpful. And if I can make it, I will for sure make it, for sure.
+
+**Jim Remsik:** This is the thing -- and to your point about putting the opportunity to show interest in sponsoring... When you give people an explicit direction as to how they could help, they can see that.
+
+\[46:04\] So if you want to speak, reach out. Let's talk. If you want to attend -- there is not yet, but by the time this airs, there will be a volunteer link as well. We're going to overstaff with volunteers. And so if you wanted to come to the event and you want to give up an hour or two at the beginning or end of the conference to volunteer, we'll let you in for free. We've always done that with Madison Ruby, and there are people that came into the community that way, that had that opportunity, that now work at Square, and Coinbase.
+
+**Adam Stacoviak:** That's cool. Yeah.
+
+**Jim Remsik:** And so what people most often need is an in. They need somebody to take a chance.
+
+**Break**: \[46:51\]
+
+**Adam Stacoviak:** You made me think about something, since you've been running these events so long and you have that thing that you do, so to speak, the way you run it... Is there anything in particular that you do, that you feel like makes an event - whether it's Madison Plus Ruby, or if it's XO Ruby, or whatever you're doing, what are some things that you try to always do as like a DNA thing for your events?
+
+**Jim Remsik:** So that's an excellent question. We always try and get people into conversations with other people at the conference. That sort of naturally happens at a hallway track. If you're going to your first time conference -- this happened to me back in... You know, I got sent to a .NET conference out in Vegas, and I methodically went to talks that I thought could help my company, my career, whatever. And what I understand now is that it's the people; that's the benefit, that's the value. You can get connected to people that are going to help you at work, you can get connected to people who you could potentially hire for your job, you can get connected to people that are going to wind up driving in a car down to Future of Web Apps, and fanboying over somebody else in the car... Like --
+
+**Adam Stacoviak:** I'm telling you... It was Obie . I couldn't help it. He was the coolest. He still is the coolest.
+
+**Jim Remsik:** You mentioned Stephen Bristol in that call, or in that conversation earlier...
+
+**Adam Stacoviak:** Oh my gosh, yeah.
+
+**Jim Remsik:** ...and that was a lesson that I learned from him, is we would walk around at a conference and he would always find a wallflower or somebody who wasn't doing anything, and he would just bring them into the fold.
+
+**Adam Stacoviak:** Yes... And he wouldn't let you go. He grabbed you -- not physically. I mean, sometimes physically. But he's like a tractor beam. "Get over here... Get into this conversation. Don't you be isolated over there." He would force you out of your shell. He really would. I know he did it to me.
+
+**Jim Remsik:** \[50:04\] In the best way. I was going to say, then you realize that "Oh, that's why I'm here." Like, he did that to me.
+
+**Adam Stacoviak:** Yeah.
+
+**Jim Remsik:** And so that is another thing that I -- just try to always make sure that if there are people who seem like they're lost, or seem like they don't know what to do, make an introduction. I can't -- especially as a host, I can't sit there and talk with everybody all day long, but I can connect them with somebody who's going to make sure that they're seen after, and introduced to people, and have a proper conference experience.
+
+Another thing that I think is inherent in the way that we've done things is I will always look for people who have never spoken before. I remember Nell Shamrell gave her first talk at Madison Ruby, I think in 2012. And there was something about Nell that caused me to want to do that. But she had never spoken before, so I didn't have anything necessarily to base that on, other than she was a great person... And it turns out she had a drama background, and she just absolutely delivered her talk, and probably the best talk of that conference.
+
+And so even at RailsConf this year, the last RailsConf, I had somebody come up to our booth and I was like "Do you want to speak?" And she's like "You know, you're not the first person to ask me that." So that was Tia. She's going to wind up speaking in Austin as well.
+
+So always just trying to -- try to find people that think that there is an explicit rule set that they have to play by, and have to either wait their turn, or wait for permission, or what have you, and say "Hey, can I take you to the front of the line? You don't actually have to wait for anybody. You can just do it."
+
+One of those things with getting everybody together is we are not catering the conference. We are strategically getting venues that are surrounded by good restaurants, and we're giving everybody $25 to go out and grab lunch. So it's a little bit of an adventure...
+
+**Adam Stacoviak:** So you spend $100 and you get $25 back to go eat lunch?
+
+**Jim Remsik:** That is a heck of a deal, if you ask me...
+
+**Adam Stacoviak:** That is an awesome deal. I love that deal.
+
+**Jim Remsik:** And that is something that we did for Madison. We had the luxury of being downtown and having really fantastic restaurants around. But the other thing is, it doesn't matter how good you are, or how good your catering staff is; it's hard to serve 100 great lunches to people, all at the same time.
+
+**Adam Stacoviak:** It is.
+
+**Jim Remsik:** Add on top of that food sensitivities, and concerns of that nature, whether your preference is vegetarian, vegan, whatever... It's hard for me to serve all those people in that way.
+
+**Adam Stacoviak:** Yeah, one thing that Alan and Stephen did - I keep seeing their names; I miss those guys so much. Obviously, I miss Stephen, but... When they did LessConf, one thing they did... And this is probably why Alan is mayor, because he went out... And if you recall this, Jim, you can share your side of the story, but... They went out to particular local venues and said "Hey, we're going to have people here for this reason", and they made a list of where you can go. I don't recall if it was free or what it was... I think it might have been free, as part of the conference. But it was like, they had already set it up. You go out into that small town, into that little city, independent all of a sudden, and you go where you want to, closest to the venue. You can walk there, you can go with the friend groups that you've made, you can make new friends, you can invite yourself somewhere, be invited somewhere... But they had already identified various places you can go, and it was just really easy.
+
+\[54:09\] And those folks were ready for it, they weren't bombarded... But you had selection, you got to see the city, and you got to come back to the venue and continue your conference experience like nothing happened. It seemed smooth to me. Maybe as an event planner not very smooth, but as an attendee, I thought it was super-cool, very smooth... And I got the choice of where I wanted to go, and I thought that was the best experience ever.
+
+**Jim Remsik:** Yeah. In downtown Madison there is a local business community called Downtown Madison Inc, and they have gift certificates. And so we did gift certificates one year. Just give them a gift certificate and you're covered. But there were so many people that went someplace where it wasn't accepted, or... That value wound up getting lost. Or they would give it back -- they would just give us the gift certificate back. That's good for me, but you didn't get the thing that you paid for as part of your deal.
+
+**Adam Stacoviak:** Yeah.
+
+**Jim Remsik:** But that experience of going out, seeing the city, finding some local hole in the wall and... I can feed you high quality salads all day long, or maybe you'd rather have a steak, or whatever... Some people -- I would just kill a hot dog. I'm not fancy. And so it's gotten to the point where people expect variety, and they have preferences that they want to hold up, and so we'd rather leave people to their preferences.
+
+**Jerod Santo:** Kind of like that awesome burrito place that we found in -- what was that, Adam? Vancouver?
+
+**Adam Stacoviak:** Oh, my God...
+
+**Jerod Santo:** Yeah, it led to a burrito place that Adam just... Loved... \[laughs\]
+
+**Adam Stacoviak:** No... I remember it like it was literally yesterday, and it was not yesterday, obviously...
+
+**Jerod Santo:** Oh, hilarious...
+
+**Adam Stacoviak:** Oh, it was bad. It was bad.
+
+**Jerod Santo:** Just some random person that we were hanging out with., and --
+
+**Adam Stacoviak:** He was being very kind.
+
+**Jerod Santo:** Yeah. "Let's go to lunch." And he lives there, and he's like "Alright, I'll take you guys to this burrito place. It's amazing." Or something, I don't know. And it wasn't that bad, but it didn't jive with Adam at all. And so we were trying to be gracious, you know... But Adam's not as good at acting like this tastes good. \[laughs\]
+
+**Adam Stacoviak:** I left hungry, let's just say. I left hungry.
+
+**Jerod Santo:** Yeah. And with a smirk -- or not a smirk. What's the opposite of a smirk on his face?
+
+**Adam Stacoviak:** A total frown.
+
+**Jerod Santo:** A total frown. Yeah, basically. You were frowning.
+
+**Adam Stacoviak:** I don't want to tangent this too much further, but you knocked a hot dog, and I had some friends over a few months back...
+
+**Jim Remsik:** Who knocked a hot dog?!
+
+**Jerod Santo:** Yeah, who knocked a hot dog?
+
+**Adam Stacoviak:** Well, you said "I can kill a hot dog", like it was not --
+
+**Jerod Santo:** Yeah, kill one.
+
+**Adam Stacoviak:** Okay, okay.
+
+**Jerod Santo:** No, like, knock it down, right?
+
+**Adam Stacoviak:** Knock it down. Gotcha, gotcha.
+
+**Jerod Santo:** Like just shove it down your throat.
+
+**Jim Remsik:** I had hot dogs for lunch yesterday, man...
+
+**Jerod Santo:** \[laughs\]
+
+**Adam Stacoviak:** Well, I discovered these Wagyu hot dogs... And these Wagyu hot dogs - I'd never had Wagyu hot dogs like this ever in my life. We had some friends over for the 4th of July this year, and they brought them. It's like a tradition for them. Like, "Sure. Sweet. I'll cook them." And so they brought their Wagyu hot dogs, and we got the long buns and all that stuff, and I got a Weber kettle, and I got them over a real fire, so they got all the flavor in it... And I promise you - I'll make it for both of you one day - some of the best hot dogs I've ever had in my whole life. In fact, those -- when I think about hot dogs that I want to have in my life, if I could eat a hot dog like that every day of my life... Or not -- every time I eat a hot dog, if it was like that, that's my example of a hot dog. It was so good. So good. Are you hungry yet, Jim?
+
+**Jim Remsik:** I mean, thankfully -- well, you all are... This is afternoon for you two. Which would be about close to lunch. Yeah, I just bought, probably three months ago, a Weber Smokey Mountain. And so I've been smoking up a storm in the backyard.
+
+**Adam Stacoviak:** \[58:01\] Well, we can we can talk about that in the after show, for sure. I will talk about barbecuing with you all day long... Because it is one of my passions. So xoruby.com... This is like pretty much tomorrow you kick off. Are your bags packed?
+
+**Jerod Santo:** Yeah, it's soon.
+
+**Adam Stacoviak:** Have you got piles of laundry that you're washing? What is your stress level in this moment?
+
+**Jim Remsik:** Stress level is pretty low. It's gonna unfold, and it will be the thing that it is...
+
+**Jerod Santo:** That's one way to look at it.
+
+**Jim Remsik:** You could always sell more tickets, you could always have more sponsors... And it's not like we're not still working on it, but... I've got backdrops, and a projector... I did a test run at my mom's 55+ community on Sunday night, just to bring everything out, put it all together, make sure that I've got all the connections that I need...
+
+**Jerod Santo:** What was your talk about?
+
+**Jim Remsik:** No, no, no, I didn't actually do a talk. I just got everything put together. But they have a community space that is large enough to put it all together. Because it turns out, it's like 30 feet of backdrops.
+
+**Adam Stacoviak:** Oh, my gosh.
+
+**Jerod Santo:** That's a big one.
+
+**Adam Stacoviak:** Yeah, don't dry run that day one, right? Test-run that at a 55+ community where it's safe.
+
+**Jim Remsik:** This was not my first test run either, because I did -- I think two weeks before I did one, and it did not go well. I'm like "I'm so glad that I don't have 100 people sitting behind me saying "Where's the show?"
+
+**Jerod Santo:** Yeah...
+
+**Jim Remsik:** You know, the one thing that does have me sweating a little bit is that I ordered the projection screen online, and it was supposed to be here on Saturday, the 23rd. And then they said, "No, it's delayed. It'll be there on August the 26th." Well, it's now August the 27th and I don't have a screen to project on yet.
+
+**Jerod Santo:** Less than a week? Or do you have more than a week?
+
+**Jim Remsik:** More than one week.
+
+**Jerod Santo:** Yeah, more than one week. You've got a week and a half... Chicago, September 6th.
+
+**Adam Stacoviak:** Well, I mean, if it doesn't arrive, dropship it to wherever you're going to be at. Don't be like "It's only at home I can take deliveries." Take delivery wherever you're at.
+
+**Jerod Santo:** Yeah, but you can't change that that late in the game. It's going to show up even later.
+
+**Adam Stacoviak:** Nah, you can'. the next venue, though. The next event gets the screen.
+
+**Jerod Santo:** You should order a backup and ship it straight to Atlanta right now.
+
+**Jim Remsik:** I have found there is a not nearly as good version, but there is a version that I can get, that should be able to get here in time. But I want the good version. I want the best experience for people.
+
+**Jerod Santo:** Sure. Well, if you are in and or around Chicago, Atlanta, NawLeans, as they call it down there, Portland, San Diego... And especially Austin, because I hear one Adam Stacoviak will be in attendance...
+
+**Adam Stacoviak:** I will be in attendance. I'll be there.
+
+**Jerod Santo:** ...it's not too late to get in on it, right? You're selling tickets right up to the day, I suppose, and...
+
+**Jim Remsik:** Absolutely. The other thing too is if you're in between those cities, I'd love to stop and have coffee, or whatever.
+
+**Jerod Santo:** Oh, cool.
+
+**Jim Remsik:** Part of making the movie or making the documentary is we want to find people along the way. I've got one stop already in Nashville, and then I'm talking to Travis Dockter, who's planning Blastoff Rails for next year in Santa Fe, New Mexico, about stopping there as well.
+
+**Jerod Santo:** Put your route up on the website. Put your route so that people can see where you'll be generally speaking. That'd be cool.
+
+**Jim Remsik:** I think, again, in that promo kit, which is in the footer of the xoruby.com, I think there is a link... But we'll make it more prominent and do a specific call on the blog.
+
+**Jerod Santo:** Yeah, that'd be cool. "Here's where we'll be, and when."
+
+**Adam Stacoviak:** \[01:01:47.12\] I would almost encourage -- I'm trying feverishly to see if you'll correct me, but... If there's already an Instagram. Because I was gonna say, just start a brand new Instagram just for this, or something like that. Anything to kind of like document as you go... Because that thing will catch steam, you know? Or even LinkedIn, I suppose. Our version of Instagram is probably LinkedIn, right?
+
+**Jim Remsik:** Yeah, I was trying to avoid --
+
+**Jerod Santo:** \[laughs\] \[unintelligible 01:02:07.01\]
+
+**Adam Stacoviak:** As I said it, I was like "Oh, that's kind of sad..."
+
+**Jerod Santo:** "Our version of Instagram is LinkedIn..."
+
+**Adam Stacoviak:** Well, I think less of our community is -- they're probably on Instagram, but not for that kind of content, maybe. I don't know. I was just thinking it might be better suited for LinkedIn...
+
+**Jerod Santo:** The problem with our community, with software developers, is that we're scattered to the four winds. That's the problem.
+
+**Adam Stacoviak:** We sure are. There is no home anymore.
+
+**Jim Remsik:** There is no home for anybody anymore. Like, there is not one place where you can go and find everybody... And that is absolutely a lesson that I knew going in, but I'm just feeling so hard. I don't necessarily know how to solve it. We are not creating our own asynchronous communication space. We're partnered with Ruby Central and we're using their Slack... So if you buy a ticket to the conference, you get added in Slack. You could always join the Ruby Central Slack from their site, but... We didn't want to create further -- we don't want to create our own Discord and have yet another walled garden that people can't get access to.
+
+**Adam Stacoviak:** Or even temporary. Like, I've been to so many conferences over the years where you have a temporary place, and all the connection you're trying to make is stuck in that past. That's a great choice to use Ruby Central's Slack, and just double up... Because - like, put them where the community's at, and not where it's temporary. That makes a lot of sense.
+
+**Jim Remsik:** Yeah. We did that with Madison Ruby, and we will be moving everybody over to that Ruby Central community Slack instead... Giving people the opportunity to move over there. We're not going to force people if they don't want to, but...
+
+**Adam Stacoviak:** How in the world do they afford that Slack?
+
+**Jim Remsik:** I think it's free. I think it's the free version.
+
+**Adam Stacoviak:** Oh man, I've got to know who they know. Well, not anymore, because we use Zulip, and it's so much better, by the way...
+
+**Jerod Santo:** I was gonna say, they probably just lose their history like we used to.
+
+**Adam Stacoviak:** Yeah, that's probably the case. Anyways, Zulip. Hey, Zulip is cool. We love Zulip.
+
+**Jerod Santo:** Zulip is cool. Another place you can go... \[laughter\] It's one more app. That's the challenge, is it's just \[unintelligible 01:04:08.23\]
+
+**Adam Stacoviak:** That is true.
+
+**Jim Remsik:** I did not join after we started talking... But that's the problem, I'd rather -- here's the thing... Early Twitter days, everybody was in one spot, and there weren't alternatives, there weren't other people doing the same thing. Now there are too many. You've got BlueSky, Mastodon, Instagram, Threads, LinkedIn, X...
+
+**Adam Stacoviak:** Wow.
+
+**Jim Remsik:** How do you find people? And that's on top of like individual Discords, and Zulips, and Slacks...
+
+**Jerod Santo:** Yeah. Well, I did with Discord what you do with Zulip. Anytime somebody says "Join my Discord", I'm like "Nope." I'm just not going to do this Discord. I just decided to nope out of all Discords. I'm not going to do it.
+
+**Adam Stacoviak:** I mean, I have an account, and I'm a part of like two, but it's such -- I don't even understand how to use it. I feel like I'm old, or something. I'm like "How does this work?"
+
+**Jerod Santo:** Discord just feels like it's more about them than me.
+
+**Jim Remsik:** Discord, I definitely struggle with. I think I have about 72 Discord accounts that don't seem connected to any other thing that I know or do... And every time I join, it's like "Oh, you're a new person." I'm like "No, I'm the same person."
+
+**Adam Stacoviak:** Yeah, their join process is horrible.
+
+**Jerod Santo:** It's like the old days XKCD about the protocols, because right now I'm about to say "What we need is one new place, where we're all going to go, and then we'll be happy."
+
+**Adam Stacoviak:** And that will be the place that no one goes. \[laughter\]
+
+**Jim Remsik:** That's LinkedIn?
+
+**Jerod Santo:** Yeah, that's what Adam just said. LinkedIn is--
+
+**Adam Stacoviak:** Maybe. Maybe. I mean, I like LinkedIn.
+
+**Jerod Santo:** It's Instagram for olds? Is that what you're saying, Adam?
+
+**Adam Stacoviak:** \[01:05:51.19\] No, I think Instagram is for everybody. I just thought that maybe more people would gravitate towards LinkedIn. I know I'm -- I probably check LinkedIn several times a week, whereas Instagram, I'm on there maybe, maybe once. And it's just really just to look at my wife sharing our life to our friends, on Instagram, because I like the way she shares it. I just want to see what she says, and look at the pictures she takes... And I saw them in the moment, and I saw them in the camera, because she's like "Hey, look at this picture", but somehow seeing the story version of it later on from the publisher, my wife, is very cool to me. So I pretty much go to Instagram to read DMs from her, because she sends me things that she finds on there...
+
+**Jerod Santo:** Same. Same.
+
+**Adam Stacoviak:** And watch what she shares on there about our life. That's the only reason I go to Instagram.
+
+**Jerod Santo:** Right. To look at her pictures. Yeah.
+
+**Adam Stacoviak:** Yeah, that's it. That's Instagram for me. So I mean, there you go. But...
+
+**Jim Remsik:** I've definitely taken the tact of trying to not add another protocol, but just drop them. So I've never even wound up on Threads..
+
+**Adam Stacoviak:** Yeah, same. I was like "No, I don't think so, man... Is this a version of Facebook I want to be involved in?" Threads just never made sense to me.
+
+**Jerod Santo:** Right. I've given Zuck enough in my life, you know...
+
+**Adam Stacoviak:** It was an also ran. It was an also ran that just didn't run. I don't know. There's a lot of people there...
+
+**Jim Remsik:** I was going to say...
+
+**Jerod Santo:** There's people there, but they're not real people, I don't think...
+
+**Adam Stacoviak:** A bunch of bots...
+
+**Jerod Santo:** No offense to those people, but they're just like --
+
+**Jim Remsik:** Kara Swisher is going to be upset with you, Jerod.
+
+**Jerod Santo:** Yes, she might.
+
+**Adam Stacoviak:** Oh, gosh...
+
+**Jerod Santo:** But you know, she's cross posting to everywhere. Come on.
+
+**Jim Remsik:** True.
+
+**Jerod Santo:** That's what people with big follower accounts do, is they just -- every new thing, they're on it, because they want to get that follower count up. And then they just cross-post the same crap across all of it.
+
+**Jim Remsik:** I feel seen... \[laughter\] I've been cross-posting stuff for XO lately, and... Yeah, you're not lying.
+
+**Jerod Santo:** No. I mean, what else are you going to do? You want to go where the people are, and now the people are in seven different places... And so you're just like "Well, I guess we'll post to seven different places... But we're not going to be happy about it." So no allegiances anymore.
+
+**Adam Stacoviak:** I know I asked you what was your motivation, but I think maybe a different version of that is... What is success of this for you? Like, when you're done with the roadshow, it's December, maybe you're sitting on your couch and you're getting ready for Christmas and you're reflecting on your year, and you're like "Man... his is what happened with XO Ruby", and you're recounting in your brain what happened... What is success? What would that be for you?
+
+**Jim Remsik:** There's a dirty secret in all of this, which is I have run out of -- we haven't heard from my wife in the other room this entire thing, but this might be the one thing... I have run out of clients under this roof, that I need to go out and meet people. And so part of this is a sales trip, right? Like, this is me, going out on the road, meeting a bunch of people in new locations, and looking for design and development work for Flagrant... So that is one goal.
+
+But the bigger goal would be for people to -- there have been people who've had whole careers over the last five years that haven't had access to local conferences. That don't understand that you can talk to somebody and get access to opportunities that they didn't know existed, because they weren't in the same physical space with one another. So if I'm sitting there and I find out "Hey, I just got a new job from somebody that I met at XO Ruby", like, that's a huge win.
+
+I just talked to somebody right before this call, and they said "I've been on the market for nine months, and I don't know what I'm doing wrong... What do you recommend?" And I sent them a ticket to Atlanta. So they're going to show up in Atlanta, and hopefully they'll walk away with a connection that leads to work, or some opportunity that is beneficial to them.
+
+\[01:09:50.05\] That's really the goal, is to get people on stage that have never spoken before. Get people to understand that there is a unique serendipity that is only available if you show up. I've got a friend and mentor, Martin Atkins. He's a musician. He always says "Awesome stuff happens when you do stuff. Nothing happens when you don't." And if you never leave the house, you're never going to meet new people.
+
+**Adam Stacoviak:** That's good.
+
+**Jim Remsik:** You're never going to find opportunity. It doesn't just show up at your door and say "Hey, do you want a job?" That doesn't happen. There was a time where that happened, in the 2010s... But that is no longer the case. And so you've got to go out, you've got to meet a lot of people, because not all of them are going to... You're not going to connect with all of them. But if you meet enough people, you're going to find enough of your people that something works out. Whatever that is for you. Maybe it's a job, maybe it's an opportunity to learn woodworking from somebody that does woodworking. But that ability to connect with other humans is so important.
+
+**Adam Stacoviak:** Can we unpack that? The "Run out of clients under this roof" comment. Can we unpack that a bit more? Can you go there with us.
+
+**Jim Remsik:** Yeah. I mean, I've been doing consulting since 2011. I've been either an owner, or part owner, going back to 2007. My wife is one of my clients, and she's in the other room. I've reached the limit of what I can sell her on... And so now I've got to go out in the real world and get other clients for us to consult. She runs a software as a service, it's successful, and she only has a certain amount of needs.
+
+So I've got to go out and find people that are looking for the skills that me and my team have. And at this point, we're largely doing Rails and Hotwire. We've got some React and TypeScript folks... But we're about half and half design and development. That is product design, UX, UI, all the way up to branding and illustration. Yeah, I mean, it's a tough market for everybody, and there's not a single consultancy owner that I have talked to that isn't struggling, or laying people off, or what have you.
+
+**Adam Stacoviak:** Is this a ChatGPT thing of three years ago thing? Like, a decline since then? I know we all went through the pandemic and there's lots of crazy change up, and then it was a whole different world, and then it was back down to the new old world again kind of thing, where it was back the way it was... Has this been a trend for a decade, or three years, or what's the trend?
+
+**Jim Remsik:** Yeah, I think the -- I mean, I don't think that we have a unique story. We grew during the pandemic, we were up to almost 20 people, we're down to 10 right now... And we are keeping our nose above water. But part of it is, unless you're going out and meeting new people, it's hard to get people to find you, and it's even harder to get them to trust you. So it's really - all of the new work that we've had this year, the bulk of it were people that we met 10 years ago, that already trust us. The other brand new work that we've gotten has been referrals from people who trust us. And so it is really -- if you spend any amount of time around me, you will likely hear this comment, but it was something that was shared in 2011 when we started my first consultancy... Somebody shared this with me and they said "It's not what you know, it's not who you know, it's who knows you." And I've come to learn that that means who knows that you are the person that's going to show up? Who knows that you are capable of solving their problem? Who knows what your values are, and who knows that they can trust you?
+
+\[01:14:21.16\] That's the part of meeting a lot of people that -- it's not that I'm out there doing a little jig so that everybody knows that I can do the jig. You just -- you've got to show up. And after a while, people just know that "That's Jim. He's going to show up. If you've got a problem with Rails, or if you're doing Ruby in Madison, you should talk to Jim. If you want some really great design work done, you should talk to Kelly... But talk to Jim. He'll get you connected."
+
+**Adam Stacoviak:** That makes sense. And so your world is shrinking because of the -- I would say probably... You run the event, but is the lack of getting out there and meeting people outside of the Madison bubble, is kind of what I hear you saying. It's like, you want to go beyond Madison.
+
+**Jim Remsik:** I think that's fair...
+
+**Adam Stacoviak:** Because you've got pretty good touch points in Madison, right? Madison Ruby has got to be giving you something...
+
+**Jim Remsik:** For sure. It is also only once a year, and so we wanted to do something that... You know, the original plan was to do 12 of these every year, and it would have been a cycle that just kept going. So it was constantly meeting new people. And we decided to settle for six, and then just focus on the fall. But if these turn out well, we do have plans to do more of them. And ideally, if folks get involved and they're successful, and they just want to keep running with it, we would be happy to do that, too. Because I would love to just go to events, rather than throw them.
+
+**Adam Stacoviak:** Yeah, putting on an event is tough business. I've resisted my entire life... And Jerod, do you want to share? Recently, we did a live show. I mean, Jerod did all that, pretty much. I didn't do anything besides show up and smile, and upset a few people, really...
+
+**Jerod Santo:** What do you want me to share about it? There's lots of --
+
+**Adam Stacoviak:** Just the fact that we did it.
+
+**Jerod Santo:** Yeah, we did.
+
+**Adam Stacoviak:** You know, the fact that we haven't done anything like that, but we did something... It was the smallest version of what we could do, basically.
+
+**Jerod Santo:** Yeah, totally. It's not like we did six of them in a row, we just did one. We're mere pikers, Jim.
+
+**Jim Remsik:** How many people?
+
+**Jerod Santo:** Oh, 50.
+
+**Jim Remsik:** That's good. I mean, that's a great audience. What sort of venue did you do it in?
+
+**Jerod Santo:** We did it at the Oriental Theater in Denver. So we picked the city first, and then we asked around about locations, and then found a location that was affordable and seemed like it was pretty cool... And it was both those things, so we kind of got lucky, I guess, in that way.
+
+**Adam Stacoviak:** Yeah.
+
+**Jerod Santo:** They were great to work with, which can always be a variable as well... And we kept it as simple as we could possibly keep it. Even our meetup the night before was completely just like "Here's where we're going to be. Come there if you want to." That kind of thing.
+
+**Adam Stacoviak:** Yeah.
+
+**Jerod Santo:** So it was just low stress for that reason. And then obviously, putting the show on is something that we normally do twice a week anyways, just not on stage. So a little bit extra, but not too much.
+
+**Jim Remsik:** Right on. Well received? Everything went well?
+
+**Jerod Santo:** I think so. There's obviously places where -- as the self-critical people that we are, there's plenty to improve. But overall, I think everybody was pretty happy about it, and the show turned out pretty well. What do you think, Adam?
+
+**Adam Stacoviak:** I think when we did our consensus recently we said the only thing we would have changed was have onstage monitors. And I think that's a win, because if all you need is onstage monitors so you can hear yourself better in the mic as you interview somebody... Because we interviewed Nora Jones; she founded Jelly, it was acquired by PagerDuty... We have a podcast about it via this live show, you can go listen to that, but... We were talking into the microphones, but we didn't have in-ear monitors, and we didn't have onstage monitors... So she had to hear us for real. The audience could hear us loud, because they heard us through the loudspeaker system... But she couldn't hear us very well, and we couldn't hear her as well as you would probably want to. I could still hear her, but it was a struggle in a couple of cases where maybe the venue was louder.
+
+\[01:18:28.01\] Just something may have happened where it was like you couldn't hear us as well, and she'd say, "Huh?" or "What did you say?" And that wouldn't have happened if we had onstage monitors or even in-ear monitors. So I feel like that's a major win, if all we walk away with is "Oh, onstage monitors." Everything else went pretty smooth. There was no real hiccups that you couldn't solve in the moment.
+
+**Jim Remsik:** Yeah, that's a huge victory. So you're going to do 12 next year?
+
+**Adam Stacoviak:** We did talk about doing several. Yeah, we did talk about more.
+
+**Jerod Santo:** Definitely more. Maybe not more than one a year... Maybe more than that. But I think we do want to do it again, I can tell you that much.
+
+**Adam Stacoviak:** We haven't circled back on it. I've kind of walked away bullish, honestly... I don't think we should do more than four. Maybe two or three in a year would be cool. Definitely not one. One's not enough. I feel like a couple.
+
+**Jerod Santo:** I don't know, one is special, because it's only one of the year, so you better get to it. Whereas six in eight weeks, or whatever it is...
+
+**Jim Remsik:** Here's the thing - the plan from the beginning was for me to drive far enough that -- somebody from Chicago is not going to drive to Atlanta, let alone the rest of the cities. And so it's local to Chicago, and there's enough people there to support that they get one a year.
+
+**Jerod Santo:** Right.
+
+**Jim Remsik:** And Atlanta gets one a year. And maybe we decide to -- rather than do San Diego, we do Denver next year. Who knows?
+
+**Adam Stacoviak:** I like that idea. See?
+
+**Jerod Santo:** So now you're talking about going on tour.
+
+**Jim Remsik:** I'm already on tour.
+
+**Jerod Santo:** Well, yeah, but you're talking about us going on tour. Or you're thinking about yourself and we're thinking about ourselves... \[laughter\]
+
+**Adam Stacoviak:** Yeah, that's true. Yeah, yeah. We're talking about us now, Jim...!
+
+**Jerod Santo:** I just thought maybe we'd go to like somewhere else, and just do another one. But you never know. I can get talked into more...
+
+**Adam Stacoviak:** I'm down for more no matter what, in terms of like doing it again. Does it have to be six or ten a year? Absolutely not. I don't think I would want to do six. But I kind of want to do more than one, not because I want to do more, but because I think it's nice to get out a couple -- like, I don't know about you, but in my life I kind of have like times when it makes sense to rejuvenate by traveling somewhere with people I want to be around, and all that kind of stuff. And so I feel like that reconnection is good for everyone... And best case, it's a regroup of our team; even better case, everyone else gets to enjoy it, too. I feel like that's what I liked about it.
+
+My favorite moments - and this is totally going TMI, Jim, and I'm sorry, but my favorite moment was really the thing itself, but bringing us together in the same place. Even Fogo de Chão, and like getting to sit down and eat together... To me, that was cool. Two times a year would be perfect. One - not perfect. Three - maybe too much.
+
+**Jerod Santo:** \[laughs\] You talked me into doing it...
+
+**Adam Stacoviak:** Maybe. Maybe too much.
+
+**Jerod Santo:** Now we're negotiating here.
+
+**Jim Remsik:** Jerod, do you find that travel rejuvenates you?
+
+**Jerod Santo:** I guess it depends on the type of travel.
+
+**Jim Remsik:** Or the company.
+
+**Jerod Santo:** \[01:21:47.24\] Yes. I like traveling, so I'm not against the travel part. I don't mind road trips. For me it's a lot more complication than is necessary. So it's like "Is it worth it?" So I think once a year it's really worth it. I think twice maybe it is. But six in eight weeks - Jim, you're crazy, man... \[laughs\]
+
+**Jim Remsik:** Nah, it's just two trips.
+
+**Jerod Santo:** \[laughs\] Oh, I'm sure you have all kinds of ways you tell yourself... I appreciate the gumption, the ambition and the drive. I think that there's people that do that. They're usually younger than all of us. They go out on tour, and they're gone for six weeks.
+
+**Jim Remsik:** Are you saying that I'm old?
+
+**Jerod Santo:** No, I said they're usually younger than all of us... So we have reached a certain, you know, age...
+
+**Jim Remsik:** I mean, I grew up -- I came from a large family, didn't have a lot of money, but we had... My family lived in Florida, and so probably six times in age five to twelve we would drive from Wisconsin down to Florida, straight through. Dad was always the only one that ever drove, and it just built that nostalgia perhaps for the road trip.
+
+**Jerod Santo:** Sure.
+
+**Jim Remsik:** But again, these are shorter stints, for the most part. We've got stops every ten hours or so... And trying to get through that in a week -- sorry, you're just listening to my thought process as to why this makes sense.
+
+**Jerod Santo:** \[laughs\] Like I said, you can tell yourself that kind of things... I think it's going to be cool, and I think it's going to be a memory... I think that you should absolutely record as much as you can, and try to make some sort of a memory video or something, so that you have a takeaway at the end, whether or not anybody else wants to watch it but you and your wife or your team... Who cares, you'll like it. And I hope lots of people come, and there's great talks, and there's great food, and there's great venues... It sounds like you still have some variables up in the air, which for me would be stress... You seem like you're very chill about it, so I'm --
+
+**Adam Stacoviak:** So chill.
+
+**Jerod Santo:** ...chill about it as well.
+
+**Jim Remsik:** It's going to work out. It's going to be what it's going to be, right?
+
+**Jerod Santo:** That's kind of how I approached this event. And I think Adam probably would have stressed about it more than I did, and this was why I kind of kept him out of the details...
+
+**Adam Stacoviak:** I would have.
+
+**Jerod Santo:** ...and I'm just like "You just get a place, and then you get a plan together, and you invite people, and then you show up and do the thing... And it'll work out." And it pretty much did.
+
+**Jim Remsik:** But again, the support from the community has been great. We've got Coraline Ada Ehmke, Scott Werner in Chicago, Seth Giddens is probably going to join in New Orleans... Tenderlove is going to come down to Portland... I'll leave out a few more names, because I haven't gotten signoff-signoff...
+
+**Jerod Santo:** Sure.
+
+**Jim Remsik:** But again, Yuri is speaking in Chicago. He's never spoken at a conference before, but reached out and he's like "I would really like to. Here's my talk idea", and we're talking through how to best pull it off... And Tia is speaking in Austin... Landon you all know in Austin... Ceci Correa in Austin... So yeah.
+
+**Jerod Santo:** Well, if you put Omaha or Kansas City on your map next time around, I will be there. I'm not going to drive to Chicago, it's eight hours. It seems like a lot. But I would definitely hop over to KC. It's like a two and a half hour drive from here. For a RubyConf.
+
+**Jim Remsik:** Tulsa?
+
+**Jerod Santo:** Tulsa is kind of far. \[laughs\]
+
+**Jim Remsik:** That's fine.
+
+**Jerod Santo:** I'm not saying I wouldn't, but --
+
+**Adam Stacoviak:** "Can I interest you in Tulsa...?"
+
+**Jerod Santo:** ...that gives me pause. Why? Is that already on your list? It has been sort of the canonical example of - I want people to say "Hey, we've got 30 people in Tulsa that want a conference." The original idea for this was to go to underserved technology cities...
+
+**Jerod Santo:** Yeah.
+
+**Jim Remsik:** ...that still have a population. Back in Hashrocket days we went to Huntsville, Alabama for a conference that Jeremy McNally ran...
+
+**Jerod Santo:** I went to that as well.
+
+**Jim Remsik:** We wound up playing --
+
+**Jerod Santo:** Yup. At the old Grand Ole Opry.
+
+**Jim Remsik:** I think we went --
+
+**Jerod Santo:** Or that was Tennessee.
+
+**Jim Remsik:** Yeah, that was Tennessee.
+
+**Jerod Santo:** Okay. Well, it was definitely a Ruby Hoedown, but maybe it was... Did he do it in different places?
+
+**Jim Remsik:** He might have. Because he also did a conference in Disney. He did Magic Ruby.
+
+**Jerod Santo:** \[01:26:18.07\] That's right. Magic Ruby.
+
+**Jim Remsik:** So he bounced aound a little bit.
+
+**Jerod Santo:** Okay. So I went to the Tennessee one, not the Alabama one.
+
+**Adam Stacoviak:** Are you talking about Jim?
+
+**Jerod Santo:** Jeremy McNally.
+
+**Adam Stacoviak:** Jeremy. I was thinking of Jim who did Lone Star Ruby... I can't remember his last name in this moment.
+
+**Jim Remsik:** Jim Freeze.
+
+**Adam Stacoviak:** Yeah, Jim Freeze. That's who I was thinking. Because he did that in... Yeah, in Florida.
+
+**Jim Remsik:** Yeah. One of those fun memories... We actually drove an RV from - I drove the RV - from Jacksonville, Florida. We rented an RV and piled eight of us into it, and drove to Ruby Hoedown in Huntsville. And we had a group of young people that were all sort of eclectically dressed, and we drive up to like a barbecue place... And we were just looking for a bathroom, and pile out of the video. And somebody asks T-Pope, "Are you guys in a band?" He's like "No..." It's like, "T-Pope, if anybody asks you if you're in a band, you say yes."
+
+**Adam Stacoviak:** Yes...
+
+**Jerod Santo:** \[laughs\]
+
+**Adam Stacoviak:** You get free food.
+
+**Jerod Santo:** The answer is always yes.
+
+**Adam Stacoviak:** They might say, "Can you come on stage?" "Oh, dang..."
+
+**Jim Remsik:** You don't want to overplay that one, that's right.
+
+**Adam Stacoviak:** That's right. "We're late. We've got a gig. It's too late, we're on our way." Well, that's cool, man. I want to call one out, because a friend of mine - a friend of ours - is going to New Orleans. And if it was in Austin, I would be much more happier, because that's where I'm going to go. I'm not going to go to New Orleans. But our friend Robert Ross from FireHydrant, who's been a friend of mine, founded FireHydrant, and then became a closer friend by being a sponsor, and being a friend... And I just love their journey... And he's talking about if I did it again. And I think he's talking about the journey of Ruby with FireHydrant, and the mistakes they made, and the things he would do again, obviously, versus the things he wouldn't do again... So I like the way Robert shares his stories, especially how it relates back to building a successful company... And so that's what I'm personally looking forward to. Will these be recorded? I mean, because I won't be able to see it. If Confreaks will answer your call maybe it will be recorded, right?
+
+**Jim Remsik:** No, no. Confreaks has answered the call. It's just, if we can find a way to get them the paltry sum that they're looking for to record.
+
+**Adam Stacoviak:** Yeah, that makes sense, too. Well, I got some ideas for you. We'll talk in post. I've got some ideas for you if that doesn't work out, for sure.
+
+**Jim Remsik:** They have been great. I've worked with Cindy for most of -- I'm trying to think... I've known Cindy for most of the decade, most of the past decade. Again, she is one of those people who shows up, and she always delivers, and she'll always try and make something happen if possible... And we're going to try and make something happen.
+
+**Adam Stacoviak:** That's a good deal. Well, everyone, go to xoruby.com. Fall in love, hugs and kisses, Ruby all around... And then go to the same domain /firestarter. And if you're not going to be there and you love everything Jim has said and you resonate with how he cares about the community, and you care about supporting Flagrant, and their future, and what they're doing, and all the ways Jim shows up - be a super-fan, buy a six ticket, buy a latte, buy some tacos, a tank of gas, whatever you can do.
+
+I would love to see some folks from our community buy a tank of gas. That'd be kind of cool. 50 bucks. Buy a tank of gas, show some love, help Jim out on this trip here... And maybe he'll do it again and again and again. Maybe in your city. Who knows? Jim, thanks so much, man. Thanks for having a great heart, and pouring in, and planting seeds, being the Johnny Appleseed we need in our world... And thanks for coming on the podcast. It's been great.
+
+**Jim Remsik:** Again, my pleasure, and thank you so much for helping me get the word out.