[
  {
    "start": 0.14,
    "end": 2.66,
    "text": "Hello friends, this is the AlphaList podcast."
  },
  {
    "start": 3.04,
    "end": 4.08,
    "text": "I am your host Tobi."
  },
  {
    "start": 4.4,
    "end": 11.5,
    "text": "The goal of the AlphaList podcast is to empower CTOs with the info and insight they need to make the best decisions for their company."
  },
  {
    "start": 11.84,
    "end": 17.76,
    "text": "We do this by hosting top thought leaders and picking their brains for insights into technical leadership and tech trends."
  },
  {
    "start": 18.16,
    "end": 24.06,
    "text": "If you believe in the power of accumulated knowledge to accelerate growth, make sure to subscribe to this podcast."
  },
  {
    "start": 24.4,
    "end": 32.2,
    "text": "Plus, if you're an experienced CTO, you will laugh at the discussion happening in our Slack space where over six hundred CTOs are sharing insights."
  },
  {
    "start": 32.4,
    "end": 33.6,
    "text": "or visit one of our events."
  },
  {
    "start": 33.88,
    "end": 36.2,
    "text": "Just go to alphalist.com to apply."
  },
  {
    "start": 51.48,
    "end": 53.32,
    "text": "Welcome to the Alphalist podcast."
  },
  {
    "start": 53.64,
    "end": 54.56,
    "text": "I'm your host Tobi."
  },
  {
    "start": 55.54,
    "end": 61.64,
    "text": "And today with me is a guy who just sold his company to another very popular company called Cursor."
  },
  {
    "start": 62.24,
    "end": 70.2,
    "text": "It's Greg Foster and Greg is the founding CTO or founder and CTO of a company called Graphite."
  },
  {
    "start": 70.7,
    "end": 75.28,
    "text": "And no, Graphite is not the visualization tools from the zeros, which I loved actually."
  },
  {
    "start": 76.76,
    "end": 81.76,
    "text": "But a tool that came up like, I don't know, when did you found like two and a half years ago or"
  },
  {
    "start": 82.4,
    "end": 83.24,
    "text": "a little bit older than that?"
  },
  {
    "start": 83.32,
    "end": 87.36,
    "text": "We founded, but we pivoted, we founded around twenty twenty."
  },
  {
    "start": 88.16,
    "end": 90.16,
    "text": "And we've been working in the DevTools space the whole time."
  },
  {
    "start": 90.22,
    "end": 91.5,
    "text": "Also, Toby, thank you so much for having me."
  },
  {
    "start": 91.7,
    "end": 92.32,
    "text": "I'm excited to be here."
  },
  {
    "start": 92.94,
    "end": 95.72,
    "text": "We founded around twenty twenty and we're working DevTools the whole time."
  },
  {
    "start": 96.8,
    "end": 104.18,
    "text": "It took us about a year, year and a half to start building what we would now is GraphSite, a code review platform."
  },
  {
    "start": 104.24,
    "end": 109.16,
    "text": "We were inspired by internal tooling from Facebook and Google, fabricator and critique."
  },
  {
    "start": 109.68,
    "end": 116.38,
    "text": "Some of these layers and styles and patterns on top of GitHub for collaborating on code changes and doing better code reviews."
  },
  {
    "start": 117.32,
    "end": 121.46,
    "text": "And then as the AI DevTools wave started picking up, we got accelerated."
  },
  {
    "start": 121.74,
    "end": 130.759,
    "text": "We got boosted by that all in about twenty twenty three and we've been flying since, but we've been working on graphite for some time now."
  },
  {
    "start": 130.919,
    "end": 130.919,
    "text": "I"
  },
  {
    "start": 131.02,
    "end": 135.7,
    "text": "think I saw you like literally exploding beforehand on product hunt."
  },
  {
    "start": 135.82,
    "end": 139.74,
    "text": "There was like one, I think that was StackPRs, which you launched, right?"
  },
  {
    "start": 140.46,
    "end": 141.3,
    "text": "Which took off like crazy."
  },
  {
    "start": 141.34,
    "end": 141.44,
    "text": "Yes,"
  },
  {
    "start": 141.92,
    "end": 142.5,
    "text": "which is still."
  },
  {
    "start": 143.74,
    "end": 144.58,
    "text": "And that's the same thing."
  },
  {
    "start": 144.64,
    "end": 146.46,
    "text": "You know, it's all part of the same platform."
  },
  {
    "start": 146.54,
    "end": 153.98,
    "text": "StackPRS was one of the better ideas internally at Facebook and some of these other companies who use fabricator, like Twitter and so on."
  },
  {
    "start": 154.6,
    "end": 163.88,
    "text": "But for us, these are all features, whether it's StackPRS or AI Code Review, these are features on top of just what should be a great code review and code collaboration platform."
  },
  {
    "start": 164.18,
    "end": 165.86,
    "text": "If you have a PR page, you should have an inbox."
  },
  {
    "start": 165.98,
    "end": 167.54,
    "text": "You should have the ability to stack your PRs."
  },
  {
    "start": 167.8,
    "end": 168.8,
    "text": "You should have a really good merge queue."
  },
  {
    "start": 169.12,
    "end": 171.32,
    "text": "You should have AI assistance helping you leave comments."
  },
  {
    "start": 171.78,
    "end": 174.78,
    "text": "But these are all features part of a platform for shipping code changes."
  },
  {
    "start": 178.86,
    "end": 180.1,
    "text": "github does these days right like."
  },
  {
    "start": 180.22,
    "end": 184.6,
    "text": "i mean that's what the window into github turned out to be like."
  },
  {
    "start": 184.68,
    "end": 186.78,
    "text": "i mean github is like kind of an operating system."
  },
  {
    "start": 186.86,
    "end": 191.46,
    "text": "there's product management or project management there's a lot there's security etc."
  },
  {
    "start": 192.26,
    "end": 195.38,
    "text": "but for me like the ultimate angle is still poor request like."
  },
  {
    "start": 195.42,
    "end": 199.84,
    "text": "what i see most is poor request and you basically reinvented the poor request."
  },
  {
    "start": 199.92,
    "end": 202.48,
    "text": "let's let's phrase it like that right for for large teams."
  },
  {
    "start": 202.56,
    "end": 202.86,
    "text": "is that?"
  },
  {
    "start": 203.06,
    "end": 203.68,
    "text": "is that correct?"
  },
  {
    "start": 204.16,
    "end": 205.16,
    "text": "we want to help people."
  },
  {
    "start": 205.98,
    "end": 208.58,
    "text": "keep those PRs moving as fast as possible."
  },
  {
    "start": 209.04,
    "end": 210.24,
    "text": "We don't own the data for what's worth."
  },
  {
    "start": 210.48,
    "end": 213.08,
    "text": "Graphite, in many ways, is a client on top of GitHub."
  },
  {
    "start": 213.56,
    "end": 219.0,
    "text": "You might have an old Gmail or Yahoo mail account, but you use a more sophisticated email client."
  },
  {
    "start": 219.36,
    "end": 221.12,
    "text": "Maybe you use a sophisticated calendar client."
  },
  {
    "start": 221.38,
    "end": 223.12,
    "text": "In many ways, we're that sophisticated client layer."
  },
  {
    "start": 223.66,
    "end": 226.3,
    "text": "On top of your existing GitHub, we treat them like a database."
  },
  {
    "start": 226.54,
    "end": 227.46,
    "text": "Your PRs are still there."
  },
  {
    "start": 227.5,
    "end": 228.28,
    "text": "Your code is still there."
  },
  {
    "start": 228.74,
    "end": 231.46,
    "text": "But Graphite is an independently adoptable client."
  },
  {
    "start": 232.32,
    "end": 233.86,
    "text": "Big teams love us like you just asked."
  },
  {
    "start": 234.22,
    "end": 235.64,
    "text": "Small teams also really enjoy us too."
  },
  {
    "start": 236.38,
    "end": 237.62,
    "text": "And we give you a lot of these superpowers."
  },
  {
    "start": 237.66,
    "end": 239.04,
    "text": "We give you an inbox for finding those PRs."
  },
  {
    "start": 239.12,
    "end": 244.18,
    "text": "If you're a big company, we give you a merge queue, which is critical the moment you exceed about a hundred people in a repository."
  },
  {
    "start": 244.8,
    "end": 257.2,
    "text": "We give you a lot of modern AI niceties, everything from leaving comments to helping fill out your descriptions to chatting with your PR and asking research questions or asking for mutations right there in the PR page and just fixing that PR up and then merging it in."
  },
  {
    "start": 258.04,
    "end": 259.019,
    "text": "All of these things we do."
  },
  {
    "start": 259.38,
    "end": 261.7,
    "text": "Again, some of these ideas are new around the AI stuff."
  },
  {
    "start": 261.74,
    "end": 263.28,
    "text": "Some of these are really timeless ideas."
  },
  {
    "start": 263.82,
    "end": 273.2,
    "text": "GitHub didn't integrate them but they've been around at some of the other companies internally for a long time and we wanted to bring these superpowers to any engineer without asking you to move off GitHub."
  },
  {
    "start": 273.62,
    "end": 284.4,
    "text": "If you're at Netflix or Shopify or Notion or at a small hundred person or ten person startup and you guys use GitHub, still anyone can just pick up an independently used graphite and you don't have to move the whole company over."
  },
  {
    "start": 284.62,
    "end": 286.48,
    "text": "That was such an important thing for us early on."
  },
  {
    "start": 287.12,
    "end": 291.72,
    "text": "As you know with engineers, it's so difficult to ask people to change their workflows."
  },
  {
    "start": 292.06,
    "end": 293.32,
    "text": "That's not how engineers like to work."
  },
  {
    "start": 293.62,
    "end": 297.3,
    "text": "Engineers like to independently play with something fiddle with it, try it, get off it, get back on it."
  },
  {
    "start": 298.08,
    "end": 300.44,
    "text": "We tried to create graphics from day one with this nature."
  },
  {
    "start": 300.92,
    "end": 308.46,
    "text": "Prior systems, Bitbucket, GitLab, so on, Fabricator, would say, hey, move your whole company's source controller to a new platform."
  },
  {
    "start": 308.58,
    "end": 310.22,
    "text": "That's impossible for these large companies."
  },
  {
    "start": 310.78,
    "end": 313.42,
    "text": "We had to create something that individuals could just pick up as play with."
  },
  {
    "start": 313.66,
    "end": 330.12,
    "text": "Even though the question is, I mean, yes, it moves into the background, but still in the age of AI and code reviews becoming more important, the question is really like, is it still the most important thing like where your code is actually located?"
  },
  {
    "start": 330.24,
    "end": 349.94,
    "text": "or is it more important than like how people are actually able to to collaborate on that or how Asians are able to collaborate with human or how people actually... how collaboration looks like these days right and no longer necessarily on the code level but one level above that right."
  },
  {
    "start": 350.78,
    "end": 353.42,
    "text": "I guess that's like also part of your vision and you will tell us about that."
  },
  {
    "start": 353.5,
    "end": 358.3,
    "text": "but maybe let's move a bit back because I didn't like finish the introduction."
  },
  {
    "start": 358.4,
    "end": 361.16,
    "text": "so you studied CS at Harvard right."
  },
  {
    "start": 362.32,
    "end": 365.46,
    "text": "If I saw that correctly and worked for Airbnb for a few years."
  },
  {
    "start": 365.54,
    "end": 367.38,
    "text": "so you have quite a lot of infrastructure experience."
  },
  {
    "start": 367.84,
    "end": 370.02,
    "text": "What initially sparked your."
  },
  {
    "start": 370.88,
    "end": 374.36,
    "text": "Interesting computers like why are you a geek or nerd?"
  },
  {
    "start": 374.74,
    "end": 375.56,
    "text": "What's your notepad?"
  },
  {
    "start": 376.8,
    "end": 377.44,
    "text": "I have been."
  },
  {
    "start": 377.64,
    "end": 378.5,
    "text": "this is such a fun question."
  },
  {
    "start": 378.68,
    "end": 381.44,
    "text": "I've been Engineering for over half my life."
  },
  {
    "start": 381.98,
    "end": 382.72,
    "text": "I'm thirty years old."
  },
  {
    "start": 382.86,
    "end": 385.04,
    "text": "I started engineering when I was maybe fourteen or fifteen."
  },
  {
    "start": 387.12,
    "end": 390.94,
    "text": "In high school, I really enjoyed graphic design."
  },
  {
    "start": 391.46,
    "end": 395.58,
    "text": "I grew up mostly in Reno, Nevada Which is kind of a random place."
  },
  {
    "start": 396.0,
    "end": 398.6,
    "text": "not a lot of coders not a lot of kids doing software engineering."
  },
  {
    "start": 398.64,
    "end": 399.48,
    "text": "There weren't classes."
  },
  {
    "start": 399.68,
    "end": 409.12,
    "text": "at my high school I had one older friend who had picked up a little bit of coding and I think at one point we collaborated and made a free Android game."
  },
  {
    "start": 409.62,
    "end": 410.28,
    "text": "I did the design."
  },
  {
    "start": 410.52,
    "end": 411.2,
    "text": "He did the coding."
  },
  {
    "start": 411.36,
    "end": 411.9,
    "text": "That was fun."
  },
  {
    "start": 412.66,
    "end": 417.48,
    "text": "When I turned fifteen, my parents kindly but firmly were like, okay, you're fifteen."
  },
  {
    "start": 417.52,
    "end": 418.18,
    "text": "You got to get a job."
  },
  {
    "start": 418.3,
    "end": 420.0,
    "text": "We're not going to cover your gas money."
  },
  {
    "start": 420.1,
    "end": 422.72,
    "text": "We're not going to pay for you to go to the movies or clothing or anything."
  },
  {
    "start": 423.16,
    "end": 423.54,
    "text": "Go get a job."
  },
  {
    "start": 425.98,
    "end": 429.22,
    "text": "I went around town with my resume and I went to Starbucks and Red Robin and all these places."
  },
  {
    "start": 429.32,
    "end": 429.94,
    "text": "I mean my resume."
  },
  {
    "start": 430.04,
    "end": 441.14,
    "text": "when you're fifteen, it's a piece of paper that says your classes and I tried to get a job and this was right after the financial crisis and college educated people weren't getting jobs a little under fifteen year old."
  },
  {
    "start": 442.14,
    "end": 443.98,
    "text": "I'm also not sure it was legal for them to hire me."
  },
  {
    "start": 444.02,
    "end": 445.3,
    "text": "It would have been had to be under the table."
  },
  {
    "start": 447.1,
    "end": 451.38,
    "text": "So as a plan B, I started coding iPhone apps."
  },
  {
    "start": 452.18,
    "end": 454.72,
    "text": "I had that one experience watching a friend do it and I was like, okay."
  },
  {
    "start": 455.22,
    "end": 455.68,
    "text": "interesting."
  },
  {
    "start": 455.86,
    "end": 456.54,
    "text": "I think I could do this."
  },
  {
    "start": 457.04,
    "end": 460.52,
    "text": "And so I would watch it iTunes University course on coding iPhone apps."
  },
  {
    "start": 460.66,
    "end": 465.08,
    "text": "I'd look up, you know, coding for dummies books, which was in Java, but I could translate it."
  },
  {
    "start": 465.42,
    "end": 467.44,
    "text": "And I would just really hackishly try it."
  },
  {
    "start": 467.5,
    "end": 471.02,
    "text": "I made a pros and cons lap and I made more and more apps and they were bad."
  },
  {
    "start": 471.38,
    "end": 472.16,
    "text": "They were aesthetically good."
  },
  {
    "start": 472.3,
    "end": 475.44,
    "text": "The code was terrible because I hardly knew how to code, but I could piece it together."
  },
  {
    "start": 475.76,
    "end": 478.52,
    "text": "And they actually, you know, this is iOS four, iOS five."
  },
  {
    "start": 479.04,
    "end": 481.92,
    "text": "It was a good time to make me making apps as an indie developer."
  },
  {
    "start": 482.36,
    "end": 485.72,
    "text": "And that that wasn't able to make enough money to help me cover some expenses in high school."
  },
  {
    "start": 486.46,
    "end": 489.76,
    "text": "And I fell in love with engineering and I went deeper and deeper into it."
  },
  {
    "start": 489.84,
    "end": 490.82,
    "text": "I kept making iPhone apps."
  },
  {
    "start": 491.66,
    "end": 492.28,
    "text": "I went to college."
  },
  {
    "start": 492.38,
    "end": 495.04,
    "text": "I was like, oh my gosh, you know, I can actually take classes in this."
  },
  {
    "start": 495.12,
    "end": 495.6,
    "text": "This is great."
  },
  {
    "start": 495.7,
    "end": 496.8,
    "text": "And then I got internships."
  },
  {
    "start": 496.84,
    "end": 497.88,
    "text": "I'm like, oh, I can do internships."
  },
  {
    "start": 497.94,
    "end": 498.8,
    "text": "They will pay me to do this."
  },
  {
    "start": 498.84,
    "end": 499.54,
    "text": "This is fantastic."
  },
  {
    "start": 500.92,
    "end": 505.2,
    "text": "And high school was tough because I was trying to just rush through all my classes so I could go home and code."
  },
  {
    "start": 505.24,
    "end": 506.72,
    "text": "And then I was like, this could be the classes."
  },
  {
    "start": 507.24,
    "end": 508.06,
    "text": "I just fell in love with it."
  },
  {
    "start": 508.2,
    "end": 509.26,
    "text": "And that was iOS engineering."
  },
  {
    "start": 509.7,
    "end": 513.64,
    "text": "When I joined Airbnb, I joined as an iOS developer because I had many years of experience at that point."
  },
  {
    "start": 514.159,
    "end": 516.299,
    "text": "But I immediately got reorged onto an infrared tool."
  },
  {
    "start": 516.34,
    "end": 516.88,
    "text": "Interesting."
  },
  {
    "start": 517.64,
    "end": 521.039,
    "text": "I mean, if you came from design and then like apps and then to infra."
  },
  {
    "start": 521.9,
    "end": 523.5,
    "text": "Yeah."
  },
  {
    "start": 524.179,
    "end": 524.64,
    "text": "Exactly."
  },
  {
    "start": 524.68,
    "end": 529.26,
    "text": "And I'm like, Oh, wow, servers, databases, online systems is very different than shipping a client to a user's phone."
  },
  {
    "start": 529.7,
    "end": 531.64,
    "text": "Um, trial by fire."
  },
  {
    "start": 531.84,
    "end": 532.38,
    "text": "I figured it out."
  },
  {
    "start": 532.5,
    "end": 532.88,
    "text": "I learned it."
  },
  {
    "start": 533.28,
    "end": 538.76,
    "text": "I got really into SRE style work and volunteering on release management and just coding a ton of backend systems."
  },
  {
    "start": 539.56,
    "end": 548.22,
    "text": "And, uh, What I realized, and this carried me into graphite, was the sense of building an interesting tool product server."
  },
  {
    "start": 548.5,
    "end": 548.92,
    "text": "That's fun."
  },
  {
    "start": 549.16,
    "end": 549.74,
    "text": "That's interesting."
  },
  {
    "start": 550.1,
    "end": 552.62,
    "text": "However, if you get good enough of software engineering, it's pretty easy."
  },
  {
    "start": 553.42,
    "end": 556.74,
    "text": "It takes time and effort, but it's not a technically unsolved problem."
  },
  {
    "start": 557.52,
    "end": 562.22,
    "text": "What I found at Airbnb on DevTools was the more interesting question was not, how do you build a server?"
  },
  {
    "start": 562.32,
    "end": 565.44,
    "text": "It was, how should all engineers work together to build things?"
  },
  {
    "start": 565.66,
    "end": 567.5,
    "text": "What is the art of software engineering?"
  },
  {
    "start": 567.56,
    "end": 568.56,
    "text": "How do you do it really well?"
  },
  {
    "start": 569.22,
    "end": 574.58,
    "text": "How should people work on pull requests and what are the patterns of committing and releasing and shipping?"
  },
  {
    "start": 575.38,
    "end": 577.92,
    "text": "That meta question was much more interesting to me and much more unsolved."
  },
  {
    "start": 578.02,
    "end": 578.84,
    "text": "And I saw that developing."
  },
  {
    "start": 579.1,
    "end": 580.34,
    "text": "Honestly, there was so much low hanging fruit."
  },
  {
    "start": 580.38,
    "end": 582.96,
    "text": "I was watching Airbnb adopt merge cues and adopt something."
  },
  {
    "start": 583.0,
    "end": 586.82,
    "text": "I'm like, oh my gosh, it's twenty, twenty eighteen and we're just starting to adopt these systems."
  },
  {
    "start": 586.9,
    "end": 588.16,
    "text": "Like this is low hanging fruit."
  },
  {
    "start": 588.82,
    "end": 589.72,
    "text": "And so I became obsessed with that."
  },
  {
    "start": 589.76,
    "end": 598.86,
    "text": "And when we went into with my co-founders creating graphite, I brought that obsession of how can we continue iterating on how engineers create code and collaborate together."
  },
  {
    "start": 598.92,
    "end": 600.24,
    "text": "That's my, I think my life's passion."
  },
  {
    "start": 600.58,
    "end": 600.8,
    "text": "Okay."
  },
  {
    "start": 601.0,
    "end": 601.18,
    "text": "Okay."
  },
  {
    "start": 601.26,
    "end": 601.66,
    "text": "Interesting."
  },
  {
    "start": 602.3,
    "end": 610.9,
    "text": "You mentioned that you pivoted, uh, but I, I'd also say that like just looking at this space, it must be a constant pivot, right?"
  },
  {
    "start": 610.96,
    "end": 621.1,
    "text": "Like, uh, whenever you approach or, uh, think that you, you, you, you, you finally got PMF, then all of a sudden something comes along."
  },
  {
    "start": 621.16,
    "end": 624.12,
    "text": "Like then you think, okay, now cursor is the hottest thing in town."
  },
  {
    "start": 624.7,
    "end": 627.42,
    "text": "And then, uh, You got acquired by a cursor, by the way."
  },
  {
    "start": 628.08,
    "end": 631.32,
    "text": "And then the next thing is like, hey, everyone's using a different tool now."
  },
  {
    "start": 631.4,
    "end": 632.36,
    "text": "Like, I don't know."
  },
  {
    "start": 632.64,
    "end": 632.88,
    "text": "I don't know."
  },
  {
    "start": 632.96,
    "end": 643.02,
    "text": "These days, like tomorrow, maybe Codex and the whole ideal setup of companies also is changing, right?"
  },
  {
    "start": 643.06,
    "end": 648.54,
    "text": "Like two years ago, maybe it was still a thing to have like two thousand engineers."
  },
  {
    "start": 648.64,
    "end": 653.88,
    "text": "Now everyone thinks about like twenty engineers and very efficient ones and a lot of agents."
  },
  {
    "start": 655.22,
    "end": 664.08,
    "text": "How do you navigate that, or how did you manage to navigate that very ambiguous world, still with huge valuations, right?"
  },
  {
    "start": 664.24,
    "end": 667.44,
    "text": "You had like two and a fifty million valuations in your last round, I think."
  },
  {
    "start": 668.06,
    "end": 673.82,
    "text": "And like with, I'm sure a product market fit, but not an easy environment."
  },
  {
    "start": 673.88,
    "end": 674.74,
    "text": "How did you manage that?"
  },
  {
    "start": 675.26,
    "end": 676.16,
    "text": "Absolutely, yeah."
  },
  {
    "start": 677.22,
    "end": 677.72,
    "text": "It's funny."
  },
  {
    "start": 678.8,
    "end": 683.18,
    "text": "DevTools wasn't always a super fast moving space, in my opinion."
  },
  {
    "start": 683.88,
    "end": 687.04,
    "text": "Tech often has a center point in the industry."
  },
  {
    "start": 687.12,
    "end": 691.2,
    "text": "that's catching fire and innovating really quickly and changing it all."
  },
  {
    "start": 691.82,
    "end": 693.36,
    "text": "AI is exceptionally fast."
  },
  {
    "start": 693.44,
    "end": 695.56,
    "text": "Maybe the fastest we've ever seen, but there's often a center point."
  },
  {
    "start": 696.0,
    "end": 697.26,
    "text": "FinTech was this for a little while."
  },
  {
    "start": 697.32,
    "end": 698.16,
    "text": "No code was this."
  },
  {
    "start": 698.3,
    "end": 699.16,
    "text": "Crypto was this."
  },
  {
    "start": 699.9,
    "end": 706.28,
    "text": "Debt tools for the two thousands in the twenty tens and into twenty twenty was not that fast moving of a space."
  },
  {
    "start": 706.36,
    "end": 708.98,
    "text": "You weren't seeing crazy innovations happen overnight."
  },
  {
    "start": 709.2,
    "end": 717.16,
    "text": "I mean, Take GitHub, one of the more interesting dev tools of our time, was founded in the mid-twos."
  },
  {
    "start": 717.8,
    "end": 723.4,
    "text": "Throughout the arc of twenty-ten didn't have major iterations or speed behind it."
  },
  {
    "start": 723.66,
    "end": 724.46,
    "text": "I don't mean as a knock."
  },
  {
    "start": 725.06,
    "end": 728.46,
    "text": "They have done some amazing infrastructure and I really like what they've done for open source."
  },
  {
    "start": 728.88,
    "end": 731.2,
    "text": "But it's not like this stuff was reinventing itself every year."
  },
  {
    "start": 731.38,
    "end": 734.66,
    "text": "You know, it's not like PRs were like, there's a new PR page every year."
  },
  {
    "start": 734.72,
    "end": 736.38,
    "text": "You weren't seeing that velocity in AI."
  },
  {
    "start": 736.64,
    "end": 739.64,
    "text": "Yeah, you actually are seeing a whole new thing every year and you're keeping up with that."
  },
  {
    "start": 739.68,
    "end": 747.88,
    "text": "So to me, it wasn't until co-pilot and cursor in about twenty twenty three, where DevTool suddenly now entered extreme velocity mode."
  },
  {
    "start": 748.12,
    "end": 750.02,
    "text": "And as you say, now we have to stay on our game."
  },
  {
    "start": 750.3,
    "end": 754.0,
    "text": "Now we have to be reinventing the wheel and constantly exploring new ideas and being in that bleeding edge."
  },
  {
    "start": 754.04,
    "end": 755.3,
    "text": "The other thing I'll say is it was."
  },
  {
    "start": 755.86,
    "end": 762.48,
    "text": "DevTools had a really interesting pattern where most good ideas have been invented within large tech companies."
  },
  {
    "start": 762.82,
    "end": 772.48,
    "text": "If there was a better way to do something to look at a code change or merge it in or leave comments or test it, usually Google or Facebook or Amazon had done it."
  },
  {
    "start": 772.66,
    "end": 774.02,
    "text": "Someone had done it at one of these companies."
  },
  {
    "start": 774.66,
    "end": 780.98,
    "text": "And a lot of this was just taking some of those good ideas and then distributing them out to the rest of the industry in a way that teams didn't have to keep rebuilding."
  },
  {
    "start": 781.22,
    "end": 786.7,
    "text": "It was one of like building a generalized version of it, marketing it, distributing it to engineers."
  },
  {
    "start": 787.98,
    "end": 788.86,
    "text": "Now it's different."
  },
  {
    "start": 789.2,
    "end": 793.02,
    "text": "Now with AI, there is a lot of new ideas that have never been done before."
  },
  {
    "start": 793.1,
    "end": 797.2,
    "text": "And low-hanging fruit, it's very exciting, but hasn't yet been battle tested."
  },
  {
    "start": 797.56,
    "end": 800.68,
    "text": "It's not just taking tenure ideas and shipping them to the rest of the industry."
  },
  {
    "start": 800.92,
    "end": 801.14,
    "text": "I also"
  },
  {
    "start": 801.24,
    "end": 809.92,
    "text": "think that this is because for LLMs, this turned out to be maybe the most viable."
  },
  {
    "start": 810.42,
    "end": 812.92,
    "text": "business case because like I mean engineers are expensive."
  },
  {
    "start": 814.3,
    "end": 817.68,
    "text": "I mean no doubt right and it's a very complex world."
  },
  {
    "start": 817.82,
    "end": 831.52,
    "text": "and if you manage to really solve this and there's a lot of documented success cases right and examples of code and you can test it deterministically or you can you can basically test the results."
  },
  {
    "start": 832.36,
    "end": 838.66,
    "text": "that makes it very comfortable for LLMs to generate code right and then many companies jumped on that."
  },
  {
    "start": 839.66,
    "end": 840.52,
    "text": "How did that feel?"
  },
  {
    "start": 840.68,
    "end": 849.22,
    "text": "I mean, you had that system running, you had that stacked PRs built for many engineers collaborating with each other."
  },
  {
    "start": 850.2,
    "end": 857.46,
    "text": "Did that feel like a present that then everyone was jumping on it and that you heard like, oh, okay, no, OpenAI is jumping on it as well."
  },
  {
    "start": 857.54,
    "end": 861.46,
    "text": "There's Codex and oh, no, there's Cursar, no, there's Claude."
  },
  {
    "start": 861.84,
    "end": 865.48,
    "text": "Did that feel like a present or did that rather feel like a hit?"
  },
  {
    "start": 865.94,
    "end": 868.14,
    "text": "It was so good for us in so many ways."
  },
  {
    "start": 868.82,
    "end": 871.18,
    "text": "There's the product side, there's the adoption market side."
  },
  {
    "start": 871.48,
    "end": 874.34,
    "text": "On the product side, it allowed us to do some cool new stuff."
  },
  {
    "start": 874.56,
    "end": 876.66,
    "text": "We can get into the game of leaving comments through AI."
  },
  {
    "start": 877.18,
    "end": 878.22,
    "text": "That's a classic simple idea."
  },
  {
    "start": 879.28,
    "end": 884.36,
    "text": "But on the just general business side, it was so helpful because a variety of factors."
  },
  {
    "start": 885.0,
    "end": 888.18,
    "text": "One factor, people were creating more code than ever before."
  },
  {
    "start": 888.88,
    "end": 896.02,
    "text": "Every year for the past few years, you're seeing almost a doubling plus in the number of PRs per engineer that are being created and shipped to these large companies."
  },
  {
    "start": 896.74,
    "end": 904.44,
    "text": "Suddenly you have a very high demand for help finding, reviewing, merging in a bunch of high volume PRs."
  },
  {
    "start": 904.48,
    "end": 911.08,
    "text": "And it turns out that all the techniques that help with stacking are also really helpful for just a large number of rapidly created PRs."
  },
  {
    "start": 911.38,
    "end": 919.82,
    "text": "Yeah, you can generate them a lot faster, but without some innovation on the other side, the outer loop, it's no easier to review, test and merge them."
  },
  {
    "start": 919.88,
    "end": 921.16,
    "text": "That's what we've been kind of obsessed with."
  },
  {
    "start": 921.2,
    "end": 924.78,
    "text": "So it made a higher demand for some of the things we were already building."
  },
  {
    "start": 925.08,
    "end": 927.32,
    "text": "whether it be stacking, merge queues, inboxes, and so on."
  },
  {
    "start": 927.48,
    "end": 928.1,
    "text": "That was exciting."
  },
  {
    "start": 928.54,
    "end": 932.92,
    "text": "It also made it easier for developers to honestly pay for and adopt tools at companies."
  },
  {
    "start": 933.6,
    "end": 945.0,
    "text": "For a long time, especially, I remember this in the in the in the twenty-tens, you had to kind of like really beg your manager to let you buy, you know, IntelliJ or like a fifty dollar tool as an engineer and it would be really scrutinized."
  },
  {
    "start": 945.4,
    "end": 948.44,
    "text": "And realistically, most of it was just free open source like the S code."
  },
  {
    "start": 949.42,
    "end": 953.74,
    "text": "And then also companies were very worried, wary about giving access to their source control."
  },
  {
    "start": 954.1,
    "end": 955.92,
    "text": "maybe for right reasons, maybe for overblown reasons."
  },
  {
    "start": 956.88,
    "end": 960.42,
    "text": "Suddenly, with cursor and twenty other tools, it blew the doors off."
  },
  {
    "start": 960.78,
    "end": 965.02,
    "text": "Everyone overnight was shipping their code bases over the wire to these startups."
  },
  {
    "start": 965.9,
    "end": 968.34,
    "text": "Everyone was expensing twenty dollar tools overnight."
  },
  {
    "start": 969.1,
    "end": 972.66,
    "text": "Our procurement discussions got so much easier and our adoption got so much easier."
  },
  {
    "start": 972.78,
    "end": 975.04,
    "text": "It helped the entire DevTools industry to adopt these tools."
  },
  {
    "start": 975.08,
    "end": 990.14,
    "text": "So we really benefited from a lot of these waves and I think we're still on the very early curve because all the conversations that are happening right now in the industry, people are viewing Code Gen as early solved."
  },
  {
    "start": 990.3,
    "end": 990.86,
    "text": "It's going to get better."
  },
  {
    "start": 990.9,
    "end": 995.52,
    "text": "It's going to get better and better, but no one disagrees that it's pretty great to generate code right now."
  },
  {
    "start": 996.08,
    "end": 999.78,
    "text": "The more open-ended question is how do you review it?"
  },
  {
    "start": 999.86,
    "end": 1000.46,
    "text": "How do you test it?"
  },
  {
    "start": 1000.82,
    "end": 1001.78,
    "text": "How do you merge it in really well?"
  },
  {
    "start": 1001.82,
    "end": 1005.84,
    "text": "How do you validate that the code that was generated is actually correct and good in what you want to be?"
  },
  {
    "start": 1006.2,
    "end": 1007.56,
    "text": "Some humans got to read it, maybe."
  },
  {
    "start": 1007.86,
    "end": 1009.18,
    "text": "Or an AI can help you read it, maybe."
  },
  {
    "start": 1009.7,
    "end": 1010.02,
    "text": "And how do you?"
  },
  {
    "start": 1010.48,
    "end": 1012.82,
    "text": "Is your testing deterministic or is an AI helping you test it?"
  },
  {
    "start": 1013.36,
    "end": 1014.28,
    "text": "And you still got to merge that."
  },
  {
    "start": 1014.48,
    "end": 1015.62,
    "text": "There still has to be emerging systems."
  },
  {
    "start": 1015.66,
    "end": 1016.62,
    "text": "That's still nothing changed."
  },
  {
    "start": 1016.7,
    "end": 1017.72,
    "text": "Deployments still have to happen."
  },
  {
    "start": 1017.84,
    "end": 1019.4,
    "text": "So that's the new bottleneck."
  },
  {
    "start": 1019.76,
    "end": 1021.2,
    "text": "And that's where a lot of the attention is shifting."
  },
  {
    "start": 1021.24,
    "end": 1023.46,
    "text": "And that's where, again, I'm very excited to help with that."
  },
  {
    "start": 1023.5,
    "end": 1026.119,
    "text": "We're not the only company doing this, but I'm excited to kind of help on that bleeding edge."
  },
  {
    "start": 1026.78,
    "end": 1032.54,
    "text": "So you just mentioned, in your answer, you just mentioned, poor requests per engineer."
  },
  {
    "start": 1033.52,
    "end": 1036.38,
    "text": "And that number has nearly doubled recently."
  },
  {
    "start": 1037.42,
    "end": 1039.099,
    "text": "Do you think this is a valid?"
  },
  {
    "start": 1039.72,
    "end": 1048.359,
    "text": "metric to like measure efficiency of engineering orgs or what do you think like I mean what lines of code is seemingly dead right."
  },
  {
    "start": 1050.14,
    "end": 1054.66,
    "text": "How did that develop on your end like your perspective on that and your perspective on Dora metrics."
  },
  {
    "start": 1055.62,
    "end": 1057.56,
    "text": "Managers have wanted this for a long time right?"
  },
  {
    "start": 1057.6,
    "end": 1062.82,
    "text": "like how do you look at your engineering team your engineering org and know how productive they are?"
  },
  {
    "start": 1062.88,
    "end": 1064.38,
    "text": "spend a timeless question and."
  },
  {
    "start": 1065.22,
    "end": 1069.46,
    "text": "Any engineer will tell you, well, hey, you know, don't just look at how much I'm writing."
  },
  {
    "start": 1069.5,
    "end": 1070.52,
    "text": "I'll gamify that metric."
  },
  {
    "start": 1070.62,
    "end": 1072.12,
    "text": "Some of my PRs are really important."
  },
  {
    "start": 1072.18,
    "end": 1077.76,
    "text": "Sometimes I write a one-line change, and it took me a week, and it saved a ton of bugs or money."
  },
  {
    "start": 1078.24,
    "end": 1079.12,
    "text": "These are all valid arguments."
  },
  {
    "start": 1079.52,
    "end": 1082.52,
    "text": "However, however, I've been looking at this for a long time."
  },
  {
    "start": 1082.56,
    "end": 1088.18,
    "text": "I've been looking at this for five years, and we're used by, like, thousands, thousands of organizations, including many of the best engineering teams."
  },
  {
    "start": 1088.64,
    "end": 1091.92,
    "text": "And I will say at scale, actually, it is kind of just the number of PRs being shipped."
  },
  {
    "start": 1092.48,
    "end": 1106.82,
    "text": "you know, if you assume that engineers in general aren't trying to game their numbers and they're writing PRs and each PR is a discrete idea and it's between, you know, fifty and a thousand lines and it's some incremental change and there's some fixed cost to releasing it."
  },
  {
    "start": 1106.98,
    "end": 1114.48,
    "text": "It actually turns out that like measuring on a PR per PR basis is a relatively strong proxy metric to the velocity of the engineering team."
  },
  {
    "start": 1114.82,
    "end": 1116.56,
    "text": "You want to look at the median size of those PRs."
  },
  {
    "start": 1116.6,
    "end": 1118.26,
    "text": "You want to look at the time to merge on those PRs."
  },
  {
    "start": 1119.08,
    "end": 1123.06,
    "text": "the rollback rates, the churn of code, how much is the same code being rewritten over and over again?"
  },
  {
    "start": 1123.1,
    "end": 1124.52,
    "text": "There's these guardrail metrics."
  },
  {
    "start": 1125.48,
    "end": 1130.58,
    "text": "Every time I've gone deep on the data science here, though, it turns out, in my opinion, you can actually just look at the PR being shipped."
  },
  {
    "start": 1131.24,
    "end": 1135.38,
    "text": "What's fascinating to me, though, here's what I disagree with as a hot tech in the industry."
  },
  {
    "start": 1136.0,
    "end": 1143.38,
    "text": "You hear these arguments like, okay, hey, if the engineers are shipping more PRs and can do more with AI systems, maybe we just need fewer engineers."
  },
  {
    "start": 1143.88,
    "end": 1147.18,
    "text": "Maybe a two thousand percent team can be a two hundred percent team or twenty percent team."
  },
  {
    "start": 1148.0,
    "end": 1148.58,
    "text": "I don't agree with this."
  },
  {
    "start": 1148.76,
    "end": 1153.86,
    "text": "I don't see evidence of this both at Graphite, at Cursor, at Larger Companies and Friends Companies."
  },
  {
    "start": 1153.94,
    "end": 1154.72,
    "text": "I don't see this happening."
  },
  {
    "start": 1155.76,
    "end": 1158.56,
    "text": "I think that the job of an engineer is not to write code."
  },
  {
    "start": 1158.6,
    "end": 1161.44,
    "text": "The job of an engineer is to solve problems and ship great products."
  },
  {
    "start": 1162.22,
    "end": 1164.22,
    "text": "You write code to do it, but I also do other stuff."
  },
  {
    "start": 1164.44,
    "end": 1167.86,
    "text": "I also write design docs and I have meetings and I try to come up with good ideas."
  },
  {
    "start": 1167.9,
    "end": 1171.04,
    "text": "I also do DevOps and I try to figure out why there's a bug in production."
  },
  {
    "start": 1171.6,
    "end": 1173.74,
    "text": "There's a lot I do in my art of engineering."
  },
  {
    "start": 1174.82,
    "end": 1177.32,
    "text": "The goal here is to solve problems and build some great stuff."
  },
  {
    "start": 1177.48,
    "end": 1179.58,
    "text": "I think part of that just got a lot easier."
  },
  {
    "start": 1179.62,
    "end": 1180.82,
    "text": "It's a lot easier to write some code right now."
  },
  {
    "start": 1181.24,
    "end": 1182.12,
    "text": "A lot of this didn't get easier."
  },
  {
    "start": 1182.28,
    "end": 1183.62,
    "text": "My DevOps isn't easier yet."
  },
  {
    "start": 1184.3,
    "end": 1185.4,
    "text": "My debugging, it's a little easier."
  },
  {
    "start": 1185.46,
    "end": 1187.78,
    "text": "I can ask some questions, but like I still got to go do that."
  },
  {
    "start": 1188.36,
    "end": 1189.94,
    "text": "Deploying code still takes thirty minutes."
  },
  {
    "start": 1190.02,
    "end": 1191.24,
    "text": "I still got to look at a data dashboard."
  },
  {
    "start": 1191.84,
    "end": 1194.44,
    "text": "My meeting for an architecture review is roughly the same as it used to be."
  },
  {
    "start": 1194.86,
    "end": 1198.7,
    "text": "So I think we made all the engineers a little more productive."
  },
  {
    "start": 1199.02,
    "end": 1200.1,
    "text": "We still need a ton of engineers."
  },
  {
    "start": 1200.18,
    "end": 1203.44,
    "text": "And if we make them all two X more productive, we're still going to want a ton of engineers."
  },
  {
    "start": 1204.24,
    "end": 1205.5,
    "text": "I don't see this appetite slowing down."
  },
  {
    "start": 1205.88,
    "end": 1220.48,
    "text": "You look historically, every time there's been a speed boost, when we moved off assembly, when we invented CI, when we created collaborative source code control, when our MacBooks got faster, there's been so many level ups in the history of engineering in the last forty years."
  },
  {
    "start": 1221.5,
    "end": 1222.34,
    "text": "I can't think of any of these."
  },
  {
    "start": 1222.44,
    "end": 1225.14,
    "text": "that then caused the industry to be like, great, we don't need a lot of engineers now."
  },
  {
    "start": 1225.36,
    "end": 1228.1,
    "text": "And I don't think that this wave will be the one to change"
  },
  {
    "start": 1228.58,
    "end": 1228.6,
    "text": "it."
  },
  {
    "start": 1228.66,
    "end": 1231.92,
    "text": "So is this like then punch cards to C++, I guess, right?"
  },
  {
    "start": 1233.26,
    "end": 1234.14,
    "text": "Or maybe a bit more."
  },
  {
    "start": 1234.96,
    "end": 1235.84,
    "text": "I think that's a better analogy."
  },
  {
    "start": 1236.46,
    "end": 1237.16,
    "text": "I think it's a better analogy."
  },
  {
    "start": 1237.3,
    "end": 1238.4,
    "text": "I think it's exactly."
  },
  {
    "start": 1238.44,
    "end": 1239.48,
    "text": "And great."
  },
  {
    "start": 1239.9,
    "end": 1241.3,
    "text": "Now I don't have to do some punch card work."
  },
  {
    "start": 1241.66,
    "end": 1246.84,
    "text": "Maybe I don't have to do as much, you know, syntax fixing as I type out my ideas."
  },
  {
    "start": 1247.4,
    "end": 1250.74,
    "text": "But we're all going, I think there's just so much good work for us all to still be doing."
  },
  {
    "start": 1251.5,
    "end": 1251.8,
    "text": "And."
  },
  {
    "start": 1252.96,
    "end": 1256.18,
    "text": "I don't have a single shred of evidence to say that it's slowing down."
  },
  {
    "start": 1256.22,
    "end": 1257.66,
    "text": "I don't see it on a team basis."
  },
  {
    "start": 1257.76,
    "end": 1258.78,
    "text": "I don't see it on a company basis."
  },
  {
    "start": 1258.88,
    "end": 1259.22,
    "text": "I'm not happy."
  },
  {
    "start": 1260.22,
    "end": 1269.84,
    "text": "Every time I hear this notification, it seems to be more of a people armchair hypothesizing this and people worrying about an AI market crash or worrying about the downsides of what might be happening."
  },
  {
    "start": 1270.72,
    "end": 1273.72,
    "text": "But I just can't, I'm just personally not finding evidence on the ground of this."
  },
  {
    "start": 1276.9,
    "end": 1279.88,
    "text": "Biggest recent aha moment when you think okay this is this is."
  },
  {
    "start": 1280.02,
    "end": 1281.2,
    "text": "this is really crazy like."
  },
  {
    "start": 1281.56,
    "end": 1283.16,
    "text": "I mean it's like we like."
  },
  {
    "start": 1283.26,
    "end": 1288.82,
    "text": "also in the discussion we had before we agreed that this is like kind of a new abstraction layer to like how we use computers."
  },
  {
    "start": 1289.56,
    "end": 1295.64,
    "text": "And I regularly still have those moments when I use I don't know Claude or cursor where."
  },
  {
    "start": 1296.98,
    "end": 1298.2,
    "text": "It really blows my mind."
  },
  {
    "start": 1298.28,
    "end": 1299.26,
    "text": "what is possible now?"
  },
  {
    "start": 1300.34,
    "end": 1302.08,
    "text": "do you have an example for one moment?"
  },
  {
    "start": 1302.16,
    "end": 1303.68,
    "text": "where like you recently thought?"
  },
  {
    "start": 1303.76,
    "end": 1305.24,
    "text": "okay this this really blows my mind."
  },
  {
    "start": 1306.9,
    "end": 1307.84,
    "text": "One that I've been thinking about recently."
  },
  {
    "start": 1307.88,
    "end": 1309.78,
    "text": "I don't know if this is a perfect answer to your question."
  },
  {
    "start": 1310.94,
    "end": 1312.14,
    "text": "I I fun fun hacks."
  },
  {
    "start": 1312.28,
    "end": 1317.12,
    "text": "I help it, you know, I use a. I'd help me read groups of books together."
  },
  {
    "start": 1317.56,
    "end": 1320.36,
    "text": "I use it to help write me quick scripts on my computer to do operations."
  },
  {
    "start": 1320.46,
    "end": 1321.2,
    "text": "I do all sorts of stuff."
  },
  {
    "start": 1321.56,
    "end": 1329.64,
    "text": "But one of the really interesting things I've been thinking about is actually in the process of graphite merging with cursor merging our engineering teams, and we're obsessed with code review."
  },
  {
    "start": 1329.7,
    "end": 1330.46,
    "text": "We love code review."
  },
  {
    "start": 1330.5,
    "end": 1331.44,
    "text": "Why can't, why wouldn't we?"
  },
  {
    "start": 1331.52,
    "end": 1332.06,
    "text": "We're a code review"
  },
  {
    "start": 1332.14,
    "end": 1332.48,
    "text": "company."
  },
  {
    "start": 1333.36,
    "end": 1341.74,
    "text": "And we built an AI code review, but we really deeply believe that it's important that humans are helping review the code and providing a second set of eyes alongside the AI."
  },
  {
    "start": 1341.92,
    "end": 1342.52,
    "text": "We believe in that."
  },
  {
    "start": 1343.36,
    "end": 1347.44,
    "text": "The cursor engineering team doesn't mandate code reviews in lots of parts of the code base."
  },
  {
    "start": 1347.5,
    "end": 1349.0,
    "text": "I don't want to speak super broadly."
  },
  {
    "start": 1350.78,
    "end": 1355.42,
    "text": "infrastructure team, certain folders of code owners, but there's a lot of product teams who it's not required."
  },
  {
    "start": 1355.74,
    "end": 1362.58,
    "text": "They can opt in to code review or they can trust the AI agent who's reviewing them and just merge in right away and maintain really fast speed."
  },
  {
    "start": 1363.3,
    "end": 1365.88,
    "text": "This was kind of shocking to me and it's not just a cursor."
  },
  {
    "start": 1365.98,
    "end": 1370.94,
    "text": "I've heard this from Exa and a few other friend startups and it's interesting."
  },
  {
    "start": 1371.28,
    "end": 1373.3,
    "text": "When I first hear that, I'm like, oh, bad engineering."
  },
  {
    "start": 1373.9,
    "end": 1375.18,
    "text": "You all are probably creating a messy code bit."
  },
  {
    "start": 1375.76,
    "end": 1376.56,
    "text": "This is going to be terrible."
  },
  {
    "start": 1377.8,
    "end": 1379.58,
    "text": "And I think that that's an overreaction on my part."
  },
  {
    "start": 1379.74,
    "end": 1383.62,
    "text": "I actually think that that's like, I need to get over some of my own biases there."
  },
  {
    "start": 1384.32,
    "end": 1401.82,
    "text": "And I'm curious and excited to see, you know, is there a portion of code changes in the world through the assistance of AI creating them, the assistance of AI helping review them, that a single author can now take responsibility and ship that, and you don't need a second pair of eyes to help review it."
  },
  {
    "start": 1402.6,
    "end": 1410.4,
    "text": "One idea I have and would be said by is, Could an LLM judge stare at each inbound code change and say hey, this one's pretty safe."
  },
  {
    "start": 1410.58,
    "end": 1414.22,
    "text": "We don't need a second human reviewer and we say hey this one's pretty dangerous this one at this one."
  },
  {
    "start": 1414.26,
    "end": 1416.36,
    "text": "We're gonna need a human reviewer and can we can?"
  },
  {
    "start": 1416.4,
    "end": 1421.44,
    "text": "we actually almost have an intake judge helping make those decisions and maybe hassle your coaching just need to review But half don't."
  },
  {
    "start": 1421.86,
    "end": 1422.94,
    "text": "it's the slowest part of the process."
  },
  {
    "start": 1423.28,
    "end": 1424.18,
    "text": "you look at the numbers here."
  },
  {
    "start": 1425.04,
    "end": 1429.66,
    "text": "writing the code change Can take you anywhere from two minutes to ten hours."
  },
  {
    "start": 1429.76,
    "end": 1432.78,
    "text": "I don't know it takes a lot of time but the AI it's getting really really short."
  },
  {
    "start": 1433.36,
    "end": 1436.26,
    "text": "Testing the code change takes about thirty minutes to run your CI, give or take."
  },
  {
    "start": 1436.68,
    "end": 1440.32,
    "text": "Releasing it thirty minutes, again, depends how fast or slow your release process."
  },
  {
    "start": 1441.04,
    "end": 1442.9,
    "text": "Human code review is super slow."
  },
  {
    "start": 1443.54,
    "end": 1450.24,
    "text": "It takes on average and on medium like twelve to eighteen to thirty hours depending on the organization."
  },
  {
    "start": 1450.46,
    "end": 1450.96,
    "text": "It's insane."
  },
  {
    "start": 1451.3,
    "end": 1453.06,
    "text": "And it's not that it took someone eighteen hours to read it."
  },
  {
    "start": 1453.28,
    "end": 1457.44,
    "text": "It took them eighteen hours to like wake up, look at the notification and actually choose to open that PR."
  },
  {
    "start": 1458.1,
    "end": 1460.68,
    "text": "It's surely definitely the slowest part of this problem."
  },
  {
    "start": 1460.72,
    "end": 1463.6,
    "text": "Even the fastest companies, it's like you're bugging the crap out of your teammate."
  },
  {
    "start": 1463.84,
    "end": 1466.18,
    "text": "It still takes a couple hours to like distract them and get them to look at it."
  },
  {
    "start": 1466.7,
    "end": 1468.04,
    "text": "If we can slim that down."
  },
  {
    "start": 1469.02,
    "end": 1473.32,
    "text": "through automated tools or reduce parts of it, I think you'll see huge performance games."
  },
  {
    "start": 1473.4,
    "end": 1474.92,
    "text": "I'm very, very interested in this."
  },
  {
    "start": 1475.82,
    "end": 1477.34,
    "text": "But like I said, you asked for something surprising."
  },
  {
    "start": 1477.38,
    "end": 1480.78,
    "text": "This is kind of surprising because it goes against my own biases that I've been working on a lot."
  },
  {
    "start": 1480.86,
    "end": 1482.74,
    "text": "And I want to find what's the correct intersection."
  },
  {
    "start": 1483.24,
    "end": 1490.02,
    "text": "There's ideas around post-merge review, for example, where AI provides a review, human merges it, and then a human later reviews it afterwards."
  },
  {
    "start": 1490.24,
    "end": 1497.72,
    "text": "So this pretty much, like if I would ask you now, like how software is built in the next two to three years, you think this is part of it?"
  },
  {
    "start": 1498.48,
    "end": 1506.5,
    "text": "Basically, we make responsibilities depending on the complexity of the task at hand or the complexity of the change."
  },
  {
    "start": 1507.2,
    "end": 1509.5,
    "text": "Something that needs to break."
  },
  {
    "start": 1509.92,
    "end": 1511.78,
    "text": "We got to make this outer loop process faster."
  },
  {
    "start": 1512.46,
    "end": 1513.7,
    "text": "It's absolutely critical."
  },
  {
    "start": 1514.5,
    "end": 1515.3,
    "text": "It's not even a question."
  },
  {
    "start": 1516.18,
    "end": 1521.04,
    "text": "The industry is going to demand it and force it to break and change and evolve because Kojen is so good now."
  },
  {
    "start": 1521.56,
    "end": 1523.34,
    "text": "And Kojen is going to keep getting better, but Kojen is so good."
  },
  {
    "start": 1523.98,
    "end": 1525.62,
    "text": "So we got to find a way."
  },
  {
    "start": 1526.84,
    "end": 1528.7,
    "text": "It's the validation and the release of these code changes."
  },
  {
    "start": 1529.52,
    "end": 1536.12,
    "text": "When I say validation, it's the human review and the deterministic testing, the CI parts of it, and the release part."
  },
  {
    "start": 1536.32,
    "end": 1538.54,
    "text": "How do you merge it in and actually get it running on a server in production?"
  },
  {
    "start": 1539.16,
    "end": 1541.0,
    "text": "That part is very clearly the bottleneck right now."
  },
  {
    "start": 1541.86,
    "end": 1547.08,
    "text": "Elements are CI, merge queue, AI code review, complex deployment infrastructure."
  },
  {
    "start": 1547.58,
    "end": 1550.5,
    "text": "I think we need to find ways to speed that up."
  },
  {
    "start": 1550.74,
    "end": 1562.84,
    "text": "Otherwise, what happens is every morning at these new companies, an engineering lead, a tech lead wakes up and their, you know, their new grad teammates who are really smart and really, really excited have shipped them fifteen code changes each."
  },
  {
    "start": 1563.04,
    "end": 1565.82,
    "text": "And now the tech lead has fifty code changes reviewed that morning."
  },
  {
    "start": 1566.36,
    "end": 1570.02,
    "text": "And honestly, like some of the engineers didn't even read them when they sent them over the wire."
  },
  {
    "start": 1570.08,
    "end": 1572.86,
    "text": "and it's chaos and it's blocking and it's like backbreaking."
  },
  {
    "start": 1573.82,
    "end": 1575.14,
    "text": "We need to evolve this process."
  },
  {
    "start": 1575.5,
    "end": 1576.14,
    "text": "There's challenges."
  },
  {
    "start": 1576.82,
    "end": 1581.56,
    "text": "Let's say we hand it all to a really smart AI code reviewer who actually is pretty good at validating and accepting these code changes."
  },
  {
    "start": 1582.4,
    "end": 1587.24,
    "text": "I have a new problem, which is that the author might not have read it and the reviewer didn't read it."
  },
  {
    "start": 1587.72,
    "end": 1591.48,
    "text": "So now there's no human at this company who quite knows exactly what's running in production."
  },
  {
    "start": 1591.96,
    "end": 1592.58,
    "text": "That's now."
  },
  {
    "start": 1592.62,
    "end": 1594.78,
    "text": "I have an accountability problem and I have an SRE problem."
  },
  {
    "start": 1595.28,
    "end": 1599.44,
    "text": "Now when the site goes down, who deeply knows how to fix it and also."
  },
  {
    "start": 1600.38,
    "end": 1601.64,
    "text": "Let's say we're running an important system."
  },
  {
    "start": 1601.7,
    "end": 1603.38,
    "text": "Let's say it's a Tesla autopilot, and it crashes."
  },
  {
    "start": 1604.42,
    "end": 1605.02,
    "text": "Who's liable?"
  },
  {
    "start": 1605.08,
    "end": 1605.78,
    "text": "A human didn't write it."
  },
  {
    "start": 1605.82,
    "end": 1606.98,
    "text": "A human didn't review it."
  },
  {
    "start": 1607.06,
    "end": 1608.12,
    "text": "Someone needs to be an owner there."
  },
  {
    "start": 1608.98,
    "end": 1612.52,
    "text": "So I'm very curious about the context question, the human context question."
  },
  {
    "start": 1612.86,
    "end": 1613.94,
    "text": "Like, we used to get this for free."
  },
  {
    "start": 1613.98,
    "end": 1616.06,
    "text": "When you wrote the code, you got the context."
  },
  {
    "start": 1616.12,
    "end": 1617.22,
    "text": "When you reviewed it, you got the context."
  },
  {
    "start": 1617.68,
    "end": 1620.42,
    "text": "Well, what if you don't write it, and what if maybe you don't review it that closely?"
  },
  {
    "start": 1621.22,
    "end": 1622.62,
    "text": "Do we lose context of what's happening?"
  },
  {
    "start": 1623.02,
    "end": 1624.38,
    "text": "And I need that context for accountability."
  },
  {
    "start": 1624.42,
    "end": 1626.18,
    "text": "I also needed to write the next code change."
  },
  {
    "start": 1626.28,
    "end": 1627.32,
    "text": "I got to know what's running."
  },
  {
    "start": 1628.0,
    "end": 1628.88,
    "text": "to write the next code change."
  },
  {
    "start": 1628.92,
    "end": 1630.04,
    "text": "So something you got to read this code."
  },
  {
    "start": 1630.18,
    "end": 1631.84,
    "text": "I don't have an answer, but I'm very curious."
  },
  {
    "start": 1631.9,
    "end": 1635.26,
    "text": "in like these new problems that are created as we pick up the velocity."
  },
  {
    "start": 1635.72,
    "end": 1641.12,
    "text": "Well, maybe it all has to then point back to some prompt, right?"
  },
  {
    "start": 1642.0,
    "end": 1656.52,
    "text": "Or like someone who initially triggered it, then that person is in a way responsible for his team of agents or her team of agents and for the changes that team of agents made, right?"
  },
  {
    "start": 1657.12,
    "end": 1671.1,
    "text": "Something I'm very interested, I love the idea that you just said, I'm very interested in bringing the providence of code changes, the prompts, the conversations, the models that were used to create that code change into the Git history and into the PRs on a more native level."
  },
  {
    "start": 1671.72,
    "end": 1673.7,
    "text": "Right now it's kind of hidden, right?"
  },
  {
    "start": 1673.76,
    "end": 1676.24,
    "text": "You look at a pull request, you just see the output, you don't really see the input."
  },
  {
    "start": 1676.3,
    "end": 1679.08,
    "text": "Maybe a person wrote some description, but you don't really know how they create that code change."
  },
  {
    "start": 1679.66,
    "end": 1683.4,
    "text": "Maybe it was half AI generated and maybe it was half hand typed, maybe it was tab completed."
  },
  {
    "start": 1683.74,
    "end": 1685.4,
    "text": "Maybe there's so many different ways it could be created."
  },
  {
    "start": 1685.76,
    "end": 1690.68,
    "text": "I think there's value in bringing that into the source control and bringing that into the record keeping."
  },
  {
    "start": 1691.04,
    "end": 1693.56,
    "text": "It creates memory, it creates accountability, creates better understanding."
  },
  {
    "start": 1694.24,
    "end": 1694.74,
    "text": "New problems."
  },
  {
    "start": 1694.92,
    "end": 1695.96,
    "text": "Now my Git repo is a lot bigger."
  },
  {
    "start": 1696.12,
    "end": 1696.76,
    "text": "That's a whole problem."
  },
  {
    "start": 1697.36,
    "end": 1702.22,
    "text": "But I'm really interested in actually creating better standards in that data model and bringing that in."
  },
  {
    "start": 1702.96,
    "end": 1704.38,
    "text": "Creating those traces, right?"
  },
  {
    "start": 1705.32,
    "end": 1707.62,
    "text": "Do you know if there's already approaches towards that?"
  },
  {
    "start": 1708.18,
    "end": 1711.08,
    "text": "If there's already an idea on how to do that?"
  },
  {
    "start": 1711.14,
    "end": 1714.52,
    "text": "Because you don't know who created the code."
  },
  {
    "start": 1715.16,
    "end": 1715.92,
    "text": "Originally, right?"
  },
  {
    "start": 1716.1,
    "end": 1717.26,
    "text": "I mean, I don't know."
  },
  {
    "start": 1717.72,
    "end": 1722.76,
    "text": "Mike Law maybe commits that he was an author as well, but that was it, right?"
  },
  {
    "start": 1723.52,
    "end": 1725.8,
    "text": "There's discussions happening on Twitter right now."
  },
  {
    "start": 1726.46,
    "end": 1727.72,
    "text": "People are really thinking about this."
  },
  {
    "start": 1727.9,
    "end": 1731.8,
    "text": "I'm not aware of any current standard or pattern that's been adopted."
  },
  {
    "start": 1732.34,
    "end": 1735.76,
    "text": "Honestly, if I have time in the next couple of weeks, I would love to push that forward."
  },
  {
    "start": 1735.82,
    "end": 1737.68,
    "text": "Like this is on the plate among many things."
  },
  {
    "start": 1738.2,
    "end": 1738.5,
    "text": "And move"
  },
  {
    "start": 1738.78,
    "end": 1739.42,
    "text": "to the next level."
  },
  {
    "start": 1740.16,
    "end": 1750.18,
    "text": "This is why I'm so excited to be working with the cursor team, because it's such a perfect balance on the outer loop review side and the cursor on the coaching side."
  },
  {
    "start": 1750.78,
    "end": 1756.32,
    "text": "I think working with them, we can help advance some of these discussions and help bring more of these new patterns to the industry."
  },
  {
    "start": 1757.26,
    "end": 1759.2,
    "text": "Yeah, I'm excited about that."
  },
  {
    "start": 1759.62,
    "end": 1768.04,
    "text": "I often had moments not so much with code because for me, the quality of code that the models produce is relatively high these days."
  },
  {
    "start": 1768.84,
    "end": 1776.72,
    "text": "But with other docs, whenever a colleague sends me that long doc with stuff in there where I think, okay, is it human written?"
  },
  {
    "start": 1777.0,
    "end": 1778.84,
    "text": "I always ask for the prompt, right?"
  },
  {
    "start": 1778.96,
    "end": 1784.04,
    "text": "And that's exactly what you would do with code as well or what you would love to do with code as well, right?"
  },
  {
    "start": 1784.56,
    "end": 1785.46,
    "text": "And I love the story."
  },
  {
    "start": 1785.96,
    "end": 1788.06,
    "text": "Personally, I love the story in Git itself."
  },
  {
    "start": 1788.9,
    "end": 1792.26,
    "text": "It's such a fair and open system."
  },
  {
    "start": 1793.12,
    "end": 1797.44,
    "text": "I don't want to lock it within a single company's data mode and gatekeep it."
  },
  {
    "start": 1797.92,
    "end": 1803.7,
    "text": "I think if we can store it in the Git repo itself, then it's portable, then every tool can read and write to it."
  },
  {
    "start": 1804.22,
    "end": 1805.8,
    "text": "It can go into the source control systems."
  },
  {
    "start": 1806.72,
    "end": 1815.98,
    "text": "I think we need to agree on a standard that people can read and write to this, and we need to be respectful that it might create a ton of bloat in the Git repos, but I think it's the right home for this to live in."
  },
  {
    "start": 1817.12,
    "end": 1817.54,
    "text": "Absolutely."
  },
  {
    "start": 1817.84,
    "end": 1819.46,
    "text": "I think so too."
  },
  {
    "start": 1820.26,
    "end": 1824.42,
    "text": "Can you already say, since the acquisition, what changed for you or what will change for you?"
  },
  {
    "start": 1825.34,
    "end": 1827.36,
    "text": "It just closed literally a few weeks ago."
  },
  {
    "start": 1829.38,
    "end": 1831.72,
    "text": "Do you already know how your future will look like?"
  },
  {
    "start": 1832.04,
    "end": 1839.54,
    "text": "You will reinvent a lot of the way we're working or many companies work, let's say at least."
  },
  {
    "start": 1841.46,
    "end": 1842.18,
    "text": "What changed?"
  },
  {
    "start": 1842.78,
    "end": 1843.82,
    "text": "In the short term, not a lot."
  },
  {
    "start": 1844.16,
    "end": 1847.46,
    "text": "In the medium term, More and more and in long term will fully mesh."
  },
  {
    "start": 1847.56,
    "end": 1853.36,
    "text": "but the way we're choosing to run it when I say we I mean graphite founders and the cursor founders the way we're choosing to execute this."
  },
  {
    "start": 1854.72,
    "end": 1855.9,
    "text": "We have our graphite office."
  },
  {
    "start": 1855.96,
    "end": 1864.98,
    "text": "It's about eighty teammates and I mean cursor is growing really really fast But as of a month or so ago, they were I don't know around two hundred plus teammates."
  },
  {
    "start": 1865.62,
    "end": 1866.96,
    "text": "So we actually make up a good chunk."
  },
  {
    "start": 1867.1,
    "end": 1869.94,
    "text": "We're like a quarter a third of the team here."
  },
  {
    "start": 1872.06,
    "end": 1873.24,
    "text": "We also are mostly New York based."
  },
  {
    "start": 1873.62,
    "end": 1874.44,
    "text": "We got this great office."
  },
  {
    "start": 1874.48,
    "end": 1875.66,
    "text": "We're moving into a bigger office."
  },
  {
    "start": 1875.98,
    "end": 1877.36,
    "text": "Cursor by chance is next door."
  },
  {
    "start": 1877.64,
    "end": 1878.82,
    "text": "They're in New York office next door."
  },
  {
    "start": 1878.86,
    "end": 1879.74,
    "text": "It's a little smaller than ours."
  },
  {
    "start": 1879.9,
    "end": 1880.44,
    "text": "It's next door."
  },
  {
    "start": 1881.2,
    "end": 1882.5,
    "text": "So that's how you match us."
  },
  {
    "start": 1884.44,
    "end": 1885.62,
    "text": "We lose so many crossovers."
  },
  {
    "start": 1885.74,
    "end": 1887.72,
    "text": "You know, we have our own infrastructure."
  },
  {
    "start": 1887.76,
    "end": 1890.94,
    "text": "We had an AWS account and we have an info team and we have our own servers."
  },
  {
    "start": 1891.04,
    "end": 1894.66,
    "text": "and we have some great patterns around engineering, product and design collaboration."
  },
  {
    "start": 1895.16,
    "end": 1896.52,
    "text": "And we're really good at some stuff."
  },
  {
    "start": 1896.58,
    "end": 1901.56,
    "text": "We're really good at interfacing with GitHub and thinking about Git and code review and online UX experiences."
  },
  {
    "start": 1901.9,
    "end": 1903.08,
    "text": "Cursor is really good at some stuff too."
  },
  {
    "start": 1903.12,
    "end": 1903.86,
    "text": "They're really good at local."
  },
  {
    "start": 1903.9,
    "end": 1907.3,
    "text": "They're really good at AI research that we have no real talent in."
  },
  {
    "start": 1909.42,
    "end": 1914.48,
    "text": "We could smush everything together overnight and probably brick our productivity."
  },
  {
    "start": 1915.06,
    "end": 1916.58,
    "text": "and just race towards integration."
  },
  {
    "start": 1917.32,
    "end": 1919.12,
    "text": "But there's not a lot of need to."
  },
  {
    "start": 1920.7,
    "end": 1923.78,
    "text": "Cursor, one thing I love about working with them, they're not trying to throw away our roadmap."
  },
  {
    "start": 1923.86,
    "end": 1926.18,
    "text": "They're not like, great, we're so glad we brought you on now."
  },
  {
    "start": 1926.76,
    "end": 1929.62,
    "text": "Now, delete everything you've created and just come work on our projects."
  },
  {
    "start": 1929.68,
    "end": 1930.92,
    "text": "It's actually very different."
  },
  {
    "start": 1930.96,
    "end": 1932.74,
    "text": "They're like, we love what you're creating."
  },
  {
    "start": 1932.94,
    "end": 1933.52,
    "text": "We love Graphite."
  },
  {
    "start": 1933.6,
    "end": 1935.72,
    "text": "We love this feature of Code Review you're thinking about."
  },
  {
    "start": 1935.98,
    "end": 1936.98,
    "text": "Please keep cooking on that."
  },
  {
    "start": 1937.48,
    "end": 1943.04,
    "text": "And yeah, of course, let's have some integrations and some buttons that navigate between Cursor and Graphite and back and forth."
  },
  {
    "start": 1943.78,
    "end": 1948.72,
    "text": "And of course, it's like, you know, graphite is going to be a product within cursor, but it's kind of like Instagram to Facebook."
  },
  {
    "start": 1948.8,
    "end": 1954.92,
    "text": "I think we're going to keep running with our own agency and patterns for a little bit and take inspiration for each other and grow closer together over time."
  },
  {
    "start": 1955.6,
    "end": 1956.68,
    "text": "Not a lot is changing day to day."
  },
  {
    "start": 1956.84,
    "end": 1959.3,
    "text": "I still wake up and go to the same office to go work with the same team."
  },
  {
    "start": 1959.4,
    "end": 1960.34,
    "text": "I work on the same code base."
  },
  {
    "start": 1960.42,
    "end": 1961.5,
    "text": "We're still evolving all of that."
  },
  {
    "start": 1962.94,
    "end": 1967.96,
    "text": "The most fun part about this is the cross-collaboration of knowledge share."
  },
  {
    "start": 1968.04,
    "end": 1975.32,
    "text": "Cursor has some really brilliant teammates, and we're getting coffees, and we're setting up meetings, and we're just trading notes on how we build these things."
  },
  {
    "start": 1976.5,
    "end": 1977.3,
    "text": "It's so wonderful."
  },
  {
    "start": 1977.36,
    "end": 1977.82,
    "text": "It's such a dream."
  },
  {
    "start": 1978.36,
    "end": 1980.8,
    "text": "Nothing's more fun than working with really smart people and learning from them."
  },
  {
    "start": 1981.44,
    "end": 1983.22,
    "text": "And that door has opened massively."
  },
  {
    "start": 1985.12,
    "end": 1985.84,
    "text": "That's an experience so far."
  },
  {
    "start": 1986.74,
    "end": 1988.7,
    "text": "Come ask me again in six months of how this develops."
  },
  {
    "start": 1988.8,
    "end": 1990.74,
    "text": "Obviously, we're weeks into it."
  },
  {
    "start": 1991.58,
    "end": 1992.54,
    "text": "I think we're picking the pace."
  },
  {
    "start": 1994.14,
    "end": 1996.2,
    "text": "How does a success metric for that look like?"
  },
  {
    "start": 1996.92,
    "end": 1998.02,
    "text": "Or for the future look like?"
  },
  {
    "start": 1998.32,
    "end": 1999.02,
    "text": "How do you ensure?"
  },
  {
    "start": 1999.64,
    "end": 2003.08,
    "text": "How do you think about stickiness of your product?"
  },
  {
    "start": 2003.16,
    "end": 2004.94,
    "text": "I mean, you change quite a lot, right?"
  },
  {
    "start": 2006.46,
    "end": 2007.92,
    "text": "You reinvented yourself quite a lot."
  },
  {
    "start": 2008.42,
    "end": 2010.22,
    "text": "How does this look like in the future?"
  },
  {
    "start": 2010.96,
    "end": 2011.9,
    "text": "What are your thoughts there?"
  },
  {
    "start": 2012.56,
    "end": 2023.82,
    "text": "How do you ensure that people love your tool and maybe even... respect that some move to a different model or to different CLI."
  },
  {
    "start": 2025.06,
    "end": 2031.92,
    "text": "Is that baked into the transaction as well or are you purely focusing or will you purely focus on Cursor in the future?"
  },
  {
    "start": 2032.2,
    "end": 2033.04,
    "text": "What are your thoughts there?"
  },
  {
    "start": 2033.68,
    "end": 2037.68,
    "text": "I think we're going to keep focusing on, what I say."
  },
  {
    "start": 2037.78,
    "end": 2044.0,
    "text": "we, I mean, the graphite team as it is, we're going to keep focusing on the outer loop of software development, which is the same thing we were focusing on."
  },
  {
    "start": 2044.36,
    "end": 2047.06,
    "text": "When I say outer loop, I think about the creation of PRs."
  },
  {
    "start": 2047.54,
    "end": 2049.639,
    "text": "the review and testing them and the merging in of them."
  },
  {
    "start": 2050.6,
    "end": 2057.88,
    "text": "Pretty much everything that happens after you hit, you know, save in cursor or cloud code or any of these tools, like what happens after you save that code edit?"
  },
  {
    "start": 2057.94,
    "end": 2058.739,
    "text": "What do you go do with that?"
  },
  {
    "start": 2058.86,
    "end": 2061.239,
    "text": "If you write, you know, you say commit and then what happens?"
  },
  {
    "start": 2061.42,
    "end": 2063.56,
    "text": "It doesn't just automatically, you know, your job's not done."
  },
  {
    "start": 2064.8,
    "end": 2066.139,
    "text": "That was our obsession six months ago."
  },
  {
    "start": 2066.26,
    "end": 2068.08,
    "text": "That's going to be our obsession this year and next year."
  },
  {
    "start": 2068.699,
    "end": 2069.739,
    "text": "How do we measure success on?"
  },
  {
    "start": 2069.8,
    "end": 2071.02,
    "text": "that is one of your questions."
  },
  {
    "start": 2071.52,
    "end": 2073.159,
    "text": "Well, luckily we've been measuring success on this."
  },
  {
    "start": 2073.92,
    "end": 2074.46,
    "text": "For years now."
  },
  {
    "start": 2074.8,
    "end": 2076.139,
    "text": "we look at you know user retention."
  },
  {
    "start": 2076.5,
    "end": 2077.139,
    "text": "We have one of."
  },
  {
    "start": 2077.219,
    "end": 2086.699,
    "text": "we have I have over a hundred slack channels with these amazing engineering teams Robin Hood and data dog and for sell and all these cool companies who use graphite every day."
  },
  {
    "start": 2087.639,
    "end": 2088.699,
    "text": "And again, we talked to them."
  },
  {
    "start": 2088.739,
    "end": 2089.46,
    "text": "We're gonna keep talking to them."
  },
  {
    "start": 2089.54,
    "end": 2093.34,
    "text": "We're gonna keep saying hi and visiting them and figuring out the roadmaps how to build them a really great product."
  },
  {
    "start": 2093.84,
    "end": 2094.6,
    "text": "We have a marketing team."
  },
  {
    "start": 2094.659,
    "end": 2096.659,
    "text": "We have a sales and we're gonna keep distributing this out to people."
  },
  {
    "start": 2097.139,
    "end": 2104.32,
    "text": "We're gonna make sure that you know in the numbers and in the in the qualitative data that people still really like what we're building."
  },
  {
    "start": 2104.36,
    "end": 2111.26,
    "text": "And we're going to keep trying to ship some really interesting innovations and keep trying to make it faster and smoother and better and adopt these AI features."
  },
  {
    "start": 2112.22,
    "end": 2112.54,
    "text": "I don't know."
  },
  {
    "start": 2113.0,
    "end": 2132.14,
    "text": "The one thing that changes a little bit, not entirely, but a little bit is like, you know, when you're operating as an independent startup, you got to grow really fast and you got to hit these killer revenue milestones and you have to balance long-term Healthy choices with really short-term gains to keep up with momentum to keep funding the company."
  },
  {
    "start": 2133.32,
    "end": 2135.7,
    "text": "What's really nice is a little bit that pressure is off our backs."
  },
  {
    "start": 2136.68,
    "end": 2142.46,
    "text": "I'm gonna you know, we're gonna keep making a product that happens to make good revenue But I don't have to hit some target this year or die."
  },
  {
    "start": 2142.92,
    "end": 2144.1,
    "text": "You know and that's always."
  },
  {
    "start": 2144.3,
    "end": 2145.5,
    "text": "you're always default dead as a startup."
  },
  {
    "start": 2145.54,
    "end": 2147.46,
    "text": "You gotta you gotta keep just running it super intensely."
  },
  {
    "start": 2148.86,
    "end": 2153.32,
    "text": "I get a little bit more affordance now to just build something that's really fantastic for users."
  },
  {
    "start": 2153.68,
    "end": 2156.46,
    "text": "Graphite's revenue is great, but it pales in comparison to Cursor."
  },
  {
    "start": 2156.8,
    "end": 2159.28,
    "text": "Almost every company's revenue pales in comparison to Cursor."
  },
  {
    "start": 2160.1,
    "end": 2166.7,
    "text": "We're here to really build something that's engaging and sticky and useful to users, more than we are to drive Cursor's bottom line revenue."
  },
  {
    "start": 2166.96,
    "end": 2167.78,
    "text": "And that's a fun mandate."
  },
  {
    "start": 2168.08,
    "end": 2173.68,
    "text": "That aligns more with where my heart is at, which is I'm here to build something fantastic for the whole industry."
  },
  {
    "start": 2173.98,
    "end": 2174.06,
    "text": "And"
  },
  {
    "start": 2174.14,
    "end": 2178.0,
    "text": "what would you say, tech-wise, if you really look at... the technology."
  },
  {
    "start": 2179.52,
    "end": 2184.42,
    "text": "How important is it these days to really have something?"
  },
  {
    "start": 2184.78,
    "end": 2188.14,
    "text": "or how does that look like for you to really have a mode and tech?"
  },
  {
    "start": 2189.06,
    "end": 2189.68,
    "text": "Do you have that?"
  },
  {
    "start": 2189.74,
    "end": 2190.72,
    "text": "Would you say you have that?"
  },
  {
    "start": 2190.84,
    "end": 2195.06,
    "text": "Or is it more the general creation, the total creation of things?"
  },
  {
    "start": 2196.16,
    "end": 2199.8,
    "text": "I'm thinking a lot about that because I'm also looking into SAS quite a lot."
  },
  {
    "start": 2201.98,
    "end": 2206.96,
    "text": "What is really the outstanding thing is a distribution and how will that look like in the future?"
  },
  {
    "start": 2207.0,
    "end": 2211.74,
    "text": "Like is a distribution of software, which really is the secret source?"
  },
  {
    "start": 2212.8,
    "end": 2214.34,
    "text": "Like how do users discover you?"
  },
  {
    "start": 2214.6,
    "end": 2218.5,
    "text": "Like, I mean, you had that product launch, which was quite successful as an example."
  },
  {
    "start": 2218.58,
    "end": 2222.14,
    "text": "And then, like, it kind of accelerated from that on."
  },
  {
    "start": 2223.24,
    "end": 2228.18,
    "text": "Is that the secret behind, like, ninety percent of the SaaS companies in the future?"
  },
  {
    "start": 2228.84,
    "end": 2236.04,
    "text": "Or do you think that the software in itself will still be a secret source to many?"
  },
  {
    "start": 2236.76,
    "end": 2237.68,
    "text": "And how is it for you?"
  },
  {
    "start": 2238.68,
    "end": 2239.24,
    "text": "What is the moat?"
  },
  {
    "start": 2239.56,
    "end": 2242.06,
    "text": "You know, it's kind of a question for us, for the companies around us."
  },
  {
    "start": 2243.58,
    "end": 2253.38,
    "text": "I think about the mental framework of Hamilton-Helmer's Seven Powers, where the powers and dynamics that can create sustainable advantage for businesses over a long time."
  },
  {
    "start": 2254.1,
    "end": 2258.1,
    "text": "These are things like network effects, economies of scale, brand."
  },
  {
    "start": 2260.16,
    "end": 2269.82,
    "text": "I think when it comes to graphite, we have slightly better mode and sustainability than the average AI startup."
  },
  {
    "start": 2270.38,
    "end": 2278.7,
    "text": "That doesn't make us better than the average AI startup, but it's interesting because we were founded pre-AI, so a lot of the advantages we had to build and serve to users."
  },
  {
    "start": 2281.5,
    "end": 2283.64,
    "text": "Couldn't just be like, you know, we slept in LL Amonto."
  },
  {
    "start": 2283.68,
    "end": 2286.26,
    "text": "We had to build, we actually had to build really complicated."
  },
  {
    "start": 2287.14,
    "end": 2292.92,
    "text": "Some of the reasons that I think Graphite's quite successful, we built extremely complicated syncing engines to work on top of GitHub."
  },
  {
    "start": 2292.98,
    "end": 2302.3,
    "text": "I said earlier in this conversation that Graphite is the only tool for code review, for wide scale code collaboration that's ever been invented that works perfectly on top of GitHub."
  },
  {
    "start": 2302.86,
    "end": 2305.5,
    "text": "And it's not, you know, you've got to migrate off."
  },
  {
    "start": 2306.0,
    "end": 2306.76,
    "text": "That's actually really hard."
  },
  {
    "start": 2307.06,
    "end": 2309.46,
    "text": "We didn't know going in, but that like killed us for like three years."
  },
  {
    "start": 2310.0,
    "end": 2325.14,
    "text": "And every startup who tries to build a graphic clone, and all these really smart companies like Anthropic and Cursor who try to build extensions on how to get up, immediately face challenges on rate limits and reading and writing and mutating the data and keeping in sync with it."
  },
  {
    "start": 2325.94,
    "end": 2328.04,
    "text": "We accidentally have a little bit of an engineering mode there."
  },
  {
    "start": 2328.6,
    "end": 2330.68,
    "text": "Engineering modes are never impenetrable."
  },
  {
    "start": 2330.74,
    "end": 2335.02,
    "text": "They're just a advantage and lead time, but that one has been surprisingly sustainable for us."
  },
  {
    "start": 2337.02,
    "end": 2345.9,
    "text": "The other thing is we've accumulated some good brand and trust in the industry by being around for five plus years and being kind and hopefully doing good by all our users."
  },
  {
    "start": 2346.46,
    "end": 2351.32,
    "text": "People continue to trust us more and more with aspects of their workflow and code bases."
  },
  {
    "start": 2351.46,
    "end": 2353.12,
    "text": "more than a brand new startup they've never heard of."
  },
  {
    "start": 2353.16,
    "end": 2356.12,
    "text": "I also think this is an advantage of take cursor and cloud code."
  },
  {
    "start": 2356.4,
    "end": 2361.2,
    "text": "Why do people use cursor or some of these alternatives instead of one of the forty other code gen tools?"
  },
  {
    "start": 2361.7,
    "end": 2362.48,
    "text": "There's some brand effect."
  },
  {
    "start": 2362.96,
    "end": 2365.5,
    "text": "There's some like, you know the name and you keep using it."
  },
  {
    "start": 2365.56,
    "end": 2371.24,
    "text": "That is a true power and the longer a company stays around and it keeps refining them, opening eyes to the extreme version of this."
  },
  {
    "start": 2371.32,
    "end": 2372.82,
    "text": "ChatTBT is an ultimate brand, right?"
  },
  {
    "start": 2373.38,
    "end": 2376.2,
    "text": "There's some accumulated power that goes within that."
  },
  {
    "start": 2376.24,
    "end": 2387.54,
    "text": "But the last thing I'll say, and I believe this would be true, as I look around the AI DevTools landscape, twenty-five, twenty-six, there's a phrase I hear on refrain, which is that speed is the moat."
  },
  {
    "start": 2389.44,
    "end": 2404.3,
    "text": "When things are changing this fast and when everyone can just plug text into an LLM, one of the only advantages you have as a team and as a company is your ability to move really fast and ship features to users faster than the next company can."
  },
  {
    "start": 2404.86,
    "end": 2408.64,
    "text": "Can you maintain an incredible pace for a sustained period of time?"
  },
  {
    "start": 2410.16,
    "end": 2410.78,
    "text": "Am I a cursor?"
  },
  {
    "start": 2410.9,
    "end": 2414.02,
    "text": "I think they do this in droves for years now."
  },
  {
    "start": 2414.6,
    "end": 2418.86,
    "text": "They've been launching crazy new features all the time, reinventing themselves, staying ahead of the competition."
  },
  {
    "start": 2419.08,
    "end": 2422.46,
    "text": "They were born into this crazy pace, right?"
  },
  {
    "start": 2423.52,
    "end": 2428.34,
    "text": "Or kind of had their inception moment, and then you kind of get used to that, I guess."
  },
  {
    "start": 2429.26,
    "end": 2435.72,
    "text": "But on the other hand, you also have to really get used to huge setbacks, right?"
  },
  {
    "start": 2435.98,
    "end": 2448.88,
    "text": "I mean, Cursor as an example, I use Cursor every once in a while, but I had my big Cursor moment, and then All of a sudden someone told me about Claude at lunchtime and then I was always Claude."
  },
  {
    "start": 2448.98,
    "end": 2451.02,
    "text": "And I think that's also a new thing, right?"
  },
  {
    "start": 2451.06,
    "end": 2453.3,
    "text": "That people change tools quite a lot."
  },
  {
    "start": 2453.72,
    "end": 2458.84,
    "text": "I see, I know many people that moved from Chatchapiti to Gemini now with the last release."
  },
  {
    "start": 2460.12,
    "end": 2469.12,
    "text": "And I think that's also an interesting reality that was a bit, yeah, was a bit unknown before, right?"
  },
  {
    "start": 2469.16,
    "end": 2478.26,
    "text": "Like you stick to your tools earlier and I thought like no one is going to reinvent that so quickly and then you just have then the new pile."
  },
  {
    "start": 2478.46,
    "end": 2478.8,
    "text": "I think about"
  },
  {
    "start": 2478.84,
    "end": 2480.36,
    "text": "this on two layers."
  },
  {
    "start": 2480.96,
    "end": 2483.24,
    "text": "So one, I think about graphite, I think about this for cursor."
  },
  {
    "start": 2483.86,
    "end": 2487.22,
    "text": "Graphite, we're a little lucky that there's not a lot of tools."
  },
  {
    "start": 2487.5,
    "end": 2491.8,
    "text": "If you like what we have, if you like inboxes and stack diffs and air code review."
  },
  {
    "start": 2492.24,
    "end": 2493.7,
    "text": "There's not a lot of tools to swap between."
  },
  {
    "start": 2493.9,
    "end": 2495.34,
    "text": "We're just lucky that we don't have a ton of competitors."
  },
  {
    "start": 2495.44,
    "end": 2501.82,
    "text": "I think the only reason we don't have a ton of competitors is it's really hard to build like a GitHub syncing engine and a complex like PR page."
  },
  {
    "start": 2502.18,
    "end": 2503.08,
    "text": "I think it's just really hard to build."
  },
  {
    "start": 2504.1,
    "end": 2510.46,
    "text": "With Windsurf and Cursor and something in Cloud Code, you could fork VS Code or you could just write a terminal application."
  },
  {
    "start": 2510.56,
    "end": 2512.14,
    "text": "There's a famous, right now."
  },
  {
    "start": 2512.58,
    "end": 2518.6,
    "text": "in last week on HN, there was a really big post that blew up which said like, hey, Cloud Code, you can kind of create a Cloud Code in like two hundred lines."
  },
  {
    "start": 2519.34,
    "end": 2521.48,
    "text": "This is why it's very easy to create these and swap between them."
  },
  {
    "start": 2521.76,
    "end": 2522.62,
    "text": "We're very lucky in graphite."
  },
  {
    "start": 2522.66,
    "end": 2526.68,
    "text": "that's just to date, maybe in the future, but to date there isn't a tool you can easily swap off onto another one."
  },
  {
    "start": 2526.92,
    "end": 2529.4,
    "text": "Not that we have a data lock and there just isn't, someone hasn't built an alternative."
  },
  {
    "start": 2530.5,
    "end": 2539.28,
    "text": "Cursor is very interesting to me because they are pretty much the only mainstream code gen tool that is relatively agnostic to the models."
  },
  {
    "start": 2539.32,
    "end": 2540.52,
    "text": "And I think this is an incredible power."
  },
  {
    "start": 2541.16,
    "end": 2543.24,
    "text": "Cloud code is great when sauna is great."
  },
  {
    "start": 2543.68,
    "end": 2547.72,
    "text": "Gemini is great when the Gemini model is great, the Gemini CLI, Codex and so on."
  },
  {
    "start": 2548.32,
    "end": 2557.62,
    "text": "The nice thing about Cursor is if you can set it up and then if each month, if a new model takes the lead, you just like pick the other model in the drop down and you keep cooking, but you can use a consistent set of tools."
  },
  {
    "start": 2558.1,
    "end": 2571.46,
    "text": "And as they layer improvements onto that harness, as they create better ways, you know, they launched kind of a Figma-esque way to interact with web pages and they have ways where you can run one prompt and actually fan out to multiple sub agents using different different models."
  },
  {
    "start": 2571.5,
    "end": 2573.2,
    "text": "They're agnostic to be underlying."
  },
  {
    "start": 2574.22,
    "end": 2583.12,
    "text": "In a world where these model companies are just taking the lead left and right and are, you know, are struggling to find differentiation, I actually think that puts Kershner in a really strong position over time."
  },
  {
    "start": 2583.36,
    "end": 2597.86,
    "text": "I think last year, sorry to some engineers that are like, there were seven different moments where different models technically took the lead on benchmarks and, you know, If you want to stay on the bleeding edge, pick a tool that can instantly and transparently swap between all these."
  },
  {
    "start": 2597.9,
    "end": 2598.42,
    "text": "They're not beholden."
  },
  {
    "start": 2598.6,
    "end": 2601.12,
    "text": "Directly swaps to the newest and hottest one."
  },
  {
    "start": 2603.36,
    "end": 2605.1,
    "text": "Claudeco does not appear to run Gemini models."
  },
  {
    "start": 2605.38,
    "end": 2607.26,
    "text": "I'm curious if they make that change, but for today."
  },
  {
    "start": 2607.3,
    "end": 2608.28,
    "text": "they're going to lock you into a model."
  },
  {
    "start": 2608.68,
    "end": 2613.36,
    "text": "So before we come to the end, quite interesting discussion."
  },
  {
    "start": 2614.48,
    "end": 2620.6,
    "text": "If you could jump like three years forward, what do you think is the default day of an engineer?"
  },
  {
    "start": 2620.66,
    "end": 2624.34,
    "text": "Is it writing codes, reviewing agents or designing policies?"
  },
  {
    "start": 2624.86,
    "end": 2632.24,
    "text": "Writing code, reviewing agents and policies, working with the agents, working on policies."
  },
  {
    "start": 2634.36,
    "end": 2635.34,
    "text": "I don't think you're wearing a lot of code."
  },
  {
    "start": 2635.74,
    "end": 2636.12,
    "text": "I really don't."
  },
  {
    "start": 2636.44,
    "end": 2637.68,
    "text": "I think you're going to prompt a lot of stuff."
  },
  {
    "start": 2637.82,
    "end": 2638.96,
    "text": "I don't think you're going to write a lot of code."
  },
  {
    "start": 2640.42,
    "end": 2641.02,
    "text": "Borderline zero."
  },
  {
    "start": 2641.26,
    "end": 2645.48,
    "text": "You might make a targeted edit, but even those targeted edits, I think you're just going to write a prompt."
  },
  {
    "start": 2645.6,
    "end": 2647.02,
    "text": "I think it's going to go close to zero."
  },
  {
    "start": 2648.2,
    "end": 2648.66,
    "text": "Policies?"
  },
  {
    "start": 2649.04,
    "end": 2650.08,
    "text": "Yes, you'll keep doing it."
  },
  {
    "start": 2650.22,
    "end": 2651.7,
    "text": "I think you're going to walk into an engineering meeting."
  },
  {
    "start": 2651.76,
    "end": 2654.04,
    "text": "You're going to go with a couple of humans at a whiteboard in person."
  },
  {
    "start": 2654.08,
    "end": 2656.56,
    "text": "You're going to decide, hey, we should have this different policy in our code base."
  },
  {
    "start": 2656.6,
    "end": 2658.94,
    "text": "You're going to shake hands on it and you're going to save that to the system."
  },
  {
    "start": 2659.94,
    "end": 2661.38,
    "text": "And yeah, you're going to delegate to a lot of agents."
  },
  {
    "start": 2661.94,
    "end": 2666.48,
    "text": "And I don't think that humans are going to stop reading the code and reviewing it before it goes in production."
  },
  {
    "start": 2666.86,
    "end": 2667.38,
    "text": "There's some real"
  },
  {
    "start": 2667.46,
    "end": 2667.98,
    "text": "issues, right?"
  },
  {
    "start": 2670.2,
    "end": 2673.28,
    "text": "No, just referring back to your bias, right?"
  },
  {
    "start": 2674.22,
    "end": 2675.94,
    "text": "Yeah, my mindsets are really mixed."
  },
  {
    "start": 2676.62,
    "end": 2685.6,
    "text": "But yeah, I hope for the industry, as a human, I hope we live in a world where my bank and my car, some human, read the code that was running on it."
  },
  {
    "start": 2685.82,
    "end": 2687.78,
    "text": "I hope that never remains true as a user."
  },
  {
    "start": 2689.26,
    "end": 2691.44,
    "text": "But also, like, you know, there's also systematic forces here."
  },
  {
    "start": 2692.2,
    "end": 2695.0,
    "text": "Most companies are like SOC-II compliant or further or federal stuff."
  },
  {
    "start": 2695.4,
    "end": 2700.52,
    "text": "Those mandate human code review, like, or some form of code review, I think to some degree, you know, human auditability."
  },
  {
    "start": 2701.14,
    "end": 2707.38,
    "text": "There's, you know, there's levels of security and regulation and important practices."
  },
  {
    "start": 2707.44,
    "end": 2709.62,
    "text": "I don't think it's going away overnight."
  },
  {
    "start": 2709.88,
    "end": 2711.02,
    "text": "There's a lot of legacy companies."
  },
  {
    "start": 2711.1,
    "end": 2712.08,
    "text": "They're not going to stop reading."
  },
  {
    "start": 2712.5,
    "end": 2713.48,
    "text": "the code that's going through systems."
  },
  {
    "start": 2713.58,
    "end": 2719.36,
    "text": "Yes, maybe the most edgy bleeding edge startup in AI will delegate it all to AI, but I think human code reviews stay around."
  },
  {
    "start": 2719.42,
    "end": 2721.34,
    "text": "So US question, will humans keep reading code?"
  },
  {
    "start": 2721.38,
    "end": 2721.54,
    "text": "Yes."
  },
  {
    "start": 2721.6,
    "end": 2722.54,
    "text": "Will they keep writing it?"
  },
  {
    "start": 2722.58,
    "end": 2723.1,
    "text": "Probably not."
  },
  {
    "start": 2723.16,
    "end": 2723.76,
    "text": "That'll fall over."
  },
  {
    "start": 2724.14,
    "end": 2726.08,
    "text": "Will they delegate a lot of sub-agents for the generation?"
  },
  {
    "start": 2726.22,
    "end": 2726.56,
    "text": "I think so."
  },
  {
    "start": 2727.18,
    "end": 2728.0,
    "text": "Will they write policies?"
  },
  {
    "start": 2728.42,
    "end": 2728.84,
    "text": "For sure."
  },
  {
    "start": 2729.56,
    "end": 2733.94,
    "text": "The other thing, are a bunch of smart engineers going to come into a room each workday?"
  },
  {
    "start": 2734.48,
    "end": 2738.4,
    "text": "and become friends and eat lunch together and debate what they should be creating and how to fix a problem."
  },
  {
    "start": 2738.56,
    "end": 2739.32,
    "text": "Yeah, that's not going away."
  },
  {
    "start": 2739.8,
    "end": 2740.92,
    "text": "I think it's a great time to be an engineer."
  },
  {
    "start": 2741.22,
    "end": 2744.32,
    "text": "I think it's a great time to be a friendly human who's really good at collaborating with other people."
  },
  {
    "start": 2746.4,
    "end": 2748.34,
    "text": "That part of the job is... Which"
  },
  {
    "start": 2748.48,
    "end": 2749.72,
    "text": "not many engineers are good"
  },
  {
    "start": 2753.46,
    "end": 2753.54,
    "text": "at."
  },
  {
    "start": 2753.72,
    "end": 2758.46,
    "text": "I read these on Twitter and they're like, companies will be run by one engineer in the future."
  },
  {
    "start": 2759.1,
    "end": 2759.46,
    "text": "Bullshit."
  },
  {
    "start": 2761.56,
    "end": 2762.66,
    "text": "Well, maybe some, right?"
  },
  {
    "start": 2762.76,
    "end": 2765.72,
    "text": "Like a small bootstrap company, it hopefully still exists."
  },
  {
    "start": 2766.28,
    "end": 2766.96,
    "text": "I hope at least."
  },
  {
    "start": 2767.06,
    "end": 2767.12,
    "text": "Yes,"
  },
  {
    "start": 2767.56,
    "end": 2767.92,
    "text": "of course."
  },
  {
    "start": 2768.0,
    "end": 2768.7,
    "text": "Of course you have small"
  },
  {
    "start": 2768.74,
    "end": 2769.74,
    "text": "and accomplished."
  },
  {
    "start": 2769.98,
    "end": 2772.72,
    "text": "But yeah, there won't be an Airbnb run by a single engineer."
  },
  {
    "start": 2773.68,
    "end": 2774.66,
    "text": "I hope so."
  },
  {
    "start": 2774.7,
    "end": 2776.5,
    "text": "I don't"
  },
  {
    "start": 2776.68,
    "end": 2777.74,
    "text": "see it."
  },
  {
    "start": 2777.78,
    "end": 2779.78,
    "text": "What will look ridiculous in twenty twenty nine?"
  },
  {
    "start": 2779.82,
    "end": 2787.36,
    "text": "Is it manual repasses, PR templates, CI CD, configured by YAML files, or copy and paste debugging?"
  },
  {
    "start": 2787.68,
    "end": 2789.04,
    "text": "Or all of the above."
  },
  {
    "start": 2789.08,
    "end": 2789.96,
    "text": "Twenty"
  },
  {
    "start": 2790.16,
    "end": 2790.5,
    "text": "twenty nine."
  },
  {
    "start": 2791.3,
    "end": 2792.1,
    "text": "is three years away."
  },
  {
    "start": 2792.62,
    "end": 2798.16,
    "text": "According to the website in paper twenty-twenty-seven, this is post the singularity and we've all died or something."
  },
  {
    "start": 2799.2,
    "end": 2799.66,
    "text": "I don't believe that."
  },
  {
    "start": 2799.7,
    "end": 2801.38,
    "text": "That's just what people say in the Turkish fear."
  },
  {
    "start": 2801.88,
    "end": 2805.28,
    "text": "But I'll give you one hot take here."
  },
  {
    "start": 2805.58,
    "end": 2806.06,
    "text": "I mean, we can make"
  },
  {
    "start": 2806.26,
    "end": 2807.06,
    "text": "lots of modifications."
  },
  {
    "start": 2807.22,
    "end": 2808.14,
    "text": "One hot take I have."
  },
  {
    "start": 2808.64,
    "end": 2810.6,
    "text": "I'm really interested to see how CI evolves."
  },
  {
    "start": 2810.9,
    "end": 2812.84,
    "text": "This is under discussed, in my opinion."
  },
  {
    "start": 2813.54,
    "end": 2815.88,
    "text": "Right now, the world of CI hasn't changed that much."
  },
  {
    "start": 2818.04,
    "end": 2826.78,
    "text": "The way it works, you spin up a sandbox, you clone the code to it, you execute your unit tests, you look at the result of those deterministic unit tests, and then you put it back to the PR and you give it a check mark or not."
  },
  {
    "start": 2827.56,
    "end": 2832.92,
    "text": "What's changed a little bit is that now most, a lot of those unit tests are being agenerated, you know, auto-completed for you."
  },
  {
    "start": 2832.96,
    "end": 2833.34,
    "text": "That's great."
  },
  {
    "start": 2833.4,
    "end": 2837.0,
    "text": "Humans like hate writing tests, so this is great for the world of like test coverage."
  },
  {
    "start": 2837.04,
    "end": 2839.52,
    "text": "But I think that it's gonna go further."
  },
  {
    "start": 2840.18,
    "end": 2854.06,
    "text": "My current guess of what this will look like in a year or two years or three years, like you're asking, is that a sandbox will be spun up, the code will be cloned to it, and then a terminal agent will just kick off, and it'll have a prompt."
  },
  {
    "start": 2854.1,
    "end": 2855.56,
    "text": "It'll say, hey, please validate this code change."
  },
  {
    "start": 2856.42,
    "end": 2860.22,
    "text": "And it'll act not like a unit test, but it'll act like you, Toby, my teammate."
  },
  {
    "start": 2860.62,
    "end": 2862.92,
    "text": "If I texted you and say, validate my code change, what would you do?"
  },
  {
    "start": 2863.22,
    "end": 2863.58,
    "text": "You clone it?"
  },
  {
    "start": 2863.66,
    "end": 2864.5,
    "text": "Yeah, you'd run the unit test."
  },
  {
    "start": 2864.82,
    "end": 2865.48,
    "text": "You'd also boot it up."
  },
  {
    "start": 2865.78,
    "end": 2866.52,
    "text": "You'd also click around."
  },
  {
    "start": 2866.6,
    "end": 2868.44,
    "text": "You'd make sure it works if you had a lot of time."
  },
  {
    "start": 2868.7,
    "end": 2870.64,
    "text": "You would really double check the code."
  },
  {
    "start": 2871.16,
    "end": 2871.88,
    "text": "You really think about it."
  },
  {
    "start": 2871.92,
    "end": 2872.8,
    "text": "You might add some tests."
  },
  {
    "start": 2873.02,
    "end": 2874.06,
    "text": "You might extend it a little bit."
  },
  {
    "start": 2874.12,
    "end": 2874.94,
    "text": "You might fix it a little bit."
  },
  {
    "start": 2875.0,
    "end": 2875.54,
    "text": "You might polish it."
  },
  {
    "start": 2875.58,
    "end": 2880.08,
    "text": "And then you're gonna save it back and you'll send me a video of you clicking through it and going through the whole loop."
  },
  {
    "start": 2880.5,
    "end": 2881.64,
    "text": "That would be the dream, right?"
  },
  {
    "start": 2881.68,
    "end": 2883.1,
    "text": "That's like a self-driving test."
  },
  {
    "start": 2883.14,
    "end": 2885.0,
    "text": "That's like delegating to a QA human, right?"
  },
  {
    "start": 2885.84,
    "end": 2896.42,
    "text": "I think that CI might start walking a little bit more towards actually... that sort of like the mixture of determinism and also QA, that's what we would do with unlimited human time and resources."
  },
  {
    "start": 2896.48,
    "end": 2898.36,
    "text": "We don't because it's like super time extensive."
  },
  {
    "start": 2898.58,
    "end": 2900.74,
    "text": "I'm curious to see how that one grows."
  },
  {
    "start": 2900.88,
    "end": 2902.34,
    "text": "I don't hear a lot of people talking about it yet though."
  },
  {
    "start": 2902.64,
    "end": 2902.84,
    "text": "Okay."
  },
  {
    "start": 2903.0,
    "end": 2903.08,
    "text": "Yeah."
  },
  {
    "start": 2903.14,
    "end": 2903.96,
    "text": "Looking forward to that."
  },
  {
    "start": 2904.46,
    "end": 2907.3,
    "text": "Like CICD deserves to be reinvented."
  },
  {
    "start": 2907.9,
    "end": 2911.38,
    "text": "I mean, the last thing, the last big thing was basically GitHub Actions, right?"
  },
  {
    "start": 2912.1,
    "end": 2914.1,
    "text": "And nothing more after that."
  },
  {
    "start": 2915.04,
    "end": 2916.52,
    "text": "So let's looking, looking forward to that."
  },
  {
    "start": 2917.38,
    "end": 2918.8,
    "text": "Now let's, let's look a little."
  },
  {
    "start": 2919.8,
    "end": 2923.42,
    "text": "Let's go back and look at history."
  },
  {
    "start": 2926.04,
    "end": 2944.78,
    "text": "Your co-founder Merrill basically told me about a secret Easter egg in your software in your stack PR view, which is a time machine feature which allows you to weigh and get, so we're versioning, and it allows us to physically travel back in time."
  },
  {
    "start": 2944.92,
    "end": 2947.56,
    "text": "Let's now imagine I'm using that Easter egg now."
  },
  {
    "start": 2948.42,
    "end": 2956.26,
    "text": "And we're moving back to the year, when you yourself worked as a freshly baked data engineer at Airbnb."
  },
  {
    "start": 2957.82,
    "end": 2963.96,
    "text": "And you know, like we observed yourself for a while, and you know, what have the chance to whisper something into young Greg's ears?"
  },
  {
    "start": 2964.24,
    "end": 2964.98,
    "text": "What would it be?"
  },
  {
    "start": 2967.64,
    "end": 2969.64,
    "text": "Yeah, you know, I had a great time at Airbnb."
  },
  {
    "start": 2970.12,
    "end": 2971.32,
    "text": "It was a really fun time to be there."
  },
  {
    "start": 2971.96,
    "end": 2973.12,
    "text": "The people were really friendly."
  },
  {
    "start": 2973.46,
    "end": 2974.28,
    "text": "The coffee was good."
  },
  {
    "start": 2974.32,
    "end": 2975.56,
    "text": "There was puppies running around."
  },
  {
    "start": 2975.94,
    "end": 2976.86,
    "text": "The culture was really nice."
  },
  {
    "start": 2976.92,
    "end": 2977.74,
    "text": "Free IPI Airbnb."
  },
  {
    "start": 2977.8,
    "end": 2978.98,
    "text": "It was a blast."
  },
  {
    "start": 2979.02,
    "end": 2979.74,
    "text": "It was a really fun time."
  },
  {
    "start": 2980.56,
    "end": 2983.36,
    "text": "And the way I chose to spend my time there is."
  },
  {
    "start": 2983.54,
    "end": 2986.28,
    "text": "I tried to be the best engineer I could."
  },
  {
    "start": 2987.24,
    "end": 2988.2,
    "text": "That was just what I wanted to do."
  },
  {
    "start": 2988.36,
    "end": 2989.58,
    "text": "I wanted to learn as much as possible."
  },
  {
    "start": 2990.04,
    "end": 2997.1,
    "text": "I woke up in the mornings like a total nerd and I would read O'Reilly textbooks, like design data, data intensive applications and streaming architectures and all this stuff."
  },
  {
    "start": 2997.8,
    "end": 3003.66,
    "text": "And I'd go to work and I'd work on Friday teams, DevTool teams, notification infrastructure teams, SRE style teams."
  },
  {
    "start": 3003.72,
    "end": 3004.3,
    "text": "I did a bunch of stuff."
  },
  {
    "start": 3004.82,
    "end": 3012.84,
    "text": "There was even a really try hard period where I would like volunteer on weekends on like the prototyping team working on like a TV, a TV iOS app."
  },
  {
    "start": 3014.0,
    "end": 3014.96,
    "text": "I did as much as possible."
  },
  {
    "start": 3015.82,
    "end": 3017.46,
    "text": "And I wanted to be a really good engineer."
  },
  {
    "start": 3017.5,
    "end": 3018.96,
    "text": "I want to learn what is top tier code quality."
  },
  {
    "start": 3019.02,
    "end": 3020.22,
    "text": "How do I give an amazing review?"
  },
  {
    "start": 3020.9,
    "end": 3022.26,
    "text": "How do I design systems perfectly?"
  },
  {
    "start": 3023.02,
    "end": 3023.66,
    "text": "I missed a thing."
  },
  {
    "start": 3023.76,
    "end": 3025.52,
    "text": "There was a thing I didn't get really good at."
  },
  {
    "start": 3025.6,
    "end": 3026.68,
    "text": "I didn't spend a lot of time on it."
  },
  {
    "start": 3026.72,
    "end": 3030.22,
    "text": "Maybe there wasn't enough time, possibly, but you asked, what would I whisper back in my ear?"
  },
  {
    "start": 3031.56,
    "end": 3035.3,
    "text": "I didn't spend a lot of time understanding how the company operated and worked."
  },
  {
    "start": 3035.58,
    "end": 3039.88,
    "text": "I was very heads down, mostly in the infrastructure department, right in gray code."
  },
  {
    "start": 3040.0,
    "end": 3046.48,
    "text": "I didn't really understand how projects got picked, and I didn't understand what it meant for over a quarter for a team or an org to do well."
  },
  {
    "start": 3046.62,
    "end": 3051.74,
    "text": "I didn't understand what it meant when my sister team's manager departed and then another manager took it over."
  },
  {
    "start": 3052.48,
    "end": 3059.66,
    "text": "Or there's these like free-floating principal engineers who are really good at getting stuff done, but like I didn't really, I don't like, you just come into a room and tell me some stuff and then leave."
  },
  {
    "start": 3059.7,
    "end": 3060.14,
    "text": "Like, who are you?"
  },
  {
    "start": 3060.2,
    "end": 3060.94,
    "text": "How are you operating?"
  },
  {
    "start": 3062.02,
    "end": 3063.66,
    "text": "There was all this stuff that was way over my head."
  },
  {
    "start": 3063.98,
    "end": 3064.42,
    "text": "Profits."
  },
  {
    "start": 3064.5,
    "end": 3067.74,
    "text": "I didn't understand how the business motivations played into what I was building."
  },
  {
    "start": 3067.78,
    "end": 3072.18,
    "text": "I was making like releases and databases better and faster, but like, I don't know."
  },
  {
    "start": 3072.5,
    "end": 3077.56,
    "text": "Everybody was trying to IPO and I didn't have a deep understanding of how these all connected and boiled down to me."
  },
  {
    "start": 3078.04,
    "end": 3079.8,
    "text": "A CTO changed while I was working there."
  },
  {
    "start": 3080.36,
    "end": 3081.4,
    "text": "I didn't understand what was happening."
  },
  {
    "start": 3081.82,
    "end": 3093.36,
    "text": "So now that I've founded a company and helped build it up and I'm watching and now, you know, helping cursor as well, I have a way, way, way better understanding, not a perfect understanding, but way better understanding of all these dynamics."
  },
  {
    "start": 3093.88,
    "end": 3102.52,
    "text": "And the best engineers I work with understand what's going on because As I kind of said earlier as well, any good engineer knows how to build a thing."
  },
  {
    "start": 3102.96,
    "end": 3107.56,
    "text": "It's much more interesting of what to build and what not to build and when to build it and who to work with."
  },
  {
    "start": 3108.12,
    "end": 3114.08,
    "text": "That seems to define upper bound excellence of problem solving and efficiency and impact within these companies."
  },
  {
    "start": 3115.28,
    "end": 3117.18,
    "text": "I was missing that gap at Airbnb."
  },
  {
    "start": 3117.3,
    "end": 3128.76,
    "text": "So I would encourage, for anyone else listening, if you're trying to be the best engineer in the world, understanding the business, talking to PMs and designers and sales folks and customer service, Understand talk to your leaders turn understand."
  },
  {
    "start": 3128.8,
    "end": 3129.44,
    "text": "What are they thinking about?"
  },
  {
    "start": 3129.48,
    "end": 3130.4,
    "text": "What do they actually care about?"
  },
  {
    "start": 3130.94,
    "end": 3134.02,
    "text": "It's a different set of motivations and understandings than just writing really good code."
  },
  {
    "start": 3134.48,
    "end": 3138.74,
    "text": "But if you boil it back down and then you and then you then you actually get in the weeds and build some things."
  },
  {
    "start": 3140.04,
    "end": 3142.12,
    "text": "That's how you'd be the best at the job."
  },
  {
    "start": 3142.88,
    "end": 3152.26,
    "text": "So you rather be interested in the what the why and the when And And the business side, right?"
  },
  {
    "start": 3152.72,
    "end": 3163.6,
    "text": "I mean, that's also what I think like many engineers like got closer to in the last years, but I think like in a nutshell, our species is also not perfect in that, right?"
  },
  {
    "start": 3163.68,
    "end": 3166.16,
    "text": "Like yet, but that will change, right?"
  },
  {
    "start": 3166.3,
    "end": 3169.24,
    "text": "And that has to change, I think, also with AI being around, right?"
  },
  {
    "start": 3169.98,
    "end": 3174.56,
    "text": "And that advice is presupposed that you're trying to be the best and most impactful in a shared company."
  },
  {
    "start": 3174.96,
    "end": 3175.74,
    "text": "Not everyone's trying to be them."
  },
  {
    "start": 3175.78,
    "end": 3176.6,
    "text": "That's totally fine too."
  },
  {
    "start": 3176.66,
    "end": 3178.96,
    "text": "I don't know Valley judgment on that, you know, do what you want to do."
  },
  {
    "start": 3179.02,
    "end": 3184.36,
    "text": "But if that's what you're trying to do, I think diversifying your understanding of the different business functions in the broader picture is key."
  },
  {
    "start": 3184.84,
    "end": 3188.7,
    "text": "Yeah, I smell that we all have to be a bit more full stack, right?"
  },
  {
    "start": 3189.22,
    "end": 3195.56,
    "text": "There's maybe no future where you just platform or just data."
  },
  {
    "start": 3196.88,
    "end": 3197.1,
    "text": "I don't know."
  },
  {
    "start": 3197.5,
    "end": 3197.82,
    "text": "Let's see."
  },
  {
    "start": 3198.86,
    "end": 3199.22,
    "text": "Let's see."
  },
  {
    "start": 3199.38,
    "end": 3203.0,
    "text": "And let's see how we live in twenty nine, right?"
  },
  {
    "start": 3204.28,
    "end": 3205.04,
    "text": "Looking forward to that."
  },
  {
    "start": 3205.36,
    "end": 3208.12,
    "text": "Let's record another podcast and latest by that time."
  },
  {
    "start": 3209.1,
    "end": 3210.6,
    "text": "Thanks a lot, Greg, for being my guest."
  },
  {
    "start": 3210.96,
    "end": 3211.72,
    "text": "Thanks a lot."
  },
  {
    "start": 3211.8,
    "end": 3212.78,
    "text": "Yeah, let's set a reminder."
  },
  {
    "start": 3213.34,
    "end": 3222.26,
    "text": "And really looking forward to see how you shape the future at Cursa now and how code reviews will look like tomorrow."
  },
  {
    "start": 3223.34,
    "end": 3228.44,
    "text": "I'll follow you there and really looking forward to see you guiding the way there."
  },
  {
    "start": 3228.7,
    "end": 3228.96,
    "text": "Awesome."
  },
  {
    "start": 3229.02,
    "end": 3230.22,
    "text": "Toby, thank you so much for having me."
  },
  {
    "start": 3230.46,
    "end": 3230.78,
    "text": "Thanks a lot."
  },
  {
    "start": 3230.86,
    "end": 3231.46,
    "text": "Always happy to chat."
  },
  {
    "start": 3231.6,
    "end": 3232.6,
    "text": "This is such a fun discussion."
  },
  {
    "start": 3232.66,
    "end": 3233.1,
    "text": "Absolutely."
  },
  {
    "start": 3233.26,
    "end": 3233.72,
    "text": "Thanks a lot."
  },
  {
    "start": 3233.8,
    "end": 3234.28,
    "text": "Have a great day."
  },
  {
    "start": 3234.52,
    "end": 3234.7,
    "text": "Bye."
  },
  {
    "start": 3236.44,
    "end": 3238.66,
    "text": "Thank you for listening to the AlphaList podcast."
  },
  {
    "start": 3238.88,
    "end": 3241.32,
    "text": "If you liked this episode, share it with friends."
  },
  {
    "start": 3241.56,
    "end": 3242.44,
    "text": "I'm sure they love it too."
  },
  {
    "start": 3242.78,
    "end": 3249.5,
    "text": "Make sure to subscribe so you can hear deep insights into technical leadership and technology trends as they become available."
  },
  {
    "start": 3249.92,
    "end": 3257.06,
    "text": "Also, please tell us if there is a topic you would like to hear more about or a technical leader whose brain you would like us to pick."
  },
  {
    "start": 3257.5,
    "end": 3264.38,
    "text": "Alphalist is all about helping CTOs getting access to the insights they need to make the best decisions for their company."
  },
  {
    "start": 3264.76,
    "end": 3268.52,
    "text": "Please send us suggestions to cto at alphalist.com."
  },
  {
    "start": 3269.1,
    "end": 3270.84,
    "text": "Send me a message on LinkedIn or Twitter."
  },
  {
    "start": 3271.26,
    "end": 3275.62,
    "text": "After all, the more knowledge we bring to CDOs, the more growth we see in tech."
  },
  {
    "start": 3275.94,
    "end": 3279.96,
    "text": "Or, as we say on AlphaList, accumulated knowledge to accelerate growth."
  },
  {
    "start": 3280.5,
    "end": 3290.96,
    "text": "See you in the next episode."
  }
]