The rise of Cursor: The $300M ARR AI tool that engineers can’t stop using | Michael Truell (co-founder and CEO)

0:00 Our goal with Cursure is to invent a new type of programming, a very different way to build software. So a world kind of after code, I think that more and more being engineer will start to feel like being a logic designer. And really it will be about specifying your intent for how exactly you want everything to work. What is the most counterintuitive thing you've learned so far about building cursor? We definitely didn't expect to be doing any of our own model development. And at this point, every match moment in cursor involves a custom model in some way. What's something that you wish you knew before you got into this role? Many people you hear hire too fast. Too slow to begin with. You guys went from zero dollars to a hundred million ARR in a year and a half, which is historic. Was there an inflection point where things just started to really take off? The growth has been Barely just consistent on an exponential. An accidental to begin with feels fairly slow and the numbers are really low and it didn't really feel off to the races to begin with. What do you think is the secret to your success? I think it's been

0:55 Today my guest is Michael Trul. Michael is co-founder and CEO of AnySphere, the company behind Cursor. If you've been living under a rock and haven't heard of Cursor. It is the leading AI code editor, and is at the very forefront of changing how engineers and product teams build software. It's also one of the fastest growing products of all time. Hitting one hundred million ARR just twenty months after launching.

1:18 And then three hundred million ARR just two years since launch. Michael's been working on AI for ten years, he studied computer science and math at MIT, did AI research at MIT and Google, and is a student of tech and business history. As you'll soon see, Michael thinks deeply about where things are heading and what the future of building software looks like. We chat about the origin story of Cursor, his prediction of what happens after code. His biggest counterintuitive lessons from building cursor. Where he sees things going for software engineers and so much more.

1:49 Michael does not do many podcasts. The only other podcast he's ever done is Lex Friedman, so it was a true honor to have Michael on. If you enjoy this podcast, don't forget to subscribe and follow it in your favorite podcasting app or YouTube. Also, if you become an annual subscriber of my newsletter, you get it a year free of perplexity, linear superhuman notion, and granola. Check it out at lenny's newsletter.com and click bundle. With that.

2:11 I bring you Michael. Troll. This episode is brought to you by Epo. Epo is a next generation A B testing and feature management platform Built by alums of Airbnb and Snowflake for modern growth teams.

2:25 Companies like Twitch, Miro, ClickUp, and DraftKings rely on Epo to power their experiments. Experimentation is increasingly essential for driving growth. and for understanding the performance of new features. An Epo helps you increase experimentation velocity while unlocking rigorous deep analysis in a way that no other commercial tool does. When I was at Airbnb, one of the things that I loved most was our experimentation platform. where I could set up experiments easily, troubleshoot issues, and analyze performance all on my own. Epo does all that and more, with advanced statistical methods that can help you shave weeks off experiment time, an accessible UI for diving deeper into performance, and out of the box reporting that helps you avoid annoying, prolonged analytic cycles.

3:07 Apple also makes it easy for you to share experiment insights with your team, sparking new ideas for the A B testing flywheel. Apple powers experimentation across every use case. Including product. Growth. Machine learning, monetization, and email marketing. Check out epo at getepo.com slash lenny.

3:25 And ten X your experiment velocity. Let's get EPPO. dot com slash Lenny. This episode is brought to you by Vanta. When it comes to ensuring your company has top notch security practices.

3:38 Things get complicated. Fast. Now you can assess risk, secure the trust of your customers, and automate compliance for SOC 2, ISO 27001. HIPAA and more with a single platform, Vanta. Vanta's market leading trust management platform helps you continuously monitor compliance.

3:58 alongside reporting and tracking risks. Plus, you can save hours by completing security questionnaires with Vanta AI. Join thousands of global companies that use Vanta to automate evidence collection. Unify risk management, and streamline security reviews. Get one thousand dollars off Vanta when you go to vanta.com slash plenty. That's V A N T A dot com slash Lenny.

4:26 Michael, thank you so much for being here. Welcome to the podcast. Thank you. Uh it's great to be here. Thank you for having me. When we were chatting earlier, you had this really interesting phrase, this idea of what comes after code. Talk about that. Just like the vision you have of where you think things are going in terms of

4:43 moving from code to maybe something else. is to invent Uh sort of a n a new type of programming. Um a very different way to build software.

4:53 That's kind of just distilled down into you describing the intent. For what you watched. In the most concise way possible. Uh and really distilled down to your just.

5:03 Defining how you think the software should work. And how you think it should look. And Yeah, with with you know the technology that we we have say in as an mature is it Uh we think you can get to a place where method of building software that's

5:15 Legions higher level and more productive. uh in some cases more more accessible too. And um that that process will be will be a gradual moving away. From

5:27 You know, what building software looks like today. Um And um I you know, I want to contrast it with maybe like the vision of you know what software looks like in the future.

5:38 on that you know, I think you know, couple visions there in a popular conscious. That we at least Um Um one is, you know, there's a group of people who think that um

5:49 you know, software building in the future is gonna look very much like it it does today, which mostly means text editing formal programming languages like TypeScript and Go and C and Rust. Uh, and then there's another group that kinda thinks like Yeah, you're just gonna type into A a bot and you're gonna ask it to build you something and then you're gonna ask it to to change the layout what you're building.

6:06 And it's kind of like this, you know, chatbot, Slackbot style where you're talking to your engineering department. And We think that there are problems with with both of those visions. I think that on the you know, on the the chatbot style of things. Um and we think it's gonna look like weirder than both.

6:21 Um The problem with the the chatbot style end of things. Is Um that lacks a lot of precision. If you want humans to have completely

6:31 you know, complete control over what the software looks like. And how it works. You need to let them, you know, gesture at what they want to be changed. Um You know, in a form factor that's more precise than just, you know.

6:42 change this about my app, you know, kind of in a text box removed from the whole thing. And then Um you know, the the version of the world where kind of nothing changes we think we think is is wrong because we think that the the technology's gonna get much, much, much better. Uh and so a world, you know, kind of after You know, after code.

6:58 Um I think that it looks like a world where you have a representation of the logic of your software that does look more like English. Right you have kind of written down You can imagine in Dauphin form, you can imagine in kind of an evolution of

7:11 You have written down you know, the logic of the software and you can you can edit that at a high level and you can point at that. And it won't be kind of The the impenetrable millions of lines of code. Um it'll instead be something that's like much tier and easier to understand and easier to navigate.

7:24 But that world where yeah. The the kind of Crazy hard to understand symbols. start to evolve towards something that's a little bit more Uh and readable.

7:32 Uh and human editable. Uh is one is one that we're working toward. This is a profound point. I think I I I wanna make sure people don't miss what you're saying here, which is that What you're envisioning in the next year essentially, uh is kinda when the things start to shift is Uh people move away from

7:46 even seeing code, think having to think in code and like JavaScript and Python. And there's this abstraction that w will Appear. Uh essentially Sudoku describing what

7:56 The code should be doing more in English sentences. Yep, we we think it ends up Ends up looking like that. I mean and we're very opinionated that that path goes through kind of existing professional engineers. And it looks like this this evolution away from code.

8:09 Uh and it definitely looks like the human still being in the driver's seat, right? And the human having both a ton of control over all aspects of the software. And not giving that up. And then also uh the human having the

8:22 Um, need changes very quickly. Like having a fast duration movement, not just like You know having something in the background that's that's super slow and takes like weeks. I go do all your work for you. This uh

8:34 Begs the question. Four people that are getting are currently engineers or thinking about becoming engineers. or designers or product manager, like what skills do you think will be more more and more valuable in this world of

8:48 The What comes after code? I think taste will be increasingly More valuable and I think often when people think about teeth in the realm of software, they think about

8:58 You know, visuals or Chased over smooth animations and Uh yeah. coloring things, UI, U S, et cetera. On kind of the visual design of things.

9:08 And I think more and more You know, the visual side of things in Yeah, a piece of software, but then Yeah, as mentioned before, I think that the other half

9:17 of defining a piece of software is the is the logic of that. And how the thing works. And Uh, we have amazing tools for specing out the visuals of things. And then when you get into the the logic of how a piece of software works.

9:30 Really the best representation we have of that is code right now. You can kind of gesture at it with Figma and you can gesture at it with writing down notes. Um but it's you know, when you have an actual working prototype. And so I think that more and more being being an engineer will start to feel like being a logic designer.

9:46 And really it will be about specifying your intent for how exactly you want everything to work. And it will last be about I'd be more more about the the what's and a little bit less about the how. Um exactly you're gonna do things under the hood. Uh and so yeah, I think I think taste will be increasingly important.

10:01 I think one aspect of software engineering and we're very far from this right now and there are lots of Yeah. uh funny funny memes going around the internet about, you know, the kind of the some of the trials and tribulations people can run into if they trust AI and for too many thought comes to engineering. Um around you know.

10:15 Uh building building apps that uh you know have have glaring glaring deficiencies and and problems and Functionality issues. But um I think we will get to a place where Um

10:26 you will be able to uh Be less careful as a software engineer. Which right now is an incredibly, incredibly important uh skill. Um And yeah, we'll move a little bit from carefulness and a little bit more towards taste.

10:40 This uh makes me think of vibe coding. Is that kinda what you're describing when you talk about not having to think about the details as much and just kinda Going with the flow. I I think it's r I think it's related. I think that by coding right now describes Um

10:53 exactly kind of this this state of creation. That uh uh is pretty controversial where you're generating a lot of code and you aren't really understanding the details. That is that is like a a state of creation that then has has lots of problems, like you don't really By by not understanding the details near the hood right now.

11:11 you then very quickly get to a place where you're kind of limited. At a certain point where you create something that's big enough that that you can't change. And so I think some of the some of the you know ideas that we're interested around. Yeah. How do you

11:22 Give people uh continued control over all the details. Um You know, when they don't really understand the code. Like I think that um Solutions there.

11:31 um are very relevant to to the people who are biocoding right now. I you know, I think that uh right now we we kind of we lack the ability to, you know, let the the team makers actually have complete control over the software. And so Um One of the one of the issues also with you know, with five coding and and letting letting taste really shine through from people is

11:51 You can create stuff, but a lot of it is the AI making decisions that you or unwieldy and we don't have control over. One more question along these lines. You throw out this word taste. When you say taste, what are you thinking? I'm thinking having the right idea for for what should be built.

12:04 And then just it it will become more and more about kind of effortless translation of here's exactly what you want built, here's how you want everything to work, here's how you want it to look, and then you'll be able to Meet that. um on a computer and it will less be about this kind of translation layer of like you and your team have a picture of what you'd want to build. And then

12:22 You have to really painstakingly labor intensive like layout. that into a format that a computer can then execute and interpret. And so yeah, I think it you know, less is less than the U I side of things, maybe Chase is a little bit of a misnomer.

12:35 Uh but Just about having the right idea for for what should be built. Awesome, okay. I'm gonna come back to these topics but I wanna actually zoom us back out. To

12:44 The beginnings of Cursor. Uh I have never heard the origin story. I don't think many people know how this whole thing started. Basically you guys are building one of the fastest growing products in the history of the world.

12:56 It's changing the way people build product, it's changing careers, professions, it's ch it's changing so much. How did it all begin? Any memorable moments along the journey of the early days? Cursure kind of started as a solution search for a problem. Um and

13:10 Uh a little a little bit. Where it very much came from Reflecting on um how AI was gonna get better. Um, over the course of the next ten years. And

13:20 Um There were there were kind of two defining moments. One was uh being really excited by using the the The first theater version of it, actually. This was the first time we had used an AI product. That was

13:34 Really, really, really useful. And Um was you know actually just useful at all. Uh and wasn't just a vaporware kind of demo thing. And in addition to being an a you know, the first AI product that we use that was useful.

13:48 Gab Cobello is also one of the most useful, if not the most useful dev tool we've ever adopted. Um And that got us really excited. Yeah. Another moment that got us really excited was the series of

13:59 uh scaling once papers coming out of OpenAI and other places. That showed that even if we had no new ideas. AI was gonna get better and better just by pulling on simple levers, like Scaling up the models and also scaling up the the data that was going into the models. And so at the end of twenty twenty one, beginning of twenty twenty two, this got us excited about how, you know

14:16 Yeah, product for now possible. this technology was going to mature uh into the future. And It felt like when we looked around There were lots of people talking about

14:26 Making models. And There it felt like people weren't really picking an area of knowledge work and thinking about what it was gonna look like as AI got better and better. And

14:37 Um, you know, that set us on the the path to like an you know, kind of an idea generation exercise. It was like you know, how are each Th these areas of knowledge work. uh gonna change in the future as this tech gets more mature. Like what is the Yeah.

14:49 And state of the work gonna look like Um how are the the tools that we use to do that work gonna change? Um how are the models gonna get? Yeah. Need to get better.

14:57 Uh changes in the work. And yeah, once scaling and pre training right now, like how are you gonna keep pushing forward technological capabilities? And The miss staff at at the beginning of first series, we actually work on

15:12 Yeah, read. Sort of did this whole grand exercise. Uh and we decided to work on You know, uh an area uh of knowledge worth that we thought would be relatively uncompetitive and sleepy and and boring. Uh and you know, no one no one would be looking at it'cause you know, we thought, Oh, c coding's great, you know, coding's totally interchangeable. But yeah, people are already doing that.

15:31 And uh so there was a period of Yeah, four months to begin with. where we were actually working on a very different idea, which was Helping to automate and augment mechanical engineering.

15:42 Uh and building tools for mechanical engineers. You know, there were problems from the get go in that. We had uh, you know Me and my co founders, we we weren't mechanical engineers. Um

15:52 Yeah, we had friends who were mechanical engineers, but uh we were we were very much unfamiliar with the field. So there's a little bit of uh blind man and the elephant problem from the get go. Uh, you know, there were problems around uh you know, how would you actually take take the models that exist today and make them useful for mechanical engineering? The way we net it out is you need to actually develop your own models for the get-go. And if the way we did that was

16:12 Uh it was tricky and you know, there's Not a lot of Uh data on the internet. of of um you know three D models of Uh of different tools and parts and the steps that I showed to

16:23 Built built after those three models. Uh and then getting them from the sources that then have them is like also a tricky process too. But um Eventually what happened was, you know, we we came to our senses, we realized we're not super excited about mechanical engineering. It's not.

16:37 Did the thing we want to take in our last two. And we looked around and uh in the area of programming It felt like you know. Despite Uh you know, a decent amount of time ensuing.

16:46 Uh not much has changed. And it felt like the people that were working on the space maybe had a had a disconnect with us and it felt like they weren't Being sufficiently ambitious about Um

16:56 where everything was gonna go in the future and how kind of all of software creation was gonna flow through these models. Uh and that's what set us off on the the path to to building Kersha. Okay. So interesting. Okay, so first of all I love that there's this this is advice that you often hear of go after a boring industry'cause no one's gonna be there and there's opportunity. And you know, sometimes it works, but I love that in this journey it's like, No, actually go after the hottest

17:17 most uh popular space AI coding app building. And it worked out. And the way you phrased it just now is You didn't see enough ambition, potentially. But you thought there was more to be done. So it feels like that's an interesting lesson if even if something looks like okay, it's too late, there's GitHub call pilots out there, some other products.

17:36 If you notice that they're just not as ambitious as they could be or as you are. Where you see Almost a flaw in their approach that there's still a big opportunity. Does that resonate? Uh that totally resonates. And I think it's um

17:49 A part of it is you need there to be a Like leap rogs that can happen, you'd be there to be things that you can do. And I think the exciting thing about AI is in a in a bunch of places, and I think this is actually Very much still true of our space.

18:03 And can talk about how we think about that and how we deal with that. But um Yeah, I think that the just the ceiling is really high. And um yes, if you're if you look around uh Yeah.

18:13 Probably even if you're you Take the best tool on kind of like any of these Any of these fields. Um, there's gonna be a lot more that needs to be done over the next few years. And so That that space having that space, having that

18:23 high ceiling I think is is unique. Um amongst areas of software, at least the degree to which it is it is high with AI. Let's come back to the ID question. So there's kind of a few routes you could have taken and other companies are doing different routes, so there's building an ID for engineers to work within and adding AI magic to it. There is another out of just a full AI agentic.

18:43 Devin sort of product. And then there's just like a model. that is very good at coding and focusing on building the best possible coding model. What made you decide and see that the ID path was the best? Route.

18:54 The folks sure from the get go Working on just a Uh A model we're working on And to end.

19:03 Uh emission. A uh programming? I think Uh They were trying to build something very different from us, which is me care about

19:11 Giving humans control over all the decisions. Um in kind of the end tool that they're building. And I think the those folks were very much thinking of a of a future where Kind of Yeah, and the whole thing is done by AI.

19:24 And maybe like the AI is making all the decisions too. And so one, there was kind of like a personal interest component. Two, I think that Uh always we try to be uh intense realist about where the technology is today. you know, very, very, very excited about how AI is going to mature over the course of

19:38 Many decades. But Uh yeah. I think that sometimes uh people Yeah, there's a there's an instinct to to see AI do magical things in one

19:48 area and then kind of anthropomorphise these models and think It's better than a smart person here and so must be better than a smart person there. But these things have massive issues. And um We

20:00 Uh from the From the very start, our our product development process was really about dog fooding. and using the tool intensely every day. And we we never wanted to ship anything that wasn't wasn't useful to us. And you know, we had the benefit of doing that because we were the end user part of our product.

20:17 And I think that that instills uh realism in you. Around where the where the tech is right now. And so Uh

20:24 That definitely made us think that we need the humans to be in the driver's seat, the AI cannot do everything. We're also interested in giving humans that control too for for personal reasons. And so that that gets you away from just you're a model company that also gets you away from just kind of this And to end stuff with without the human having control. And then The way you get to an IDE versus maybe a plug into an existing coding environment.

20:44 Is Uh the belief that you know, programming is gonna flow through these models and the act of programming is gonna change a lot. Over the course of the next few years. And that the extensibility that existing coding environments have is so, so, so limited.

20:57 So if you think that the UI is going to change a lot, if you think that the form factor programming is going to change a lot, necessarily need to have control over the entire application. I know that you guys today have an IDE and Uh And that's probably the bias you have of this is maybe where the future is heading, but I'm just curious, do you think A big part of the future is also going to be

21:14 AI engineers that are just sitting in Slack and just doing things for you. Is that something that fits into Cursure One Day? I think you'll want the ability to move between all these things fairly effortlessly. Mm. Sometimes I think you will want to have the thing kind of go spin.

21:30 Off on its own for a while. And then I think you'll want the ability to pull in D AI's work and then work with it very, very, very quickly. Right. And then maybe have it go spin off again. And so these like

21:42 Background versus for round form factors, I think you want that all to work well in one place. And Uh I think the background stuff There's like a segment of programming that it's especially useful for.

21:53 Which is Type of programming tasks where we're going to be able to do It's very easy to specify Exactly what you want. Um

22:00 With you know, without much description and exactly what correctness looks like without much description. And often that's the bug fixes are kind of like the Are are a great example of that. But it's definitely not all of programming. So I think that w you know, what the IDE is, uh will totally change around.

22:17 You know, having our own editor. I was Paristine. It's gonna have to evolve over time. And I think that that will both include You can spin off things from different surface areas like Slack or your issue tracker or whatever it is.

22:27 And I think that will also include like, you know. the pane of glass that you're staring at is gonna change a lot. Um And you know, we just mostly think of an ID as the place where you are building software. I think something people don't talk enough about w with talking about agents and all these

22:43 Uh yeah. engineers are gonna be doing all the stuff for you. Basically we're all becoming uh engineering managers. with a lot of reports that are just like not that Not that smart and you have to do a lot of Reviewing and approving and specifying.

22:56 I guess thoughts on that and is there anything you could do to make that easier?'Cause that sounds really hard. Like anyone that has a large team has had a large team being like, Oh my god, all these Uh junior people just checking in with me doing Not high quality work over and over. It's just like What a life.

23:10 Yeah. Maybe eventually one on ones with uh. So many one on ones. Uh yeah, so the the customers we've seen have uh most success with AI I think are still fairly conservative. About some of the ways in which

23:27 in which they they use his stuff. And so I do think today The most successful customers. really lean on Things like um, you know, our next act edited prediction.

23:37 Where we you know, your coding is normal and making the next scenes of actions you're gonna do. And then they also really lean on like scoping down the stuff that you're gonna hand off uh to the bot. And You know, there's

23:47 for a fixed percent of your time spent reviewing code. You could Um from from an agent. Um or from an AI overall. You can you know there's

23:56 Kind of two patterns. One is you could Yeah. Spend a bunch of time specifying things up front. AI goes and works. And then you then go and review the AI's work. And then you're done. That's the whole task.

24:07 Or you could really chop things up, right? So you can You know, specify a little bit AI write something, review. Specify a little bit AI write something review. And that's kind of, you know, auto completes all in the way of Gas spectrum. And um still we see uh

24:20 Often the most uh successful people. Um using these tools. are are are chalking things up right now. And can you fairly scale. This sounds less less terrible. I'm gr I'm I'm glad there's a solution here.

24:31 I'm gonna go back to You guys building cursor for the first time? What was the point where you realize this is ready? What was kind of a moment of like, Okay, I think this is time to put it out there and see what happens. So when we started building cursor Um

24:45 We were Uh fairly paranoid about spinning for a while. Without releasing to the world. And so to to begin with, too, we actually the the first version of Cursure was was hand rolled.

24:57 Yeah. Uh now we we use uh VS Code kind of as a base, like many browsers use Chromium as a base. Um in his for top of that. Uh to begin with, we we didn't and built the proto shape of cursor from scratch.

25:10 And That involved a lot of work. We had to build our own Uh, you know, there are a lot of things that go into you know, a modern code editor. Uh including um you know, support for many different languages and um navigation support for moving amongst the language, you know, error tracking support for things.

25:26 There's you know, things like, you know, in an integrated command line, you know, the ability to use like remote servers to Uh to you know to the the ability to capture remote servers to to view and and run code. And so we kind of just like went on this blitz of building things incredibly quickly. building kind of our own uh editor from scratch and then also the AI components. And um it was after like a couple of months That we just

25:48 Yeah. Uh. It was after maybe five weeks that we were Living on the editor full time. And you know, had thrown away our previous editor.

25:56 Uh and we're we're using new one. And then once it got to a point where we found it. A bit useful. Then we put it in other people's hands and had this like very short beta period. And then we launched out to the world within uh a couple of months from the first flying code.

26:10 I I think it was probably probably three months. And it was definitely a like You know, let's let's just get this out to people and build in public quickly. The thing that took us by surprise is we thought we would be building for a couple hundred people. For a long time.

26:22 And you know, fr from the get go there there was kind of an immediate Prussia's interests. And a lot of feedback too. Uh and you know, that was super helpful. We learned from that. And that's actually you know, why we switch to being based off of

26:33 Yes, code. Instead of just, you know, this hand roll thing. Uh, a lot of that was motivated by kind of the initial user feedback. And uh you know, and then have been iterating in in public. Uh from there.

26:44 I like how you understated uh the The traction that you got. Uh I think you guys went from zero dollars to a hundred million ARR in like a year, year and a half, or something like that, which is uh historic What do you think was the key to the s to success?

27:00 of something like this. He's talked about dog fooding being a big part of it. Like you build it in three months. That's insane. Mm-hmm. Uh What do you think was is the secret to your success? The first version was n was not

27:12 You know, the three month version wasn't very good. And so I think it's been, you know, a sustained paranoia about, you know, there are all of these Ways in which this this thing could get better.

27:22 Yeah. The end goal is really to uh invented a new form of programming that involves automating a lot of coding as we Uh, no no today. And um no matter

27:32 You know, where we are with Cursor, it feels like we're very, very far away from that angle. And so there's Uh, there's always a lot to do. But I think it's been kind of uh A lot of it hasn't been r over rotated on kind of the initial push. But instead is like the continued evolution of the tool and just making the tool consistently better.

27:47 Was there an inflection point after those three months where things just started to really take off? To be honest, it felt fairly slow to begin with. Um And Yeah, maybe maybe comes from some in impatience on our part. Um

27:59 But uh one one I think you know, there's the the overall speed of the growth, which is um Uh you know, continues to take us by surprise. I think one of the things that uh has been most surprising too is that the growth has been

28:13 Fairly just consistent on an exponential. of just consistent month over month growth. Accelerated at times by Um launches on our part and other things. But

28:24 Uh you know But an exponential to begin with feels feels fairly slow and the the numbers are really low and Uh so it didn't it didn't really show after the races to begin with. To me this sounds like build it and they will come actually working. You guys just built an awesome product that you loved. yourselves as engineers. You put it out and people

28:40 Just loved it, told everyone about it. Essentially all just Uh You know, the the team working on the product and making the product good. In lieu of

28:51 you know, other things one could spend one's time on. Yeah, we we definitely spent time on tons of other things. For instance, building the team is incredibly important. Yeah, you know. Um

29:00 I Doing things like uh support rotations are very important. But some of the normal things that people would maybe uh Reach four. uh in in building the company early on.

29:10 Um, we really let those fires burn for a long time, especially when it came to things like like sales and marketing. And so just working on the product and building a product that you like for Yeah, your team likes and then you know, also then adjusting it for a some set of users. That can kind of sound simple.

29:24 But then Yeah, it's har hard to do that well. And uh there are a bunch of different directions one could have run in. Bunch of different product directions. And I think that um you know, one of the difficult things

29:35 Yeah, I think Focus and kind of strategically picking the right things to build. And prioritizing effectively is tricky. Think another thing that's tricky about this Uh Mr. Maine.

29:46 Is It's kind of a new form of product building. where um it's very intradisciplinary in that we are something in between a normal software company. Uh and then in between a normal software company and then a a foundation model company?

30:00 In that um Yeah. Uh we want to develop a Yeah. We're developing a product for millions of people.

30:06 And that you know, that side of things has to be excellent. Then also one in is doing more and more on the science and doing more and more on the model side of things. Uh in places where it makes sense.

30:17 And so that element of things doing that well too. H has been tricky. But yeah, you know the overall thing would note is yeah. Uh maybe Yeah. Some of these things sound

30:26 It's sound simple to specify, but then like doing them well is is hard and there are buff different ways you can run it. I'm excited to have Andrew Luo joining us today. Andrew is CEO of One Schema, one of our longtime podcast sponsors. Welcome, Andrew. Thanks for having me, Lenny. Great to be here. So, what is new with one schema? I know that you work with some of my favorite companies like Ramp and Banta and Watershed. I heard you guys launched a new data intake product that automates the hours of manual work that teams spent importing and mapping and integrating CSV and Excel files. Yes? So we just launched the 2.0 of One Schema File Feeds. We've rebuilt it from the ground up with AI.

31:02 We saw so many customers coming to us with teams of data engineers that struggled with the manual work required to clean messy spreadsheets. Filefeeds 2.0 allows non-technical teams to automate the process of transforming CSV and Excel files. with just a simple prompt. We support all the trickiest file integrations, SFTP, S3, and even email. I can tell you that if my team had to build integrations like this, how nice would it be to take this off our roadmap and instead use something like one schema?

31:30 Absolutely, Lenny. We've heard so many horror stories of outages from even just a single bad record in transactions, employee files, purchase orders, you name it. Debugging these issues is often like finding a needle in a haystack. One schema stops any bad data from entering your system and automatically validates your files, generating error reports with the exact issues in all bad files. I know that importing incorrect data can cause all kinds of pain for your customers and quickly lose their trust. Andrew, thank you so much for joining me. If you want to learn more, head on over to onschema.co. That's oneschema.co.

32:05 What is The most counterintuitive thing you've learned so far about building cursor building AI products. I I think one thing that's been counterintuitive for us uh hinted at At it a little bit before. But is we we definitely didn't expect to be doing any of our own model development when we started.

32:21 I as mentioned, you know, when we when we got into this, there were companies that were immediately from the get go Going and Just focusing on kind of training them all from scratch. And We had done the calculation for it.

32:32 To d to train G D four and Just knew that that was not something we were gonna be able to do. And also felt a little bit like Uh focusing one's attention in the wrong area. Іказ

32:43 There are lots of amazing models out there and why do all of this work to replicate What other players have done. Especially on the pre training side of things. You know, taking a uh a neural network that knows nothing and then teaching it the whole internet.

32:55 And Uh, so we thought we were we were gonna be doing that. uh at all. And it seems uh clear to us from the start that the the existing models, there were lots of things that They could be doing for us that Um they weren't doing'cause you know, there wasn't the right tool bill for them.

33:09 In in fact though, we do a ton of model development and internally it's a it's a big Um focus for us on the hiring front. uh and have assembled a a fantastic team there. And it's also been a big win on the the product quality side of things for us. Uh and at this point

33:23 Every magic moment in cursure. involves a custom model in some way. And Uh so that that was definitely counterintuitive and and surprising and Uh

33:33 It's it's been a gradual thing where you know there was an initial use case for for training our own model where it really didn't make sense to use any of the biggest foundation models that was incredibly successful, kinda moved to another use case that worked really well. Uh and has have been going from there. And one of the You know, the the helpful things and And doing this for model development.

33:52 Is is picking your your spots carefully. Not trying reinvent the wheel, not trying to focus on places. Maybe where the the The best. foundation models are are excellent, but instead kind of focusing on their weaknesses and

34:04 How you can complement them. I think this is gonna be surprising to a lot of people hearing that you have your own models. When I You know, when people talk about cursor and all the folks in the space, they would kinda call'em G P T rappers, they're just sitting on top of Chat GPT or Sonnet and what you're saying is that you have your own models.

34:19 Talk just like the stack behind the scenes. Yeah, of course. Um so we definitely use uh the biggest Chinese models uh You know, a bunch of different ways. Um the really important components of uh bringing the cursor experience to people. The the places where we use our own models.

34:33 So Sometimes it's to survey use case. that a foundation model wouldn't be able to serve at all for cost or speed reasons. And so one example of that is um Uh the auto melee side things. And so

34:46 This can be a little bit tricky for Uh People who don't code to understand. But code is this weird form of work. Or sometimes

34:56 really the next five, ten, twenty, thirty minutes of your work. is entirely predictable from looking over your shoulder. And I would contrast this with writing. So writing Yeah.

35:06 Everyone or a lots of people are familiar with Yeah, Gmail's autocomplete and Different forms of that show up when you're Right those text messages or emails or things like that. They can only be so helpful.

35:17 'Cause often it's just really not clear what you're gonna be writing. Just by looking at what you wrote in before. But in code sometimes when you edit a part of a code base. It's just you're gonna need to change things and in other other parts of a code base. And it's it's entirely clear how you're gonna need to change things. And

35:31 I said one core part of Crusher is this really suitable autocomplete experience. where you predict like the next set of things that you're gonna be doing across multiple files, across multiple places for that file. And You know, making models good at that use case. One, there's the speed component. Uh those models need to be really fast. They need to give you a completion within three hundred milliseconds.

35:50 There's also this cost component of we're running tons and tons and tons of molecules, you know, every keystroke we need to be Yeah, changing our creation for what you're gonna do next. And then it's also this really specialty use case of You need models that are really good not at completing the next token, uh just to like a generic text sequence.

36:07 But are really good at auto completing a series of deaths. um, you know, looking at what's changed within a code base and then clicking the next set of things that are gonna change. Yeah, both debuted and added and all of that. And we we found a a ton of success in training malls specifically for that task.

36:22 So that's w a place where you know no foundation models are involved. It's kind of our own thing. We don't have a lot of labeling or branding about this in the app. That that you know cores of

36:31 Uh you know, power's a very core part of cursure. And then Uh an you know, another set of places where user models are to to help things like Sonnet or Gemini or GPT. And Uh those sit both on the input of those big models and on the output.

36:45 On the input side of things, those models are searching throughout a code base. Try to figure out the parts of a code base to show. To one of these big malls? Um, you can kind of think about this as like a mini Google search that's specifically built. For finding the

36:59 you know, relevant parts of the code base to show one of these big models. And then on the output side of things, you know, we take the sketches of the changes that these models are suggesting you make with that code base. And then You know, we have models that then kind of fill in the details.

37:13 of like you know the high level thinking is done by these smartest models. They spend a few tokens on doing that. And then these smaller specialty incredibly fast models coupled with some inference tricks. Then take those high level uh changes and turn them actually into volcode discs. And so it's been super helpful for

37:28 Um pushing on on quality. In places where you need specialty tasks. And it's been super helpful for pushing on speech, which is such an important dimension of product quality for us. Uh two. This is so interesting.

37:40 I just had Kevin Wheel on the podcast, CPO of OpenAI and he calls this the ensemble of models. That's the same way they work. To use the best feature of each one and to your point the cost. advantages of using cheaper models. Uh these open these other models are they Based on like Llama and things like that, just open source models that you guys Plug into and build on.

38:00 Yeah, so uh again we try try to be very pragmatic about the place that we're gonna do this work and we don't wanna reinvent the wheel and so Um starting from the the very best. Um you know Uh pre trained models that exist out there.

38:13 Often open source ones. You know, s sometimes in collaboration with these big model providers that that don't share their weights out into the world. Um,'cause the thing we care about less is Yeah, the ability to lead read.

38:25 Line by line, you know. the you know, the the matrix of weights that then, you know, go to give you give you a certain output and it's we uh we just care about the ability to kind of Uh To train these things, to post train them.

38:36 And so uh by and large, by and large, yes, open source models, uh, you know, sometimes working with uh the closed first providers too to tune things. This leads to a a discussion that a lot of AI founders always think about and investors, which is motes. And defensibility in AI. Uh so it feels like one is

38:52 Custom models is is emote in the space. How do you just think about long term defensibility in the space, knowing there's other folks, as you said, launching constantly trying to take you trying to eat your lunch. I think that Yeah, there were ways to

39:06 Inertia and Um Yeah. T traditional notes, but

39:13 Uh I you know, I think by and large we're in a space where, you know It is incumbent on us. Uh to continue to try to build the best thing. And and and everyone. in this industry.

39:24 And I you know, I truly just think that the ceiling is so high that Yeah. Entrenchment you build. Um you can be leap frog.

39:33 And I think that this Uh. resembles markets that are maybe a little bit different from normal Software markets, normal enterprise uh markets of the past. You know, I think one that comes to mind is is the market for search engines at the end of nineteen ninety nine.

39:48 Are you know at the end of the the nineties and beginning of the two thousands? I think another market that comes to mind that resembles this market in many ways. It's actually just like the development of Um The perform computer and mini computers?

40:01 You know, in the seventies, eighties, nineties. And um I think that yes, m you know, in each of those markets, the ceiling was incredibly high. You know, it was possible to swish. you could keep getting value for like the incremental hour of a smart person's time, the incremental RD dollar for a really long time, you wouldn't run out of useful things to build. And Then you know, in in in search in particular, non link computer case.

40:23 Hading distribution was was helpful for making the product better too, and that you could tune Yeah, algorithms you could tune the learning. Based off of the the data, the feedback you're getting. From users.

40:35 And I think that you know, all of those dynamics exist exist in our market too. And so I think that's the thing. maybe the the sad sad truth for people like us, but then like the amazing truth for the world is I think that there are many leaf frogs that exist. Uh there's many, yeah.

40:48 more useful things to build. We're a long way away from where we can be in, you know, five, ten years and It's kind of incumbent on our state to keep that entry going. So what I'm hearing is this sounds like a lot more like a consumer sort of moat where it's just be the best thing consistently so that people stick with you versus creating lock in and things like that where they're just for like Salesforce where it's just contract for the entire company and you have to use this product.

41:10 Yeah, and I I think the the important thing to note is Yeah. If you're in a a space where like you kind of run out of useful things to do very quickly, then that's You know, that's not a great situation to be in. But if you're in a place where you know

41:22 Big investments in Um Yeah, ha having more and more great people working on the right path can keep giving you value. Then you can get kind of these And

41:33 you can kind of have, you know, you know, deeply work on the technology in the right direction and and get to a place where that is defensible. Uh but yes, it is it is yeah. I think there's there's a consumer like tendency to it and I really think it's just a you know, about building the best thing possible. Do you think in the future there's one winner in this space, or do you think it's gonna be a world of a number of

41:52 Products like this. I think the market is just so very big. And this is also one thing that um You know, you asked about the IDE thing early on. And One thing that I think a trip of some people that were thinking about the space is like

42:06 They looked at the IDE market of the past ten years. And they said, you know. Who's making money off of that, or is like you know, there's all these It's this super fragmented space where everyone kind of has their own thing. with the wrong configuration and

42:21 You know, there's one company that commercially Like actually makes money off of uh making great great editors. But like that company's only so big. And uh, you know.

42:31 Then Like the conclusion was it was gonna look like that in the future. And I think that the thing that people missed was That you know. There is only so much

42:40 You could do. Building editor on the twenty tenths. For cutters. And you know, the the company that made money off of editors was doing things like making it easy to navigate around a code base.

42:50 And you know, doing some some error checking and type checking for things. And you know, hav having good debugging tools, but like Which were all uh very useful, but I think that the the set of things you can build for programmers. I think the set of things you can build for knowledge workers in many different areas.

43:05 just goes very far and very deep. And I think that really kind of like the the problem in front of all of us is is like the automation of a lot of busy work and knowledge work and really changing All the areas not working for us to be um much available and more productive. So uh that that was all, you know, a long winded way to say I think the market's really, really big that we're in.

43:24 Uh I think it's much bigger than people have realized. Uh Yeah. uh you know building tools for developers in the past. And I think that there will be a a bunch of different solutions. I think that there will be one company and to be determined if it's gonna be us.

43:37 But I do think there will be one company. that builds the the general tool. That builds almost all the world software. And that will be a very, very generationally big business. But I think there will be a kind of niches you can occupy in

43:50 doing something for a particular segment of the market or for a very particular part of the software development life cycle. But the general like programm shifts from just writing formal programming languages to something way higher level. This is the application you you purchase and use to do that.

44:05 Uh. I think that there will be generally one winner there and and it will be a very big business. Juicy. Uh along those lines it's Interesting that Microsoft was actually like right at this set like at the center of this first

44:19 With an amazing product. Amazing distribution co pilot, you said was like the thing that got you over the hum of like, wow, there could be something really big here. And it doesn't feel like they're winning. It feels like they're falling Behind.

44:31 What do you think? What what do you think happened there? I think that there are like specific historical reasons. Why Co pilot might not have

44:39 Lived up right so far. uh lived up to the expectations that some people have for it. Then I think that there are structural reasons. I think the structural reason is

44:50 And to be clear, you know, Microsoft uh you know, in the Cobala case Uh obviously uh Big inspiration for our work. Um in in general I you know. Think they do lots of awesome things and

45:00 We're users of many Microsoft products. Um But I think that This is a market that's not super friendly to incumbents.

45:09 In that Um You know, a market that's friendly to incumbents might be one where there's only so much to do. it kind of gets commoditized fairly quickly and you can bundle in with other products. And where the ROI between

45:20 Yeah, different products is You know, quite quite small. And you know, in that case Perhaps it doesn't make sense to buy the innovative solution, it makes sense to just kind of buy the thing spundled in with other stuff. Another market that might be

45:33 Yeah. Particularly helpful for incumbents. It is one where there's you know From from the get go, it's just like you have your stuff in one place and it's like really, really excruciatingly Hard to switch.

45:43 And you know, for better or for worse, I think in in our case you can try out different tools and you can decide which product you think is better. And so that's not super friendly, uh, to June comments, and that's more friendly to whoever you think is gonna have the most innovative product. And then the specific historical reasons, like as I understand them, are the group of people that worked on the first version of Copilot.

46:04 have by and large gone on to do other things at other places. I think it's been a little hard to kind of coordinate among all the different departments and parties that might be involved in in making something. I like this. I'm gonna come back to Cursor.

46:16 A question I like to ask everyone that's building a tool like this. If you could. sit next to every new user that uses Cursor for the first time and just whisper a couple of tips in their ear. To be More successful, most successful with cursor.

46:29 What would be like one or two tips? I think right now and we'd want to fix this at a product level. A lot of being successful with Kircher is Kind of having a taste for like What the models can do

46:42 Both what complexity of a task they can handle and like kind of How much you need to specify. Yeah, things things to that. that model, but like having a d a taste for the quality of the model and where its gaps exist and what it can do and what it can't.

46:55 And Uh Right now we don't do a good job on the product of like You know. for educating people around that.

47:02 Um and maybe g giving people some swim lanes, giving people some guidelines. But so um to develop that taste. Um we give kind of Two two tips. So one is As mentioned before

47:13 Uh would bias less toward like Hey trying to have the model. Like Trying in one go.

47:20 To tell The model, hey, here's exactly what I want you to do. Then seeing the output and then either being disappointed or accepting the entire thing for an entire big task. Instead what I would do is I would chop things up into bits. And you can spend basically, you know, the same amount of time specifying things.

47:34 Overall. But chopped up more. So you're specifying a little bit, you're getting a little bit of work, you're specifying a little bit, getting a little bit of work. And you know, not doing as much the like, let's write a giant thing. Uh telling Maul exactly what to do. I think that will be a little bit of a recipe for disaster right now. Uh and so biasing toward chopping things up.

47:51 At the at the same time And sh it might make sense to do this on a side project and not under professional work. Yeah. I would encourage people to especially, you know Developers who are kinda uh used to

48:02 existing workflows for building software. Yeah, I would encourage people to Explicitly try to fall on their face? And try to discover the limits. Uh

48:11 Uh what these models can do. By f you know, being ambitious and like kind of a a safe environment, uh like perhaps a side project. And and trying to kind of go hand to hand you can use AI to the fullest. Because, you know, sometimes we do run or a lot of the time we run into people Um

48:26 who haven't given the AI yet a a fair shake. And are kind of underestimating its abilities. So Generally biasing towards chopping things up and making things smaller. But like to discover the limits of what you can do there.

48:36 Like explicitly just kind of try to go for broke. In a safe environment and you know. Get get a taste for it. You might be surprised in some of the places where the model doesn't break. What I'm essentially hearing is

48:47 kinda build a gut feeling of what the model. can do and how far it can take an idea. versus just kind of guiding it along. And I bet that you need to rebuild this gut every time there's a new model launch, like when it's on it. I don't know, four point oh comes out, you have to kind of do this again, is that generally right?

49:04 Yes. Uh it's not You know, for the past few years it hasn't been as big as like, I think the The first kind of experience people have had with some of these big models.

49:14 But um yeah. You know, it this is also a a problem we would hope to solve much better. Just for users and T Bury off of them. But yeah, e each of these things have Slightly different quirks and different personalities.

49:26 Kind of along these lines, something that people are always debating. Tools like Cursor, they More helpful to junior engineers or are they more helpful to senior engineers? Do they make senior engineers ten X better? Do they make junior engineers more like senior engineers? Where do you think most of the Who do you think benefits most today from Cursor?

49:43 I think across the board. Uh both of these cohorts benefit in big ways. It's a little hard to say on the relative ranking. I will say the Fall into different anti patterns.

49:54 So I would the junior engineers we see going a little too wholesale relying on AI for everything. And We're not yet in a place where you can kind of do that end to end on a professional tool, you know, working with tens, hundreds of other people within a long lib code base.

50:12 And then the senior engineers For many folks, it's not true for all. And we actually Uh Uh Often.

50:20 you know, one of the ways these tools are adopted is there's developer experience teams within companies often those are stopped by incredibly senior s you know, senior people. Because often those are people who you know, are building tools to make the rest of the engineers within an organization more productive. And we've seen some very, very on you know boundary pushing.

50:37 Kind of uh Uh Yeah, like we've seen s people who are, you know, on on the the front lines of like really trying adopt the ta technology as much as possible there. But by and large I would say On average, as group, the senior engineers.

50:50 Underrated. what AI can do for them and stick to their existing workflows. And so the relative ranking's a little hard. I think they both have they they fall into to different Different anti patterns. Uh but they both by and large get get big benefits.

51:03 Those tools. That makes absolute sense. Uh I love that it's like two ends of the spectrum, like expect too much, don't expect enough and It's like the uh The three bears? Is that the allegory?

51:15 Yeah. Yeah, okay. But not staff, you know, right in right in the middle. Um Interesting. Okay. Just a couple more questions. Um

51:28 What's something that you wish you knew before you got into this role? If you could go back to Michael at the beginning of Cursor, which was not that long ago. And you could give him some advice. What's something that you would tell him? The tough thing with this is Feels like so much of the

51:42 Uh The hard one knowledge is tacit. And a bit hard to communicate for a blade. And uh The sad

51:51 fact of life feels like for you know for some areas of human endeavor, like you kind of do need to fall on your face to Uh Either either need to fall on your face to to learn the correct thing or you need to be kind of around someone who's a great example of kind of excellence in the thing. And One area where we I felt this is

52:09 Is is higher. Um I think that Uh We actually were

52:14 So we try to be incredibly patient on the higher front. Um It was really important to us. That

52:23 Yeah, both for personal reasons and also for I think actually for the company's strategy. having a world class group of engineers and researchers to work on. Cursure with us? was going to be incredibly important.

52:35 Also getting people who Fit. You know, a a sort of mix of Yeah. Intellectual curiosity.

52:42 And experimentation because there can be so many new things we need to build. And then also kind of an intellectual honesty. And maybe micro pessimism and bluntness. Because you know, with all the noise and Yeah, especially as the company's grown and

52:54 And the business has grown. Yeah, keeping a level head, I think, is an incredibly important too. Um but getting the right group of people into the company Yeah. Yeah, the th the thing that maybe more than anything else

53:06 apart from apart from building the product, we really, really Uh Uh You know, fussed over. And uh

53:14 Yeah, I we actually waited a long time to grow the team because of that. And I think that most Yeah, many people you hear hire too fast. Think we actually hired Too slow to begin with. I think it could have been remedied. I think we could have been better at it.

53:27 And um You know, the the method of Uh uh of recruiting that we ended up uh eventually been working really well for us, which which isn't that novel of like going after people that we think are really world class and like recruiting them over the course of in some cases mo many years. Uh I ended up working for us in the end, but I I don't think we were very good at it to begin with.

53:49 And so I think that there were hard won lessons around Both who was the right profile. Like who actually meets us on the team, like what did what did greatness look like? Uh and then how to, you know. Um talk with someone.

54:00 Um about about the opportunity and you know Get them excited if they really weren't looking for anything. Um There there were lots of kind of uh learnings there about how to do that well.

54:10 Um and that's because of Philosoph. What are some of those learnings for folks that are, you know? hiring right now. What's something you missed or or learned? I think, you know, to start with

54:19 Uh we maybe We actually biased a little bit. Too much towards Um

54:28 well known school very young had done the things that were like You know, high credential. Um in those well known school environments. And

54:38 Um And actually like you know I think found uh were lucky early on to find a lot of you know Uh To find uh fantastic people who are willing to

54:48 You know? To do this with us. Uh who were who were later career. And so yeah, I think we should kind of spend a a bunch of time on maybe a little bit the the wrong profile to begin with. And part of that was a seniority thing. Part of that was like, you know, kind of an interest and experience thing too.

55:02 Uh, we have hired people who are excellent, excellent, excellent and very young. But they maybe look uh in some cases slightly different from Yeah. being straight out of central casting. Yeah, another lesson is just like

55:13 We very much evolved our interview loop. And so now we Uh You know, we have

55:21 Like a hand rolled set of interview questions and then you know, kind of core to our Um Court of how we interview too is is actually we had people on site for two days and do do a project with us, a work test project. And um that has worked really well, but increasingly you're finding that.

55:37 And then yeah, I think how to to learn about what people are interested in. And you know, put our best foot forward and and letting them know about the opportunity when they're really not looking for anything and have those conversations. Uh there's definitely been you know, gotten gotten better at that over time.

55:53 Do you have a favorite interview question they like to ask? I think this two day work test. Which we thought would not scale past a few people has been has had surprising staying power. And the great thing about it is

56:05 It lets someone go end to end on it, like a a a real project. It's it's not, you know, word to we use, it's kind of a can Candlist of projects. Um, but it gives you two days of seeing like a real work product. And um It doesn't have to be incredibly time intensive on the teaching front uh time. You know, you can take the time you would spend in like a half day or one day on site and you kind of spread it out over those two days and

56:28 Give someone a lot of time to do. On their projects. And so that can actually help it help it scale. Uh and then it really helps you.

56:38 It helps to enforce You know, do you want to be around this person type test. Um because you are around this person. Uh, you know, for a few days and so the bunch of meals with them.

56:49 And uh so that one we didn't expect that one to stick around, but that has been really, really important to our value to process. And then also important to getting people excited at the Especially the very early stages of the company. Because before people are using the product and know about it. And yeah, when the the product is comparatively like not very good. Really the only thing you have going for you is you know, a team of people that you know some people find special and and want to be around.

57:13 And yeah, the two A It would would give us a chance to just like Yeah, have this person Uh meet us and uh in some cases hopefully get get convinced that they they want to throw in with us.

57:23 And so yeah, that one that one was unexpected. Not exactly an interview question, but kind of like a you know a a forward interview. The ultimate interview. So just to be very clear about what you're describing, it's a you give them an assignment like build this feature in our actual code base, work with the team to Uh how would it end ship it? Uh yes. Not not like so we don't use the IP.

57:43 Not should end to end. But yeah, it's like a mock like Yeah, very often in our code base. Here's a real mini two day project. You're gonna do it hand to hand. largely being left alone. Yep, there's there's collaboration too.

57:55 Uh and then you know, we're we're a pretty improving company, so And in almost all cases, yeah, it's actually just sitting in office with us, too. And you've been saying that this has scaled to even today's how how big are you guys at this point? Uh, so we are going on sixty people. So small.

58:11 For the scale and impact. That's I was I was thinking it'd be a lot larger than that. Yeah. And I imagine the largest percentage is engineers. Yeah, the thing that's more than anything And to be clear, you know

58:24 the the work ahead of us is is is building a group of people that is is bigger And awesome and can continue to make the the product better and the service we give to customers better. And so you don't plan to stay that small. Uh for longer. We wouldn't wouldn't hope so.

58:38 But uh yeah, uh part part of the reason that um that number is is small is uh the percentage of of engineering and and research and design is very high within the company and So uh many software companies when they have, you know, roughly forty engineers Would be over a hundred people.

58:53 because there's lots of operational work and often they're very, very sales led from the get go. Uh and that's just Labor intensive. And you know, we started from a place of being like incredibly lean and product led. And like, you know, we now serve lots of market customers and it built that out. But

59:08 You know, there's much more to do that. Question A. Wanted to ask you. There's so much happening in AI. There's things launching every there's like newsletters. Like many newsletters whose entire function is to tell you what is happening in AI every single day.

59:22 running a company that's at the center, kind of the white hot center of the space. How do you Stay focused and how do you help your team stay focused and heads down and just build and not get distracted by all these shiny things. You know, I think hiring is a big part of it. And if you get people with uh The right attitude. Um

59:40 And you know all of this should be asterisked in Like Yeah, I think we're doing well there. I think that like Yeah, we'd probably be doing better there too.

59:49 And um you know, it's something that we should probably talk even more about as a company. But I think that you know hiring people with the right disposition, you know, people who are less Focused on External validation, more focused on building something really great, more focused on doing really high quality work.

1:00:04 And people who are just generally kind of uh level level headed and Uh maybe like you said, the highs aren't very high, the lows aren't very low. I think hiring can can get you through a lot here and Think that's that's actually like, you know, a learning throughout the company.

1:00:17 Is that Yeah. For any You you need process, you need hierarchy, you need lots of things. But for for any kind of organizational tool that you're intro introducing into a company.

1:00:27 You know. the the result you're looking to get from that tool. Also Yeah, you can go pretty far on like hiring people. with the right behaviors that you want like, you know, to result from that or organizational thing. And

1:00:39 Yeah, the specific example that comes to mind is We've been able to get away with not eight Ton of process yet on the engineering front. I think we need a little bit more process. But for our size not a time process.

1:00:51 By hiring people who I think are really excellent. Yeah, the one is you know, hiring people are level headed. I think too is just talking about it a lot. I think three is hopefully leading by example. Uh and yeah, for us personally, you know.

1:01:03 We've you know, since twenty twenty one, twenty twenty two, been professionally working on I'm this and work on I. And we've just seen a sea change of The comings and goings of

1:01:14 um various technologies and ideas of Yeah, if you're to transport yourself back to it. End of twenty twenty one, beginning of twenty twenty two. This is Chief G Three. You know, instruction PC doesn't exist. There's no Dali, there's no stable diffusion.

1:01:24 And then You know, we've gone through all of those image technologies existing, ChatGBT and that rise. And you know, G four, all these new models, all these different modalities, all the video stuff. And

1:01:39 Only Yeah, a very small. number of these things really kind of affect affect the business. So I think We've kind of just

1:01:46 And built up a little bit of an immune system and kind of know know when when an event comes around that actually It's really gonna matter for us. And this is, you know, this dynamic too of there being lots and lots and lots of Chatter.

1:01:58 But then maybe Only a few things that really matter. I think has been mirrored in AI over the last decade. Where um There have been

1:02:07 So many papers. on deep learning. in academia. So many people are in academia. Then the amazing thing is

1:02:15 There it Really a lot of I mean. A lot of the progress of AI can be attributed to some very simple, elegant ideas. in the vast majority of of ideas that have been put out there.

1:02:26 haven't had staying power and haven't mattered a ton. And so the dynamics a little bit mirrored in kind of the evolution of deep learning as a field overall. Last question. What do you think people still most misunderstand or maybe don't fully grasp a w about where things are heading.

1:02:41 With AI. In building And the way the world will change. People are Still

1:02:48 A little bit, you know, occupy too much. either end of a spectrum of Uh, you know, it's all gonna happen very fast and you know, the this is all You know, bluster and Type and super snake well.

1:03:02 And you know, I think we're in the middle of Eh technology shift that's gonna be incredibly consequential. It's gonna be more Casapaturaline. Yeah. Any shift in tech.

1:03:13 That we've seen since since the advent of computers. And I think it's gonna take a while. And I think it's gonna be a multi decade thing. And I think many different groups will be consequential in pushing it forward. And um you know, to get to a world where

1:03:27 Computers can increasingly do more and more and more for us. There's all of these independent problems. That need to be knocked down and progress needs to be made on them and Some of those are on the the science side of things of getting these models to understand different types of data.

1:03:40 Be faster, cheaper, smarter. You know, conform to the Uh the modalities that we care about, you know, take actions in the real world. And then some of it's on like how we're gonna work with that.

1:03:50 Yeah, what's the You know, what's the experience a human should actually be seeing and and controlling on a computer. And working with these things. But I think it's gonna you know it's

1:03:59 Gonna take decades, I think that there's gonna be lots of amazing work to do. I think that also Yeah, one of the most And like a pattern of a group that I think will be especially important here.

1:04:10 You know. Not not to talk our own book, but I think his like, you know The company that works on Um automating and augmenting a particular particular area of knowledge work builds the both the technology under you know under the surface for that.

1:04:24 Um integrating the best parts from providers, sometimes doing it in house. And then also builds the you know, the product experience for that. I think people who do that and you know We're doing it in trying to do it in software. People do that in other areas. I think those folks will be really, really, really consequential.

1:04:38 Not just for like, you know, the end value that users see. But then I think as they get to scale, they'll be really important for pushing forward. Mm-hmm. the you know the technology. Because I think they'll be able to build

1:04:48 you know, the most successful of them will be able to build very, very big businesses. Um And um Yeah, so excited to see the rise of You know.

1:04:57 Other companies like that in other areas. I know you guys are hiring uh For folks that are interested in hey, I wanna go work here and build this sort of stuff, what kind of roles are you looking for right now, anyone? Specifically you're trying to any roles you're most excited about filling? ASAP. What should people know if they're curious?

1:05:12 There are so many things that this group of people need to do that like we are not yet equipped to do. And so uh Yeah. kind of generic across the board, first of all. And so If you uh

1:05:24 don't think we have a role for something. Maybe if you reach out that that won't actually be the case. Um and maybe we can actually learn from you. And kind of decide that we we need something that we weren't yet aware of. But um you know, by and large, I think that you know T two of two of the most important things for us.

1:05:39 have the best product in the space. And then grow it. And we're kind of in this land grab mode where Almost everyone in the world

1:05:48 is either using no tool like ours or they're using One that's maybe Developing less quickly. And uh so so growing growing um curvature is is is a big goal and Um

1:06:01 Uh I would say Yeah, uh especially always on the hunt. for for folks uh who Excellent engineers, designers, researchers. Um

1:06:10 But then folks in uh all on across the business side too. I can't help but ask this question now that you talk about engineers, there's kind of this question of just like, you know, code's gonna write up all our co uh. AI is gonna write all our code, but everyone's still hiring engineers like crazy. All the foundational models, so many are good. Uh yeah. Do you think there's gonna be an inflection point of like engineering roles start to kinda slow down? Uh I know this is like a big question, but just it's

1:06:36 Do you see engineers being more and more needed across all these companies, or do you think At some point. There's all these Cursor agents running. Building for us.

1:06:45 Again, we we kind of have the view that like There's this You know. Both long messy middle. Uh

1:06:53 Uh Yeah, it it not jumping to a just like you step back and you ask for all your stuff to be done and you have your engineering department. And you know, very much. Like

1:07:02 You want to evolve from programming is that exists today. We want humans to be in the driver's seat. And you know, we think even in the end state, like that's you know Giving folks control over everything is is really important. And you will need professionals to do that and kinda decide what the the software looks like.

1:07:18 So both both I think that yes, like you know Uh like you know, engineers Uh are are definitely needed. Uh, I think that engineers will be able to do much more. I think the demand for software is very lasting, which is, you know, not the most novel thing, but I think it's

1:07:31 It's kind of crazy to think about. How expensive And Libre intensive. It is

1:07:39 pretty simple and easy to specify, or it would look like it to the outside observer. And you know, just how hard those things are to do right now. And so if you can you know, all of the stuff that exists right now that's you know justified by the cost and demand that we have now, if you could bring that down by orders, you'd I think you would have tons and tons and tons of more stuff that we could do our computers, tons of more tools.

1:08:00 And You know, I've I've felt this where, you know, one of my early jobs actually was Working for a biotechnology company. And it was building internal tools for them. And the off the shelf tools that existed were horrible.

1:08:13 And did not fit their use case at all. And then the internal tools I was building There was definitely a ton of demand there. Uh for things that could be built. And you know, that far outstriped just the things that I could I could build in the time that I was with them. But yes, I think that uh

1:08:27 It's still so, you know. The physic of working on computers are so great it should be able to you should be able to kind of basically just move everything around. Do everything that you want to do. There's still so much friction.

1:08:38 I think that there's much more demand for the software then. uh for software than what we can build today. With you. things costing like a blockbuster movie to make kind of simple productivity software. And so I think long into the future, yes, there will actually be more demand for engineers. Is there anything that

1:08:52 We didn't cover that you wanted to mention any last nugget of wisdom you wanted to leave listeners with. You could also say no because we've done a lot. We think a lot about Uh how How you set up a a team.

1:09:06 To Be able to make new stuff. In addition to like continuing to improve the stuff that you have right now. And if we're to be successful, like

1:09:15 Yeah. I D is gonna have to change a ton, what for feel like Looks like to change the time going into the future. And um Yeah, if you look around Uh the

1:09:25 The companies we respect. Uh there are definitely examples of companies that have continued to really wa like, you know. ride the wave of many leapfroths and continue to kind of actually push the frontier. But you know. Uh

1:09:37 There Th they're kinda rare too. Uh, like it's a hard thing to do. And um So you know part of that is is just kind of

1:09:45 thinking about the thing and trying to reflect on it, you know, in our our A days and you know The first principle side of things. Part of it is also, you know, trying to get in and and study past examples of of greatness here. And Um

1:09:58 You know, that that's that's something that we think about a lot, too. Yeah, the what you just told is if we were before we started recording, you had all these books behind you and I was like, What's that over there? Like the history of some. old computer company that was influential in a lot of ways that I've never heard of. Uh, and I think that says a lot about you of where a lot of this innovation comes from is studying the past and studying history and

1:10:17 What's worked and what hasn't. Okay. Uh, or can folks find you online if they want to reach out and maybe apply. You said that there may be roles you they may not even be aware of. Well, where do they go find that? And then how can listeners be useful to you? Yeah, I you know, if if folks are um you know interested in working on this stuff, would would love to speak and

1:10:34 uh they can find uh if they go to Chris dot com they can kind of both find the product and find out how to reach us. So Easy. Michael. Thank you so much for being here. This was incredible. Wonderful. Thank you.

1:10:46 Bye, everyone. Thank you so much for listening. If you found this valuable, you can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app. Also, please consider giving us a rating or leaving a review, as that really helps other listeners find the podcast. You can find all past episodes or learn more about the show at Lenny's Podcast.com. See you in the next episode.

What you see above is a preview of the first minutes. One unlock costs 10 credits and covers this episode forever: full segment and word-level timestamps on this page, plus .txt, .srt, .vtt and word-level JSON downloads, as many times as you like.