“Engineers are becoming sorcerers” | The future of software development with OpenAI’s Sherwin Wu

0:00 95% of engineers use Codex. 100% of our PRs are reviewed by Codex. For engineers, I don't know what job has changed more in the past couple years. Engineers are becoming tech leads. They're managing fleets and fleets of agents. It literally feels like we're wizards casting all these spells, and these spells are kind of like going out and doing things for you. What do you think people aren't pricing in yet? In our third order effects of the one person billion dollar startup. To enable a one person billion dollar startup, there might be a hundred other small startups building bespoke software. So I think we might actually enter into a golden age of B2B SaaS. I've been hearing more and more there's this stress people feel when their agents aren't working. There's a team that's actually doing an experiment right now with an OpenAI where they are maintaining a 100% code. Codex written code base. They run into the exact problems that you're describing. And so usually you're like, all right, I'll roll up my sleeves and figure it out. This team doesn't have that escape hatch. You've shared that listening to customers is not always the right strategy in AI. The field and the models themselves are just changing so, so quickly. They tend to like disrupt themselves. The models will eat your scaffolding for breakfast. advice to folks that are like, okay, I don't wanna miss the boat. Make sure you're building for where the models are going and not where they are today. There's a quote from Kevin Whale, our VP of science here, and he likes saying this is the worst the models will ever be. Today my guest is Sherwin Wu, head of engineering for OpenAI's API and developer platform.

1:17 Considering that essentially every AI startup integrates with OpenAI's APIs, Sherwin has an incredibly unique and broad view into what is going on and where things are heading. Let's get into it after a short word from our wonderful sponsors. Today's episode is brought to you by DX, the developer intelligence platform designed by leading researchers. To thrive in the AI era, organizations need to adapt quickly. But many organization leaders struggle to answer pressing questions like, which tools are working, how are they being used? What's actually driving value? DX provides the data and insights that leaders need to navigate this shift.

1:53 With DX, companies like Dropbox, Booking.com, Adion, and Intercom get a deep understanding of how AI is providing value to their developers and what impact AI is having on engineering productivity. To learn more, visit DX's website at get dx.com slash Lenny. That's get dx dot com slash Lenny. Applications break in all kinds of ways, crashes, slowdowns, regressions, and the stuff that you only see once real users show up. Sentry catches it all. See what happened, where and why. Down to the commit that introduced the error.

2:27 the developer who shipped it, and the exact line of code all in one connected view. I've definitely tried the five tabs and slack thread approach to debugging. This is better. Sentry shows you how the request moved. what ran, what slowed down, and what users saw. Seer, Sentry's AI debugging agent, takes it from there. It uses all of that Sentry context to tell you the root cause, suggest a fix, and even opens a PR for you. It also reviews your PRs and flags any braking changes with fixes ready to go.

2:59 Try Sentry and Sear for free at Sentry.io slash Lenny and use code Lenny for$100 in Sentry Credits. That's S C N T R Y dot Io slash Lenny. Sherwin, thank you so much for being here and welcome to the podcast. Thank you, thank you for having me. I wanna start with what's feeling like a barometer of progress in AI, especially in engineering. What percentage of your code, if you even write code anymore, and your team's code

3:31 Is written by AI at this point. I do write code occasionally now, still. I actually say for managers like myself, it's way easier to use these AI tools. uh than to manually code at this point. And so I know for myself and some of the other EMs. Engineering managers at OpenAI?

3:47 Uh All of our code is written by by Codex, uh, at this point. But more broadly, there's just been this there's just so much energy. There's like a tangible energy internally around just how far these tools have gotten and how good Codex is a tool has gotten for us. And uh it's it's a little hard for us to exactly measure. how much of the code is is written because

4:06 the vast majority of it, I'd say like close to a hundred percent is is usually generated by AI first. Uh what we do track though is is you know, at this point, uh the vast majority of engineers use codecs on a daily basis. So ninety five percent uh of engineers um use codex. Um one hundred percent of our PRs are reviewed by Codex daily as well. So basically any code that goes into production that's merged in, Codex kind of has its eyes on and uh suggests improvements, suggests changes uh uh in the PRs. And so uh that's kind of what we're seeing internally, but by and large, the most exciting is just the energy that that there that that there is. Um

4:40 Another observation that we've had is uh engineers who tend to use codex uh more. Uh open way more PRs. So uh they're actually opening seventy percent more PRs uh and uh than than the engineers who aren't using codex as much. Uh and the gap is widening. So I feel like, you know, the people who are opening more PRs um are starting to you know learn how to use the tool more and more, get more efficient and that seventy percent gap keeps uh uh growing over time. And so might have actually increased since I last looked at the at the number.

5:07 Okay, so just to make sure we hear what you're saying, you're saying all of the code of these ninety five percent uh engineers at at at OpenAI is written by AI. It's written and then they review it. Yep. Yeah. It's it's like crazy that that's almost like not crazy anymore. That we're just like getting used to this.

5:26 I think there's still some getting used to, to be clear. Uh there's also I think some You know, uh engineers who I think trust uh Codex a little bit less, but um basically every day I talk to someone who who Uh Is blown away by something that I can do and and kind of like the their bar of of trust kind of uh Uh or like how much they trust the model to do on its own goes up over and over.

5:48 Uh over time and There's a quote from Kevin Will, our our um uh VP of of science here. He likes saying this is the worst the models will ever be. And so this is the worst that the models ever be for software engineering as well. And so over time we just see people trusting it more and more. And then we'll see the models get better and better as well. Yeah, Kevin Wheel, former podcast guest, uh he he said exactly that line on this podcast in a few times.

6:09 Yeah, uh Peter the clawed bot slash maltbot slash open claw is what it's called now. Uh developer uh recently shared that he uses codex for his work and he feels like any time it does things he just trusts that it has done the right job and he's just like almost certain he could just commit it to master and it'll be great. Yeah, yeah, he's a great um user of Codex. I know he's in close touch with the team, gives us great feedback. Um

6:34 Uh, not surprised that he uses it. I mean uh Sorry, it's called open And then I saw that this mor I mean this is very recent, but this morning I think l most Uh kind of like uh Uh with Sherry as well and seeing all of the uh AI agents talk to each other is pretty uh pretty surreal. It's basically her is happening in real life, is what I'm hearing. Yeah.

6:54 So just like coming back to this crazy moment we are living through four engineers. In particular We've gone from You write every line of code to now AI is writing all of your code. I don't know what job has changed more in the past couple years.

7:09 Like job that we didn't expect to change this much, where just like the job of an engineer is So different. In the entire lifespan of an engineer, like in the past couple of years, it's now shifted. Two I don't write any more code. How do you imagine the role of an engineer and the job of a software engineer?

7:25 looks in the next couple of years, just like what is that job? Yeah, it's I mean, it's always been really cool to see. Um Uh, and it's part of where the excitement is because uh like the job is likely gonna change pretty significantly over the next one or two years. It kind of feels like we're still figuring things out though. And so there's like this excitement I know, especially from some of the software engineers, of like, We're in this rare moment, you know, maybe over the next twelve to twenty four months where we'll kinda gonna figure things out ourselves and set our standards for

7:51 ourselves. In terms of where I see uh I see this moving, so I think there's the common thing that everyone's saying, which is uh you know, people are generally Like I see engineers are becoming tech leads. They're basically like managers now. They're managing fleets and fleets of agents. Um, I know many of the engineers on my team Basically have like ten to twenty

8:11 uh threads kind of being pulled on at the same time. Obviously not active running codex uh jobs. But uh just a lot of parallel threads. They're checking in on what they're doing, they're steering the agents uh and codex and and and and giving it feedback. And so Their job has kind of really changed from Just writing the code itself into being almost like a manager.

8:31 In terms of where I think this will go one to two years from now. So one uh kind of metaphor that that I kinda always come back to here is is actually from this uh It's from this uh programming textbook uh that I read back in college called Sickbee. I don't know if you've heard of it. uh structure and uh interpretation of computer programs. So S I S I C P

8:51 Um At at MIT it was it was really popular and and it was actually used as the uh uh introductory it was the textbook for the intro programming course for A very long time. Um and it kinda has this cult following. Um it it teaches you programming, uh it teaches you a dialect of list. I call scheme.

9:09 Uh and so it like introduces you to like functional programming. It's like very mind mind opening that way. But the thing that was memorable for me about that book, so I I I kinda read it in college. Um the very beginning of it kind of describes programming as a discipline. And draws this metaphor to basically like sorcery. Like it says like software engineers are like wizards, and you're like you're like programming languages are like incantations, and you're like,

9:32 You know, you're you're saying you're issuing me spells. And these spells are kinda like going out and doing things for you and and the challenge is like what incantation do you have to say to make the The program do what you want. And this book was written in nineteen eighty, so this is this is a while ago. And I think that metaphor has actually s like kind of persisted over time. And I think it's actually playing out as we move into this uh

9:52 new era of vibe coding or just like what software engineering will look like because Programming languages were basically these incantations. They've changed over time. And the challenge is always and and and the trend has been that these it's been easier and easier to kinda get them the the the computer to do what you want uh via programming. And I think The current wave of AI is is

10:10 probably the next stage of that evolution. It is now literally incantations because you can tell, you know, your uh you can tell codex, you can tell cursor. uh exactly what you want to do. And then it'll all go do it for you. Uh and I particularly like the wizard and like the the the Cersei analogy'cause uh I think our current state is is starting to move towards

10:28 Kinda like the the sorcerer's apprentice, uh, you know, from Fantasia. Uh where Mickey Mouse is like, you know. He find the sorcerer's hat and he tries to do all these things and I actually think it's a really apt analogy because one Uh it's just it's really powerful now. These incantations you can do can is is extremely high leverage.

10:45 But you kinda have to know what you're doing. Right? Like in Sorcerer's Apprentice, the whole plot is like Mickey goes wild, the the brooms like go crazy and everything's flooding. I think he literally Sets the like sets the uh the brooms off on a task and then goes to sleep. Uh and and so you know, it's like bi coding at its at its at its greatest.

11:02 And then eventually the the the old sorcerer comes back and like cleans everything up. And um you know, when when I see engineers kind of like doing these these these these twenty different uh codex threads at a time. There there is some skill and there's some seniority and like, you know, uh um a lot of thought that needs to go into this because you want to make sure that the the the models aren't going off the rails. Uh you definitely don't want to just like completely

11:24 uh go away and and you know, like ignore ignore the thing. But it's also extremely high leverage. Like, you know, a a a a very senior engineer who's who's really prolif uh Proficient with these tools. Uh can now just do way more things via Oh what they're doing.

11:38 And I think this is also what makes it fun. Like it literally feels like we're wizards now. You know, we feels like we're closer to to to to having uh uh uh to to making making it feel like this like magical experience where we're you know, casting all these spells and having software do all these things for you. I was thinking of the Sorcerer's Apprentice exactly as the metaphor as you were describing that, so I'm glad you went there. Uh a previous podcast guest described it as you have a genie that you can that grants you wishes. And it's a useful frame because you have to be very clear about the wish you want. Like if you want to be big. How it might be like the monkey's paw type thing where, you know, it's like you call what you want, but what are the side effects?

12:12 Um yeah, yeah, I think that and the analogy is great and Um yeah, the crazy thing for me is just the staying power of that book. Sick be like It's called the Wizard book. You know, people call it the Wizard book because that is the metaphor that they kind of weave throughout the the book and Um we're we've basically reached that point now, which is which is which is really cool. There's two kind of threads I want to follow here. One is I've been hearing more and more there's this like stress that people feel when their agents aren't working.

12:35 You fire off all these, you know, codex agents and then you have to keep stay on top of them. Oh shit, one's not working, I'm wasting time. Uh do you do you feel that do you feel that across your team at all? Yeah, yeah, I mean it happens all the time. And I actually think like th this is where the interesting part of all of this lies right now because these models aren't perfect. These tools aren't perfect. And we're still trying to figure out how to best interact with these uh with with with codecs or with these AI agents to to get work done. We see this come up all the time. There's a particularly

13:03 interesting team that we have internally. So there's a team that that's actually doing an experiment right now uh with an open AI where they are basically Maintaining a one hundred percent codex written Code base. Uh so you know, like You know, uh uh some you know, you you'll have the AI write code, but you'll obviously end up like rewriting a lot of it and and and you might need to like double check and change things.

13:23 But this team is just fully codex filled and just like leaning in entirely. And they run into the exact problems that you're describing, which is like, you know their challenge is, you know, uh, you know, I want to get this thing, this feature built. But I can't get the Asian to do it. And so usually there's an escape hatch where, you know, then you're like, All right, I'll roll up my sleeves and like figure it out. And then instead of using codex, I might use like tab complete and and cursor and and and things like that.

13:45 But this team uh uh uh for the experiment this team doesn't have that escape hatch. Uh and so then the challenge like how do I get the the the the agent to to to do this? And Um, I actually think we're gonna be publishing a blog post from some of our learnings here. Um but a lot of fascinating like paradigms and best practices are falling out of this.

14:03 Um, one interesting thing that we've noticed, I I don't know if this is what you you kinda feel, but we definitely feel it here is A lot of the time uh when the coding aid is not doing what you want. It's usually a problem with context and just like information that you've given it. It's just you've either under specified or there's just not enough information around how to do something. Available to the agent. Available to codex.

14:25 Uh and so uh when when you have to solve it through through that, uh the challenge is then to to to add documentation. And actually work around this this limitation. And basically encode more tribal knowledge that's in your head somehow into the code base, either via you know, code comments itself or code structure itself.

14:43 Or the uh text files like you know.md files, skills, any type of additional resources within the repository. So that the model can um uh can better do its task. There's a whole bunch of other learnings from this uh this group, which I think is fascinating uh to to explore. But yeah, kind of giving removing that escape hatch of of no longer using the AI has allowed them to start piecing together a lot of the problems that

15:06 Uh, we'll have to solve if you really want to lean into agents. Another uh issue people ran into. You talked about how people are shipping PRs like crazy, a lot more PRs if they're working with AI. Uh obviously code review is becoming a a bigger challenge. Is there anything you've figured out in your team to help speed that up to make that scale as and not just create this terrible job for people where they're just sitting there reviewing PRs all day?

15:27 Yeah, I mean one thing is codex reviews one hundred percent of all RPRs at this point. And so Uh I actually think so One one really interesting thing that's happened is the things that tend to we hand we tend to hand to the models immediately tend to be the things that annoy us or like are the most boring parts of uh software engineering. It's also why it's more fun now because we get to do more You know, more of the fun things.

15:50 Um, for me, um, speaking more for myself. I really hated code reviews. It was like one of the worst things for me. And then I remember on in my first job uh uh out of college, uh it was at it was at Quora. Um, I owned I was working on the newsfeed. And so I own the code for the Neesfeed.

16:06 And so I was a reviewer for Newsfeed, and uh it was just like the central piece of code that everyone would touch. And so I would just Every morning I'd log in and be like Like twenty to thirty code reviews. Like, Oh my goodness, I gotta like, you know, get through all of these. Um I would procrastinate and then it grows to like fifty and so there's just like a A lot of cover views.

16:24 Codex is really good at reviewing code. Uh so actually one thing that we've noticed that five two and five two four has gotten extremely strongly adept at. is reviewing code and especially when you kind of steer it in the right direction. And so uh for code reviews, yeah, we create a lot of PRs, but Codex reviews all of them. And it makes you know code reviews go from

16:41 a you know, I don't know, ten, fifteen minute task to sometimes even just like a two to three minute task because you have a uh a bunch of suggestions uh already Already baked in. Uh a lot of the times people will uh especially for small PRs, like you you actually don't even need people to review. You kind of trust codex in this way. Um, the original author kind of looks at codex. It is, you know, the the benefit of coder view is to have a second pair of eyes to make sure that you're not doing anything dumb.

17:04 Codex is a pretty smart second pair of eyes at this point. And so Uh that's something that that we've heavily leaned into. Um The general CI process and like the post uh kind of push and like deployment processes also have been heavily automated via codex internally at this point. If you talk to a lot of engineers, the thing that annoys me the most is after you've written your beautiful code, like how do you get it into production? You know, you got to

17:25 You gotta run through all these tests, you gotta like, you know, limp errors, you gotta all have the code review. Um, there's a lot of automated stuff you can do with codex. And so we've actually built some tools internally that that help automate that process, automate the lint. Yeah, if there's like a linked error, it's a very easy codex fix. Uh and then just it could just patch it and then kind of restart the CI process. Um so all of that is it w we're trying to collapse as as into as as little work for an engine as possible.

17:48 Which and and the byproduct of which is uh um uh they can they can now merge and push out a lot more peers. Codec writing the code, codec reviewing its own code. I'm curious if you are open to using other models to review your model's work. Is that is that a path or is it just it's good enough, we don't need anything else? So I will say there's there's definitely a circular thing here and like going back to Sourcer's Apprentice, like you wanna make sure you're not letting the brooms go crazy here.

18:09 Um and so You know we we're very thoughtful, I'd say, around which PRs kind of are completely just codex uh reviewed. Most people still obviously take a look at their PRs, uh, and so it it's not like it's going to zero. It's more like going from, you know, hundred percent attention to like thirty percent attention, which which just helps

18:28 Things pushed through. Uh in terms of like multiple models, uh so we we obviously test a lot of models internally, and so we have a lot of those. Um, we use uh external models less. Um it's we we think it's important to kinda dog food our own models and kind of like get feedback there. But uh you can also, you know, there are a lot of like internal variants of models that you can use to give you different perspectives um here as well. And and we found that to to work quite well. Okay, so just to

18:53 Just to make sure we get a like a barometer of today's world at OpenAI in terms of AI and code. Uh just so I understand and then I wanna move on to a different topic. Uh A hundred percent of code across OpenEI is written by Codex at this point? Is that

19:07 The way to frame it. I wouldn't make the statement that a hundred percent of code running in production today was is written by AI. Uh and and and just it's kinda hard to to to do attribution there. But the like almost every engineer heavily uses codecs in all of their tasks at this point. And so I you know, if I were to guess make like the vast majority of code at this point, it's it was probably authored by yeah.

19:28 Incredible. Okay, so there's A lot of talk and we've been talking about kind of the I C role, the work of an I C engineer. There's less talk about the changing role of a manager, especially an engineering manager. How has your life as a manager changed with the rise of AI and just what do you

19:45 Where do you think the managers What's the role of a manager in the future? Is Development Changed less than an engineer? Uh there's no, you know, codec for managers, just uh however I use codex quite a bit for for some of the Um uh some of some of the like kinda more manager tasks that I do. I'd say a couple of things are are changing. They're like some trends. So I don't think it's changed that much yet, um, but I see trends and I think if you play it out, you can kinda see where where a lot of this is going.

20:11 becoming increasingly clear. Is codex really empowers like top performers to to get a lot. lot I like to be a lot more productive. And so it it really like And I think this is maybe true for AI m more broadly, like across society, which is like the people who really lean in are like the people who

20:30 have high agency or like will really get get Get good at these tools. Will kind of supercharge themselves. Uh and so I'm kinda noticing this now as well, which is like the top performers kind of end up uh uh uh being a lot more a lot more productive.

20:46 Uh and so you see a a broader spread uh in in team productivity in this way. One so one thing that I've always done as as a management philosophy is to spend uh actually the majority of my time with top performers. Just like make sure they're unblocked, make sure they're happy, make sure You know, they're they feel productive. And they feel hurt. I think this is even more true uh in an AI world where, you know, your top firmers are gonna just like really be shooting ahead, uh, using these tools. I think I think one example is the is the the team that's, you know, maintaining a one hundred percent codex generated code base, like just letting them kind of rip and and and see what's happening there.

21:18 Is something that that's paid dividends. So I think that that's kind of one one trend that I'm seeing where you where you where Um spending even more time with top performers for managers, I think is is likely gonna um Uh continue.

21:30 The other thing is Mm. So this is more uh An observation, but My sense is with a lot of these AI tools available to managers, so less like writing code, but just

21:42 Things like chat GPT with organizational knowledge, like being able to do research and understanding organizational context a lot better. Another good example is uh um We're doing performance reviews right now, and it's actually really easy to use chat GPT with internal knowledge hooked up to GitHub and micronotion docs and Google Docs. To give a get a really good sense of what this person has done over the last twelve twelve uh months uh in writing a little, you know, deep research report for it. My sense is I think managers will be able to manage much larger teams in this world. Kinda like how, you know, like software engineers are

22:13 managing twenty to thirty codexes. Um, my sense of these tools will allow managers, uh people manager to be higher leverage. Um and uh It will allow them to to to manage, you know, teams of Way more than than the current best practice of I think is like six to eight, right, for software engineering. You kinda see this apply to, you know, like uh

22:32 the non uh engineering domains like support or uh operations where it's like Yeah, previously. Um Uh w previously like the the size of a support team might be limited, but like as you can pass off more things to agents. You can actually

22:48 Do more work and also manage more people this way. I think the same thing might happen for um people management as well, especially in tech companies. Um And uh we're already seeing this. There's some teams uh where uh their em is managing, you know, quite a few people and they're doing it pretty adeptly because of some of these tools where they can get higher leverage and understand what their team's doing, understand organizational context a little bit better and operate in that way. I love this advice that the way you described it is you've always leaned into top performers and spent more time with them, unblock them, make sure they're happy. The way Mark Anderson he was just on the podcast, the way he phrased it is AI makes good people Better.

23:22 And it makes great people exceptional. Yeah. Yeah. And what you're saying here is just Just doing this more and more is probably the right move. Spending more time with the best people on your team to unblock them, make sure they have everything they need. Yeah, a very good example right now is uh there are I would say like a a group of engineers internally who are really codex filled and are thinking through what the best practices are for interacting with this model.

23:45 And that is just an extremely high leverage thing for them to do. And so just like as a manager, I'm just like, Yeah, go explore this. you know, uh whatever best practices come out of this, you know, we we have to share with the org. Well we'll you know, uh we'll we'll uh we do all these knowledge sharing sessions, we'll we'll like share documents and like best practices everywhere. So things like that just uh you know elevate everyone.

24:06 And uh and so I I view that as like, you know, another example of this trend um uh that um that we're seeing where the top performers really get exceptional. People just like have a sense. This is big. AI is changing so much. The world is changing. Uh, it's gonna be a huge deal. What do you think people aren't pricing in yet into what will change into where things are heading. Just like what's an example of something you think are like, Okay, we're not realizing this yet. So one of my favorite kind of uh uh

24:33 like phrases or like things that have come out of uh this whole AI wave is is the idea of the one person billion dollar startup. I think Sam may have It's fascinating to think about, right? It's like yeah, if if you know, if people are so high leverage, at some point there will likely be Um a one person billion dollar startup.

24:53 Um And while I think that's really, really cool, I think people aren't really pricing in the second or third order effects of this. And and really what You know 'Cause what the one person billion dollar st startup implies is that There's you know, one person can just have so much more agency and so much more leverage using one of these tools.

25:10 Um, that it is just super easy for them to get everything done that they need to for for their business to, you know, ultimately create something that's a billion dollars. But I think there are a couple of other implications of this. So one of them is uh Uh If it's easy for a person to create a one person build or if if it's possible for a person to create a one Person billion dollar startup.

25:29 It also means it's way easier for people to just create startups in general. Like I actually think this will Like one second order effect of this is I think there's just gonna be a huge like startup boom. And like small, like SMB style boom. Um, where anyone can build software for anything, right? Like uh Uh, one, uh you're kinda starting to see starting to see this play out in the AI startup scene where

25:51 Software's became a lot more vertical oriented. Where like these verticals uh like creating some AI tool for some vertical tends to work quite well because you know, you really lean into uh that particular domain, you like really understand the use case for it. And so if you play out

26:07 AI There's no reason why you can't have like Hundred X more of these these startups. Uh and so I think I think one world that we might end up seeing Happen is

26:17 In order to enable a one person billion dollar startup, there might be like a hundred other small startups building bespoke software that works extremely well. to support uh other types of you know small small one person You know, billion dollar startups. And so I think we might actually end it uh enter into a golden age of like B2B SAS.

26:35 Uh and just like soccer and Sarsman General. And so I think I think that's that's a really interesting trend to to kinda see because As it's as it's really as it gets easier and easier to build software Um, as it's easier and easier to uh, you know, uh run a company.

26:50 Um, you might actually just end up seeing way more of these these these stars. So the way I I I've been thinking about is like, yeah, there might be one A one person billion dollar startup. But there might be like a hundred, you know. Uh hundred million dollar startups. There might be tens of thousands Of ten million dollar startups.

27:06 And as an individual, it's actually pretty great to have a ten million dollar business. Like that's like enough for You're set for life at that point. And so You know, we might really see see an explosion. In that way. And I and I feel like people aren't aren't really you know pressing that in.

27:19 Um There's another kind of like third order effect to this. You know, and again, uh all of these, like as you get to the further and further out predictions, I think uh are there's a lot of uncertainty. I think if we end up moving to this world where you end up with these like kind of micro companies building software that works for one or two people. who own the company and and and and are working there.

27:39 Um, I think the startup ecosystem will change. I think the V C ecosystem will change. Yeah, it might we might end up in uh in a world where there's just like a handful of big players that are offering platforms and supporting all of these startups, but You know, the types of venture s scale return startups that can really hundred or thousand X your your investment might actually end up shrinking.

27:59 If you end up having a bunch of these, you know, smaller ten to fifty million dollar uh companies. Uh, which are not great for venture seller returns, but are great for the individuals, the high agency individuals who are now, you know, really leaning to AI to to to build these businesses for themselves. I love how many uh order like uh order effects we've been through. Uh when I've heard the fourth order effect now. Sure. I'm just joking. I uh I can't I it's too fourth order is too too is too gigabrained for me. I can't I can't think that far ahead. It's like inception where just everything gets slower every time you go deeper into someone. Yeah. Every layer.

28:32 Uh Okay, so the billion dollar startup I've been I think about this a lot'cause I I'm not gonna be a billion dollar startup'cause what I'm doing is not venture scale in any way and not super high leverage, but Just could see how many support

28:44 tickets I get from just like the most ridiculous things. It's hard for me to imagine one person Like I'm bearish on this billion dollar startup. I just want to share this thought. Uh simply because of the support costs, even if AI is helping you. At a billion, just like Unless your A C Vs are

29:03 Yeah, very high and you have very few customers. I just dealing with support and people are like, you know, like they can solve their own problems, but they're like, I'll email support ask about this thing. Just dealing with that is Hard to scale is in my experience. So unless you have in my opinion, unless you have a bunch of contractors, which I don't know, does that count as a single person?

29:21 Company, I feel like it's very difficult to scale. billion dollar startup and not have someone helping you with at least the support work and AI, I think will only take you so far. So I I I think that's true. Uh and actually I think my view on it is is is slightly different, which is I think that your, you know, Lenny's podcast might end up becoming a billion dollar Start up, but

29:41 Um what I think might happen is Uh instead of you kind of being the one person who has to dispatching AI. to

29:50 solve and fix those support tickets. I think what might end up happening is there might be a whole smattering of other startups. that are building software And super and and like super tailored towards what you might need. And so You know, uh there might be like

30:05 Ten or twenty startups that build support software for Podcasts and newsletters. And uh That might be a one person startup. Like it doesn't need to be a big one. And

30:15 Uh it's it's and you know, they might be able to just code up this product very, very easily. They are able to kinda like build their own thing. And because it's so tailored and unique and hopefully, you know, useful for you, it might be something that you purchase Um as the one person billion dollar start. I would buy that. I would buy that. Yeah, there's like a question of like what you in house and what you what you like kind of

30:33 uh outsource and what I think might happen is because the cost of writing software and building products is is is is collapsing so much, you might end up outsourcing a lot of this. And in doing so, reducing the size of your company. Uh and so that's kinda the world that I think might end up happening. Again, there's like high uncertainty in what might play out here, but the end result still might be a one like one person driving this like height. high massive leveraged company. That might actually reach a billion dollars. I can see that. I also think about Peter at ClawedBot slash Multbot slash open claw.

31:01 Of just like how he barrage he is right now by all these ask and emails and pings and DMs and PRs just like I'm curious to s and he's not even making any money out of this thing. Um yeah, I I can't imagine what it's like to be him right now. It's must be like absolutely insane. It it pr it's probably like uh Uh you know, like the the the months after we launched ChatGBT, the craziness that was uh as one as one man. Uh he's coming out of the part, by the way in in a week.

31:25 Oh, that's exciting. Yeah. Uh, maybe the fourth order effect is distribution becomes increasingly Important. because there are so many freaking things trying to get your attention. So people with an audience platform I think become more and more valuable, which is

31:38 Good good stuff. Okay. Uh I wanted to come back actually to your management stuff. So I I really loved your insight about Spending more time with our performers has been really successful to you? Just thinking about you as a manager of a team that is building the platform that

31:53 Powers Basically the entire AI economy, like every AI startup, is building on your API. Uh. Clearly you're doing a great job. What other

32:03 kind of core management lessons have you learned. What do you find is really important and and and key to your success as a manager of engineers and just people. Yeah. Um I I think a lot of the lessons that I've learned here, I I don't know how specific it is to the OpenAI API or or some of our enterprise products in in particular. I think my my management philosophy has obviously changed over time, but I think it it's uh Well it stayed the same more than it's changed uh over time. Uh one one of these principles is is kinda what I talk talked to you about before, which is you know spending a lot of time with with top performers, like actually spending

32:37 And like to be very concrete, like it's like more than fifty percent of your time. with your top performers, with maybe your top like ten percent uh performers. And really, really trying your best to empower them. The way that I think about it is um Is is it has kinda come back to this analogy of software engineer as as as a surgeon.

32:56 Um, which comes from the the mythical man month book. So it is actually it it's funny, so I I pull it from the book, but in the book they actually Describe this world where um I think they were like predicting the future No,'cause'cause I think the book was written like in the seventies or something. Um they said that software engineering might end up moving into a world where that software engineers are like surgeons.

33:16 Or like in a surgery room, there's like one person doing the work. Um, and uh you know, there's the one person like cutting or whatever and like doing all the surgery. And everyone else in the room is there to just support them, right? It's like the nurse and like the assistant, the the resident and the fellow. And then the surgeon's like, I need a scalpel and they give him a scalpel. And then uh uh they're like I need you know, this

33:35 Everyone's there to just like, you know, support the one uh surgeon. And so the the the Mythmal Mammoth actually predicted that that is kind of the direction that software's gonna go. I don't think that's exactly played out where like, you know, it's much more collaborative and like It's not only one person doing the work. But I've always really liked that analogy. And and and and uh that analogy is actually what I strive to uh uh kind of like emulate in my own management philosophy, which is Um software engineering isn't really like surgery.

34:03 Where it's not just one person doing work, but the way in which I like treating the people on my team and the way that I act as a manager is I want to. uh empower them, make them feel like they're a surgeon. Um and in insofar as like as like making sure that I'm supporting them and making sure they have everything that they need to to do their work. And it feels like they have an army of people kinda supporting them. Um and looking around corners and giving them everything that they need. When it's really just near the

34:26 As the manager and so like The the the example I that I give is is looking around corners and unblocking people, especially from an organizational perspective, is extremely, extremely useful. And again, going back to the AI conversations, even more important nowadays, right? Like uh If if people are just like Cranking PR after PR.

34:42 The main theme bottlenecking uh progress and and you know shipping something tends to be organizational or like process oriented. And if you as a manager can kind of look around corners and kind of unblock the team. If you can you know, like if if the surgeon

34:55 Need scalpel. But you know, the manager kind of already has a scalpel ready for them. That that's the best case scenario. That's kind of the the way that I approach uh um um management and and especially uh engineering management. And so That's something that that's really, really um stuck with me over time. And uh even though, you know

35:12 software engineers aren't exactly surgeons. That metaphor has always kind of stayed in my mind as of as of uh uh rest of my career. I love that and I I feel like I wonder if that's something AI can help with is look around corners and predict Here this engineer's gonna be blocked by this decision. We need to figure this out. Yeah, that's actually a really good uh point. I haven't tried this yet, but I wonder what would happen if I ask uh Chad GPT hooked up to company knowledge. You know, like what are the active blockers? Uh look through all the notion docs. What are that maybe Slack messages. You know, it's probably in Slack somewhere.

35:41 What are the active blockers on my team and is there something I can do to To help. Um I have not thought about that, but you're right. Yeah. Yeah. Uh and it's I think even more interestingly, what do you anticipate will be a blocker for this engineer or this team in the in the coming months or Yeah, you ask the you ask the model, you ask the AI to do the second and third order things. Anticipate that, and anticipate what the blockers will be next month too. I think we've got a we've got a good idea right here. Yeah. Yeah. This episode is brought to you by Data Dog, now home to Epo. the leading experimentation and feature flagging platform.

36:15 Product managers at the world's best companies use Datadog, the same platform their engineers rely on every day, to connect product insights to product issues like bugs UX friction, and business impact. It starts with product analytics, where PMs can watch replays, review funnels, dive into retention, and explore their growth metrics. Where other tools stop, DataDog goes even further. It helps you actually diagnose the impact of funnel drop offs and bugs and UX friction. Once you know where to focus, experiments prove what works. I saw this firsthand when I was at Airbnb, where our experimentation platform was critical for analyzing what worked and where things went wrong.

36:52 And the same team that built experimentation at Airbnb built Epo. Beta Dog then lets you go beyond the numbers with session replay. Wat exactly how users interact with heat maps and scroll maps to truly understand their behavior. And all of this is powered by feature flags that are tied to real-time data, so that you can roll out safely, target precisely, and lear continuously. Datadog is more than engineering metrics. It's where great product teams learn faster.

37:20 Fick smarter and ship with confidence. Request a demo at datadog HQ dot com slash Lenny. That's datadoghq.com slash Lenny. Okay. I'm gonna shift to talking about the API and the platform that you all build. Some so you work with a lot of companies implementing Your API, your platform building on

37:39 on your on your tools. You told me that you find that A lot of companies actually have negative ROI on their AI deployments. Which uh I think is what a lot of people read about and Feeling.

37:51 And it's interesting actually seeing that. What What's going on there? What are they doing wrong? What do you what what's happening in the world of AI and deployments in ROI? Yeah, so so to be clear, I I I don't like explicitly see quantitative numbers around this. Uh, you know, uh it it's actually really hard to measure. These things

38:08 But especially from observing some companies kind of trying to do AI, I I would not be surprised if uh a lot of AI deployments are actually, you know, negative ROI. I mean, part of this too is that I think there's also general sentiment um from uh folks uh around the country, um, like basically outside of tech. That AI is being forced onto them. Um, and I think part of this is is is uh uh uh probably a symptom of some negative ROI. Uh AI deployments.

38:35 A couple of things I've observed around this. So one one thing is and I think I I come back to this again and again. Like I think we in Silicon Valley just forget that we live in a bubble. Like we are so like Twitter's a bubble, sorry, X is a bubble. Um Silicon Valley's a bubble, software engineering's a bubble. Most people uh in the world, most people in the US are not software engineers.

38:55 Or not very AI pilled. Um are not following every single model release. And so uh uh and so we're just like highly out of the loop on how to use this technology. And so You know, like we um we always talk about all these like best practices for codex, all these like codex pill people within open AI. I'm sure everyone on X two posts are like crazy power users of of these AI tools.

39:16 You know, they they lean into skills, they lean into agents.md M Cs Uh yes, yeah, all all all of that. And uh When I talk to some of these companies and I and I talk to the the actual employees using these It's like the most basic thing that they're trying to do.

39:33 And they like have very little understanding of exactly how this technology works. And so that That's that's kind of like one big observation for me, which is like They're asking very simple questions uh of these things. They're really not not pushing it just yet. And so

39:49 That kinda goes back to uh that kinda ties into to to what I what I think Um more companies do or like what could do or or what what a more ideal AI deployment setup looks like. Um and this is kind of how we've run things with an open AI too. Um, the companies where I think it's it started to work really well have a combination of both top down buy in. So it's like the C suite's like, you know, we're we want to become an AI

40:10 AI first company. And so there's buy in, they buy the tools. They have, you know, exec support. But it also has bottoms up. adoption and buy in.

40:18 And so what I mean by that is it has like actual employees doing the work. Who are really excited about this technology and are willing to learn, evangelize, build best practices. And kind of like knowledge share within the organization.

40:32 We've we've seen this a lot internally. So like Obviously OpenAI has always wanted to be uh a very AI centric company. But where when it really started taking off was when was with the introduction of codecs and these tools where like people them like actual employees themselves could start applying it to their work. Uh and I think you really need this because

40:50 At the end of the day, everyone's work is like very different. It's like very unique. Uh software engineering is different than finance, is different than operations, different than go to market and sales. And so there's like a lot of these like last mile intricacies of Work that needs to really be done in a bottoms up. Fashion.

41:07 And so My sense is a lot of these these A deployments don't have Like don't have bottoms up adoption. Like it was like an exact mandate. And it's extremely top down.

41:17 And it's very divorced from what the actual work looks like. And as an end result, you end up with a giant workforce that doesn't really understand the technology is like, I know I'm supposed to use this and maybe it's like on my performance review too, but Um I'm not sure what to do. Uh and they look around, no one else is doing it, there's no one else to learn from. Uh and so my my uh you know my recommendation for a company's kinda pushing this is is find or maybe even staff a full time team internally that is this

41:40 kinda tiger team internally that can Um explore the full extent of the capabilities, apply to specific workflows, do the knowledge sharing. uh create excitement uh within folks uh who might want to use this technology.

41:53 Uh,'cause in the absence of that it's very difficult to it's actually very difficult to pick up. And who who would you put on this Tiger team? Is it like engineer led, do you find in your experience? Is it a cross functional sort of team? Yeah, it's it's interesting because so um also a lot of companies don't have software engineers. Uh and so uh the the pattern I've seen is it tends to be these like software engineering adjacent like basically technical people, but are not software engineers. I think they those are the ones who get tend to get most excited.

42:21 Uh around this. It's like, you know, maybe the It's like maybe the like You know, support team operations lead. Who doesn't code.

42:30 but loves using these tools and, you know, is like an Excel wizard or something. And so it's like technical adjacent or like coding adjacent and like, you know, pretty technical. Those are the kinds of like those are the kinds of people I've seen in these companies who just like really light up and get excited around this. I mean you can usually build a team uh a team around that. But yeah, it it's like oftentimes not software engineers. Software engineers, I think, will understand this, but not every company has. Software engineers, um

42:53 It's actually kind of a rarity. They're they're they're hard to find, they're expensive. Uh and so it's it's these other other types of folks. What I'm hearing is the anti-pattern is Top down. This is very the CO found exec team just like we are gonna go AI first, we're gonna lead into AI, everyone's gonna be judged on their performance using AI tools. How much your Productivity is increasing thanks to AI.

43:14 And Without With that being just top down and not creating a team that is Bottom up. Spreading the the gospel.

43:22 You find that doesn't work. Yeah. Yeah, exactly. Exactly. And the advice is Find the people that are Most excited.

43:28 And Instead of kind of having them spread out through the organization, your what you find works is Create a little T A I kind of evangelist team that yeah finds ways to use it and kind of spreads it across the work. Yeah, I mean another uh it's kinda like hearing you you play back to me, another way to think about it, kinda tying back to my own

43:45 Imagine a philosophy is just find the high performers in AI adoption and empower them. You know, let them build hackathons, let them you know, hold seminars, do knowledge sharing. kinda create the seeds of uh of excitement internally. Okay, amazing.

43:59 There's a couple of hot takes I wanna hear uh from you something that I've seen you talk about and share. One is um You've shared that. Talking to customers and listening to customers is not always the right strategy in AI and it might often lead you astray. I don't know if it's that hot of a take. I think the main thing here is so obviously you should talk to your customers. Like it's it's like useful to talk to customers.

44:20 I just think the AI field Uh especially what I've seen over the last kind of like three years. Um uh working on the API and and and seeing kind of all that evolve. Is the field and the models themselves are just changing so so quickly. They tend to like disrupt themselves.

44:38 Especially around the like tooling and the scaffolding space. So Uh there there's this quote that I read actually uh earlier this week from a it's from an ex article uh by this guy named Nicholas, who's the founder of a startup called FinTool. Uh, where uh I think he was he was sharing a lot of the best practices that he has learned. Through building AI agents for financial services, I think at a at a start thin tool. Um this phrase that I thought was really good, which is uh the models will eat your scaffolding for breakfast.

45:04 Like if if you look if you rewind back to twenty twenty two. Right when ChatGPT launched. Um these models are pretty raw. And there was like all this product scaffolding and and things, especially in the developer space. To basically

45:17 try and steer the model and build a scaffolding around it. To get it to do what you want. Like agent frameworks, there's like like vector stores, I think was like really popular back then. Uh and just like a whole smattering of tools here. And as you've kind of seen the feel play out, that the models have just changed so much.

45:34 uh that uh and had gotten so much better That they ended up yeah, literally eating some of some of the scaffolding. Um And I think this is even true today. So I think the the the article from Nicholas um actually you know the the current scaffolding, which is uh fashionable, is skills files based context management. I could see a world where at some point, you know That's no longer useful. Uh, where the model can actually

45:56 you know, manage all that themselves or like You know, uh uh or or or there might be you know it's hard to predict, but like might move on to some new paradigm where you no longer need this file based like skills skills type thing. You have literally seen this play out, right? Like the agent frame or something are a little less useful now. Um, there's a period of time like twenty twenty three where we thought vector stores and is is is gonna be like the main way For you to

46:16 you know, bringing organizational context into the models. And you need a you know Uh vectorized and embed every bit of your corpuses and then you do all this work to like figure out the vector search to like optimize that to full out the right information at the right time. All of that is scaffolding.

46:31 Because the model, you know, was not good enough. And turns out, you know, in this case, it turns out as the models get better. A a better approach is actually to take out a lot of that logic and trust the model. I give it a set of tools for search. It doesn't need to be a vector store. You could actually just hook it up to any type of search. It could literally be files on a file system like skills.

46:50 uh an agents M D Uh to kind of steer it uh as well. Obviously there's still a place for vector stores. I I know a lot of companies still be using it. But the the the entire scaffolding around that and building an entire ecosystem around that and assuming that's the only scaffolding that you need has has really changed. And so

47:05 Tying this back to the like, you know, uh it it's you know, you don't always have to listen to your customers. Because the field is changing so much, at any point in time, you know, a lot of people are kind of in this local Local maximum. And If you just blindly listen to your customers, they'll they'll be like, Yeah, I want a better vector store.

47:20 Like I want a better Uh I want a better, you know, agent framework for this. And uh if you had just kind of only chased down that path, it actually would have led you to, you know, build something that again is the local maxima. Whereas as the models get better. we've had to reinvent and kind of rethink the right right uh uh abstractions and the right tools and frameworks to to to to build. Around these models.

47:42 Um and the cool slash exciting slash kind of crazy annoying part is It's a moving target. And so yeah, like the current current smattering of of tools and frameworks right now. We'll likely need to evolve and change pretty significantly over time. Um as the models get smarter and better.

47:57 But that is just the nature of building the space. I think it's what makes it exciting. Uh but it also means when you talk to customers you kind of need a Balance the exact feedback that they want. uh with uh where you think the models are going and where you think things will uh trend over the next one or two yeah. It's interesting how this is um The bitter lesson is uh you know, this big lesson that

48:16 AI and ML folks learned, which is just like Uh don't the less you overcomplicate the less logic you add to to machine learning TI the more it'll be able to scale and grow and just like take it all away and let it just just compute, basically. Just Give it more power to to get

48:31 Yeah, there's yeah, there's literally a version of the bitter lesson applied to like building with AI where, you know, we were trying to architect all this stuff around and turns out the models will just kind of, you know, eat it all away. And and and and honestly like open AI API team has like been guilty of this, uh, where we kind of like took some, you know, left and right turns. Uh when we shouldn't have Um but uh yeah, the models still end up

48:54 Models get better. And uh we're all learning the bitter lesson. Down. So what would be the the key takeaway for folks building on, say, the API or just building agents and You know, having to build a little bit of this around for now is it just yeah, what would be the advice?

49:07 My general guys and I've been giving this to people for a while and I think it's still true today is make sure you're building for where the models are going and not where they are today. Um Uh you know, the the It's it's clearly a moving target. And I think a lot of the companies that I've seen startups that I've seen really, really do well.

49:24 Is they build a product For an ideal like Type of capability. That is like maybe eighty percent of the way there. Today.

49:33 And it like they end up you know, having a product that like kind of works, but is like just almost there. But then as the models get better you know, suddenly it might click and then their product now is incredible because it works you know, like Uh uh like maybe with like oh oh three at some point it's only worse. But five point one, five point two, suddenly it unlocks it.

49:51 But they're building these products with the in like the model capability improvements in mind. And with that you end up creating an experim experience that's way better than if you had assumed that it's it's static in the first place. Um and so that would be my my general uh advice, which is you know, build for where Where where the models are going and not not where they are today. You end up building a better product.

50:10 You may need to, you know, like wait a little bit, but like, you know, the models are getting so much better so quickly you you often don't need to wait um that long. So to follow that thread. Where are like in the next six to twelve months Where is the API heading? Where's the platform heading? Where are the models heading? As much as you can share, I know there's a lot of secret here. that maybe you're most excited about, or do you think that people should start to Prepare for and however much you can share.

50:34 I mean, so the obvious one is um how long of a task uh these models can do coherently. Um so there's like the the meter benchmark that that I think track software engineering tasks and how long You know, like how long of a task can these models do, uh, fifty percent of the time, eighty percent of the time. Uh I think we're at something like

50:53 multi hour tasks being able to be done by uh sovereign earn casks being able to be done by Um uh these frontier models. uh fifty percent of the time and then I think eighty percent is something like just under an hour. But the the the sobering thing about that that chart is they plot all the uh previous models uh on this chart as well. You can really see the trend of this.

51:14 That's something that I'm really excited about, which is, you know I actually think products today. Really optimized for tasks that the model can do for like minutes at a time. Like even codecs and like the coding tools, I'd say like You know, it's it's in the cly, you're kinda like seeing it be interactive. It's really, you know

51:29 quite optimized well for like maybe at most ten minutes. Type has. I have seen people push codex to the limit and do like multi hour long. Uh tasks. Uh, but again, I I I I I think that that's more of the exception.

51:42 But I uh if you follow this trend, like I think like in the next twelve, eighteen months we could see models that could do multi hour long tests very, very coherently. At some point it might reach like, you know, six hours a day long task. Where you kind of like dispatch it and have it do, you know, do things on uh on its own for a while. The types of products you build around that will look very different. You wanna give the model feedback. You obviously don't want it to completely run wild for a day. Maybe you do. But but you probably don't Um and and then the the universe of things you can have the model do really expand. So that's something that I'm really

52:13 I'm really excited about seeing. Another uh thing over the next twelve to eighteen months where I think would be really cool is uh crewments in our in the multimodal models. So uh And and actually by by multimodality, um, I'm mostly thinking about audio. Here we're Uh the models are pretty good at audio. I think they're gonna get a lot better.

52:31 Um at audio over the next six to twelve months. Especially the likes. You know, the um native multimodal Modeled the speech to speech ones. I think there's also interesting work uh being done around

52:42 um new types of models and architectures on the uh multimodal audio side. Uh as well. But uh Audio, especially in the enterprise and in a business setting, I think is a hugely underrated uh domain still. Like everyone talks about coding, it's all text. Uh but uh we're talking in audio.

52:59 A lot of the World's business is done. The audio. Uh a lot of services and operations are done via Uh talking and audio. And so

53:07 Uh, I think that that area is gonna look very exciting in the next twelve, eighteen months. And I think there will be uh even more unlock for uh what we can do uh with with audio models uh there as well. Amazing. So Quick summary, uh expect agents and uh

53:22 AI tools to run longer to that that trajectory to continue to increase. And then audio and speech becoming a bigger deal, more First party and and native and better and And core to the experience. Yeah.

53:35 Extremely cool. Okay, I wanna go back to one of your hot takes. Another hot take that I've seen you discuss you're big uh you're very bullish on business process automation. as an opportunity in the world of AI. Talk about that. Yeah, this go this goes back to the thing that I said previously, which is um we we we live in a bubble in Silicon Valley. And um a lot of the work that we do

53:56 that we're used to software engineering, you know, product management, building products. Uh is very differently shaped than the work that goes on um that runs our entire economy. And I see the say in and now when I talk to customers. Uh if you if you talk to any like you know, company that's not based in uh it's not a tech company. Um there's a lot of business processes.

54:17 And so what what I mean by this is is you know, I I generally delineate it as You know, there's like uh like software engineering is kind of like open ended knowledge work. Right, it's and this is why I think uh tools like Codex tend to be quite quite good because it's exploring and and you're giving it these like open ended things. But software engineering fun is fundamentally like pretty open ended. And it's not very repeatable.

54:38 Right. So like uh you build a feature, you're not trying to build the exact same feature over and over again. And a lot of like tech Jobs are in the space. I think like data science is kind of in the space as well. Even some of the like strategic finance stuff. But as you move further and further away from software engineering and like what what is current tech. A lot of jobs are just business processes. They're like repeatable things.

55:00 Uh repeatable operations. um that you know some manager at a company has kind of like iterated on Um, there's usually a standard operating procedure. That people want to do Uh and you don't want to deviate from it.

55:13 that much, you know, there's like in software engineering, the ingenuity isn't isn't isn't deviating. But a lot of a lot of the the the work being done in the world i is actually just um running through these procedures and and operations. Like if I you know if I call Um a support line. They're running through one of these. If I call my utility company

55:31 There's a bunch of processes and things that they can and cannot do. Um for me. Uh, and so I'm I'm just extremely bullish on this general category of like And and and I think it's underrated because it's so different from what we think about in in Silicon Valley, people tend to not think about it. But how can we apply um AI

55:47 uh and and some of the tools and frameworks that we have towards this business process automation. Towards automated Automating and making easier Um repeatable business processes. With high determinism.

56:00 Um that is fully integrated with business uh data and business decisions. And and and different systems within an enterprise. Um and how can we actually make that that process better? Uh, because I actually think there's a lot of opportunity and a lot of work to be done uh in that area. And we just we just don't talk about it'cause it's it's uh a little bit less uh uh in our wheelhouse.

56:20 So your take here just to make sure I fully understand it is you think there's a much uh bigger opportunity outside of engineering. Impact. uh productivity of companies and also jobs of these folks that are doing these kind of

56:33 Repetitive, easily automated tasks. Impact jobs and also just impact how work is done. Like so much of work is done in this way. Like you think about You know, like what a Like basically we I I talk to customers all the time, big enterprises like like how how will AI transfer my company? Like how will it run in in in in a world uh with AI in like twenty years? Um and and you know, software engineering is part of the story. But there's so much more on the business process side.

56:58 And I think it might look even more different on the business process side and And the work there is is pretty substantial. It's actually interesting. I don't know like from an absolute percentage or absolute basis. I don't know if it's bigger or smaller than software engineering. Like software is pretty huge and pretty extensive. Uh as well. But it is pretty massive and it's definitely bigger than, you know, uh uh Alright.

57:17 It's it's bigger than you would think it is based off of how how people talk about it or don't talk about it on X or Twitter. Okay, uh In going in a slightly different direction. Uh Uh having built the platform building an API.

57:29 Uh People building on API, the biggest question on people's minds is always just Uh. How do I not have open AI squash my idea and build their own thing and then, you know? destroy this this market I created.

57:42 What's the general policy, what's the general philosophy of how startups should think about where open AI is unlikely to go. My my general answer here is is um The market is so big. And so massive.

57:56 Like I actually think You know, startups should just not overly think about where Open AI where these labs are going. I've talked to a lot of startups, you know, that have you know not worked out, starts that are doing really well. Every startup that I've seen that has kind of fizzled out is not because OpenAI or you know a big lab or Google or something has has come to watch them. It's because they built something and it like really didn't resonate with with the customers. Whereas the ones that take off,

58:20 Like even in very competitive spaces like coding. Like cursor is huge at this point. And it's because they build something that people really love. And so My general advice is like don't, you know Don't overly stress about this. Just build something that people like and you will you will have a space in this. I can't overstate how big of an opportunity there is right now. Like

58:39 The the the opportunity space of building with AI is so big. Like the space is so big that the overturn window of what is acceptable and not acceptable for VCs to do has completely changed here. VCs are like investing in like competitive companies left and right. It's just like the space is so big because because the opportunity is is is is unlike anything that we've seen before. And while you know,

59:01 Uh that that affects how VCs operate. From a starter perspective, it's like the most empowering thing in the world because the Like even if you just build something that that some people really, really love, you will you will end up with a massive, massively valuable business. Uh and so Okay.

59:14 That's why I told you like don't don't over anything about it. The other thing like I also think is important to remember, uh, at least from an open AI perspective, one thing that that that we've always held very near dear, which both Sam and Greg helped You know, reinforced from the top as well. Is we I actually view ourselves fundamentally as a like ecosystem platform company. The API was our first product.

59:34 We think it's really important for us to foster this ecosystem and continue to, you know uh support it and and not squash it. And so if you kinda look at the decisions we make, it this is all we've Weave through it. Every single model we've released in one of our products gets released in the API. Like even you know, we rele release these codex models now that are a little bit more optimized for the codex harness.

59:53 But they always find their way into the API and like all of our, you know. Never using those, we don't hold back on any of that. Uh, we think it's really important to keep our platform neutral. And so you know we don't block competitors.

1:00:05 Um we allow people to have access to our models. Um Uh we also want, you know, like uh we've recently been testing more of like the sign in with chat GPT, you know, uh product as well. And so we we we want to foster this ecosystem. I think it's really important that we do so. Uh the general like thinking about this is like, you know, a rising tide r like lifts all boats and You know, we might be a aircraft carrier like pretty big at this point. But we think it's important to raise the tide, uh, because everyone kind of uh benefits and I think we'll benefit as well. Like our API itself has grown.

1:00:34 And so I'd really encourage people not to view open AI as this kind of like You know thing that'll just uh uh shove people out of the way. But instead focus on on building something valuable. Uh and we, you know, remain committed to to to providing an open ecosystem.

1:00:50 Why why is that important to open AI, just this focus on Building a platform, creating a way for people to build businesses just like Is that just that's been the vision from the beginning? We want this to be a a platform. It's been the vision from the beginning. It comes goes back to our charter, actually, like our our mission.

1:01:09 Um so the open air's mission has always been to one to build AGI, so you know, where I was getting that. But then the second thing is to like spread the benefits of it to all of humanity. And there's kind of like a lot of, you know, uh uh the main part there is all of humanity. Like uh and obviously Chad GPD is trying to do this, you know, we're trying to reach however many, you know, the the whole world But very early on, and this is why we we launched the API you know back in I think it's like twenty twenty or something like really early.

1:01:33 We don't think we as a company will be able to reach all of humanity, right? Like there's I don't know, every s every corner of the world's like like pretty pretty pretty deep. And so We actually feel like in order for us to fulfill our mission, we need to have some platform. style thing here where we can empower other people to build

1:01:50 You know. The customer support bot for podcasters and newsletter hosts. Uh because we're not gonna be able to do it ourselves. Uh and so we've largely seen this play out with the API. Uh, this is why we we you know we we we we talk to so many of our customers and and and really you know love seeing the diversity of of things built on. But yeah, it it's been there since day one because it's it's it's kind of we view it as an expression of our mission.

1:02:12 You haven't even mentioned the uh the app store the guys are launching, the chat GPT app store. Yeah. Is Is that under your umbrella, by the way, or is that a different Oregon team? It's a it's a different team. So it's under Chat GPT. We obviously collaborate very closely with them and uh you know, they built like an apps SDK, uh, which is uh built in close collaboration with our team. Uh but that is more within the Tat GPT. Umbrella.

1:02:31 Uh, but that is also another like that's another example of this, right? It's like Chat GBT is like We we we we we kind of like have these eight hundred million weekly active users who are just coming over and over again like It's a great asset to have as a business, but like, man, would it be better if we could somehow

1:02:49 allow, you know, uh other companies to come in and and and and uh take advantage of this as well and and build for this this audience as well. And and then ultimately we think it'll help us expand that that that group as well, right? And so It's all it all kinda comes back to the mission and Uh, we find that being a platform being open tends to help here. Just that number, eight hundred million I think it's M M A's just like weekly. Weekly acting. We uh billion people using weekly

1:03:16 Like it's absurd having no how these numbers we're just used to now, but that's i in insane. Unprecedented. Yeah, it's it's mind bottling for me to think about from a scale perspective, uh, honestly. Yeah, and the way I think about it is like ten percent of the world. Uh and growing, by the way. Like it's just it's it's shooting up. Um uh come to Chat GPT uh um and and use it every day. Or sorry, every week.

1:03:37 At this point, I just wanna double down on this point you're making Open AI's mission was to make AI available to all humanity. And I think some people just that. They're like, Oh, you know, it costs money and it's like uh Like the fact that it It's there's a free version of chat GPT that anybody can use.

1:03:54 that is not so different from the most powerful AI model that exists in the world. For free. that's not gated, that anyone could use. Like if you have uh if you're a billionaire, there's only so much more you can get out of AI than what someone, you know, in a village in Africa can can get. And I know that's always been really important to open AI.

1:04:11 Yeah, yeah. I mean Look, uh that that's why I think we've lean into the health work. We've lean into like like uh like uh education's gonna be a very interesting thing here. Um the other In insane kind of trend here is is the free model has gotten so smart over time. Like the free model back in twenty twenty two was You know, like uh

1:04:29 Well it's good at the time, but it's like nothing compared to what you get today'cause you get two P D five today. Uh and so the like, you know, raising the floor across the world is kind of, you know. something that we're really we're trying to do and and we view it as as part of our Mission. The other flip side of this, by the way, is like, you know, kind of talking about like the billionaires or or whatever.

1:04:46 I know people love saying like you're using the same iPhone that like you know Steve J or sorry, uh like Mark Zuckerberg's probably using or like the billionaires are using. For like twenty dollars a month. You're basically using you know, like using the same AI. That You know, the billionaires are using

1:05:00 Uh for like Two hundred dollars a month. uh you get the same pro model that you know all the buildingers are using but they're probably not using pro for everything. They're probably just using the the plus tier ones. uh for their day in and day out. And so yeah, this kind of like democratization and just like Spreading of this this benefit like across all of the world is

1:05:17 And that's really meaningful to us and something that Um Uh drives a lot of of of what we do. One last question, just for folks that are thinking about building on the API are just like, Oh wait, I could do cool stuff with Open AI's models and APIs.

1:05:29 What What does your API and ma and platform allow people to do? Like I know you can build agents on top of the platform, just talk about what you allow. So fundamentally the API offers a bunch of developer endpoints. Uh and and uh and these developer endpoints basically let you sample from our models. The most popular one that we have right now is one called responses API.

1:05:49 Uh, and so this is a an endpoint and it's optimized for building long running agents, so agents that'll work for a while. So where you can basically use can you can you can Yeah, I had a very you know. uh uh low level You're basically just giving the model text.

1:06:03 The model will work for a while. You can kind of You know, pull it to see see what it'll do and then you'll get the model response back at At some point. lowest level primitive that we have uh for people. And that's actually what a lot of people use. That's the most popular way of building on top of API.

1:06:18 With that, it is like Super unopinionated. And you can do Basically whatever you want. It's like the lowest level thing. We've also started building more and more kind of like layers of abstraction on top to help people.

1:06:29 Build uh some of these. Uh and so next layer up we have this thing called the agents SDK. Which has also gotten extremely, extremely popular. Um, this allows you to use, you know, the responses API or some other API endpoints that we have. To build uh what you might more traditionally think of as an agent, like uh, you know, an AI kind of working in an infinite loop. It might have subagents that it delegates to. It starts building all this framework, all this scaffolding, actually. You know, we'll see where this all goes.

1:06:53 Um But it makes it a lot easier for you to build these these these these kind of agents. Giving it guardrails, allowing it to like farm out sub tasks to other agents and and kind of like orchestrate a swarm of agents. Uh they agents S C K

1:07:06 Uh kinda lousy to do that. And then above that Uh, we've now started building tools to help Uh also with kind of like the meta level of deploying an agent. Uh so we have this product called uh

1:07:18 Um agent kit. Uh uh. Uh and widgets, uh, which are basically a bunch of UI components that you can use to very easily Um build a very beautiful UI. Um on top of uh uh either our API or agents SDK.

1:07:31 Um, because you know, a lot of times these agents kind of look very similar from a UI perspective. Uh and so there's Asian kit. We also have a smattering of like uh eval's products, like eval's API where you wanna test and like, you know, see if your models were or your your agent or your workflows working. Uh you can test it in a very quantitative way. I'm using our Edol's product.

1:07:50 And so yeah, the I view it as like these these various layers, they're all kind of helping you build Um what you want um with our uh AI. with our models, um and with increasing levels of abstraction and and and and and uh you know how opinionated it is. And so Um you can stuff you can do the you can use the whole stack and and it it uh very quickly allows you to build an agent. Um or you can go down down the stack as low as you want to to basically the responses API and and build

1:08:15 Um whatever you want, uh because of how low upload is. Sherwin, is there anything else that You wanna share anything else you wanna leave listeners with? Anything we haven't touched on that you think might be helpful before we get to our very exciting lightning round. The only thing I'd I'd leave folks with is yeah, I I think

1:08:29 Um I think the next like two to three years are gonna be Some of the most fun. uh in tech and in the startup world, uh that that will have in a very long time. And uh I would just encourage people to not uh not take it for granted. Like I I entered the workforce in twenty fourteen. It was great for like a couple of years. I felt like there was like a period of like

1:08:49 five to six years where it wasn't very exciting in tech. Uh and then in the last three years has just been the most insanely exciting, energizing period uh of my career. And I think the next two to three is gonna be a continuation of that. And so Uh when Kurtville not take it for granted, I'm trying to not take it for granted.

1:09:05 At some point, you know, this wave's gonna play out and it's gonna be a lot more, you know, incremental. Uh, but in the meantime we're gonna get a explore a lot of really cool things, invent a lot of new things and change the world and change how we work. And so uh that's the main thing I'd I'd leave folks with. I love this message. I want to spend a little more time on it. Um when you say don't miss it, is it What do you recommend people do? Is it just build, lean in, learn?

1:09:26 join a company building really interesting things like what's What's your advice to folks that are like, Okay, I don't wanna miss the boat? Yeah, I would just say engage with it. So it's basically like what you said, um lean in. Um building uh tools on top of this is is part of the you know, is part of the story. I'm just using the tools. Like you don't you know, you don't need to be a software engineer to to lean into this. Um

1:09:46 All I think a lot of jobs are gonna gonna gonna change here. So just using the tools, understanding the limitations of what it can and cannot do. So that you can kind of watch the trend of what It can start to do um as the models improve. And yeah, and so it's basically like getting used and getting getting used to the technology and getting familiar with it, instead of kind of like laying back and uh uh uh letting it letting it pass you. On the flip side of that, there's a lot of

1:10:09 I think stress and just anxiety around like there's so much happening. How do I keep up? I gotta learn a Clotbod this week. Oh god. Wha is there something you learned about it just not like you're at the center of this. How do you not get overly stressed and worried about missing things that are going on and just key stay on top of news what what are some things you've done learned. Yeah, so I I think I'm personally a bad example of this because I am I'm basically chronically online.

1:10:33 Uh on X and uh uh our company Slack. So I I I actually try and absorb. I end up absorbing a lot of it. What I will say though is just like from observing other folks who are less, you know, addicted to this stuff like I am. Um Yeah, a lot of it is noise. Like you don't need to you don't need to have like one hundred and ten percent of this kind of pass your mind. Like go into your mind.

1:10:56 Honestly, just leaning into like one or two different tools starting small is already like, you know, more than you need. Here. I think just the combination of like the frenetic pace of the industry X as a product. just creates like this insane kind of like um Uh uh Like yeah, this insane like pace of of news.

1:11:15 Which is honestly very overwhelming. Uh the main thing is like you don't need to be you don't need to know all of that to to really engage with what's happening right now. And even something as simple as just like Install the Codex Clyde and play around with it. Install Chad GPT and connect it to a couple of your Uh, you know, internal uh uh data sources, notion, Slack, GitHub.

1:11:34 And see what it can and cannot do. Um all of that I think is uh a a part of it. Amazing. Sherwin, with that, we reached our very exciting lightning round. I've got five questions for you. You ready? Yeah. Yeah, absolutely. First question, what are two or three books that you find yourself recommending most to other people? Oh I'll talk about one nonfiction one in one fiction book. Uh the fiction book was I just finished reading it. I I it it was really I I really recommend it. It it's uh

1:11:58 Uh there is no anti mimetics division by QNTM. Uh say uh I think he's like an online author, but I saw it being shared on X, uh this this uh it's like a science fiction y kind of book. Um, and it was I basically devoured it in like Two days. Um it was it's super, super well written, super fascinating. It's about

1:12:18 A government agency that's fighting, you know, things that make you forget it. Um and so it's just a very like smart, like creative both that that and fresh uh honestly in terms of like source material. Uh that that I really like. So I'd I'd recommend that one. Uh the book is also unintentionally hilarious. So like it's like meant to be like this like sci fi almost like horror. Solid book, but it is it was it was uh It made me laugh a couple times.

1:12:40 So uh that's the that's the um fiction book. Nonfictions, I'm gonna cheat and I'm I'm gonna recommend two of them. So in the last year I've been reading a lot more about China and kind of like the US China relations. And I think there are two books that came out in the last year that have been, you know, really, really eye opening for me in in in that regard. First one is the Dan Wing book Breakneck. That one was really, really good. I really liked his analogy of like the lawyerly US is the lawyerly society. China is the engineering society.

1:13:05 Uh and there are pros and cons to each. I read it and I was like, Hm, yeah, does Does seem like we're run by lawyers uh in the US. So then that's one. Uh and the other one is the Patrick McGee book on Apple in China was super, super interesting. I'm a huge Apple fanboy.

1:13:20 Like if you could see my uh desk right now, it's it's all Apple stuff. But just like One, it was just super fascinating learning about Apple's relationship to China. And then two It just like had a lot of inside information about Apple as a company that I found fascinating. So it was also quite a page turner and

1:13:36 Um also, you know, a very, very timely a timely book as well. The anti mimetics book sounds amazing. I'm buying it right now as you're talking. Yeah. Yeah. It's it's like I think it's only like a couple hundred pages. I literally finished it in two. It was just like so so good. Okay, great tip. Okay, uh favorite recent movie or TV show you have really enjoyed? Yeah, that one's tough'cause you know, with I've I've two kids and uh uh a busy job and so I really haven't had much time.

1:14:02 Um To watch T V shows, uh I will say in the last couple of weeks I watched a couple of episodes. I'm I'm actually a big anime guy. And so uh I s I I watched a couple of episodes that there's a new season of this anime called Jujutsu Kaisen. Uh that's out.

1:14:15 Uh so season three of J J K uh was was was really good. Um in general, uh, I'm a huge uh fan of uh Japanese anime. I think they create the most uh novel and unique uh plots uh in universes that uh Western media has shied away from. Um and so Uh generally big fan of that. But yeah, it haven't really watched much, but saw a couple of episodes of J J K recently.

1:14:39 Extremely understandable in your role. Yeah. Favorite product you recently discovered that you really love? Yeah, okay. So so um so I recently uh had to set up uh Wi Fi and like home networking. And I went all in on ubiquity, uh routers, um and cat security cameras. I had never heard of it before I had how to do this. I always just had a very simple setup.

1:14:59 Uh and it is just such a well built product. Uh I don't know if you've used it before, but it's basically like the apple of like Home networking. So uh beautiful products. Uh, but the thing that actually makes it extremely good is its software is good. Uh, and so they have a really great um mobile app. To help manage, you know, uh all of the the home networking.

1:15:19 Um and so basically uh Ubiquity you can use it to buy uh wireless routers Um you need Ethernet uh wiring throughout your house to use it. Um, but I actually think what makes it really good are security cameras. So if you have security cameras that are plugged into the Ubiquity ecosystem, they have an incredible mobile app. Uh an Apple TV app and iPad app, um to kind of see the live feed of your cameras and

1:15:40 And so uh they're they're they're a little pricey, but not that pricey. Uh, but it's been a just an incredible product experience. All right. I went Eero, so I made a mistake. Euros are pretty good too, but uh ubiquity fully converted to ubiquity at some point, good tip. Okay, two more questions. Do you have a favorite life motto that you find yourself going back to in work or in life?

1:15:59 Yeah, uh the one that I always, you know, repeat to myself is uh uh never feel sorry for yourself. There's a lot of things that are gonna happen. you know, uh, at work, uh in life, uh, and reminding yourself to never feel sorry and that you always have a sense of agency to kinda pull yourself up. It's something that I've had to tell myself a lot and um also something that I repeat to to to to uh a lot of other folks as well. Last question.

1:16:23 So in your previous life you worked at Open Door, where you led work on basically figuring out how much to Uh pay for houses. You basically built the model that told the company, here's how much we'll pay for this house. What's like a variable in the price of a house that you didn't expect is really important and impacts the price of a house? There's a bunch that were surprising. I'll I'll maybe list the the the the

1:16:43 The the couple of mo most uh uh interesting ones. Um Power lines? And like uh high voltage power lines like are super super Uh they actually impact your price quite a lot.

1:16:55 I didn't really fully internalize this until I went to like Dallas and observed like when your house sits next to one of these giant like you know voltage lines like buzzing. And most people have families, you don't want your kids kinda near there. Uh, so I think that was one that really Really uh kinda surprised me. That makes sense. Yeah. And then the other one which which uh uh was something that uh was always something really difficult for us to

1:17:17 Uh quantify Oh, was floor plans. Uh and so it is very important, like yes, of course it it's really important. But just like quantifying what a good floor plan is like and what a really bad floor plan is like. We were doing all these things of like how wide is the kitchen and like is it a what style of kitchen is it? And then like where's the master bedroom and

1:17:36 And so it was just really, really hard to quantify. But I remember floor plan was a big one because like we'd have a home that like wouldn't sell and then our uh ops team would go in and be like, Yeah, that's the floor plan issue. So like how do you how how could you tell? It's like you go inside, you just feel it. It feels you know, the floor point feels feels off. Uh so yeah, th th those are ones that were uh surprising. And then the last one that was more impactful than I thought is um General like curb appeal and like even like the front door. Uh and so I actually think there there's a Zillow book on on this where um the front door replacement tends to be the highest ROI.

1:18:06 Uh for homes. Um but just like the feel of like As you walk up to the home As a buyer. what you're interacting with and the first moments of the house I think was uh I'd underrated its importance.

1:18:18 That is extremely interesting. Uh and I love that you had to f figure out a deal all this uh in code and not. Yeah, yeah, floor plans. I have a bunch of stories around like for floor plans, there's like there's like uh it's not digitized, so there's like a handful of people who have like paper floor plans. uh of like all these homes in like Phoenix and Dallas. Um yeah, a lot a lot of fun fun stories from the open door days. Okay, Sherwin, uh thank you so much for doing this. This was incredible. Uh, where can folks find you online and uh and how can listeners be useful to you? Yeah, so I'm uh online on on Twitter on X, I'm just at Sherwin Woo.

1:18:49 And uh yeah, I mostly just tweet about uh Open AI and the API and some of the products that we're launching. Uh, and then how folks in the industry uh can be useful to me. Uh, I love hearing about things that people are building. And so if you're working on a startup, if you're hacking on an idea. you know, would love to uh uh just reach out to me on X. Um, I would love to hear about uh what you're building and And learn about how OpenAI can help support you.

1:19:11 Amazing. Sherwin, thank you so much for being here. Yeah. Thank you, Lenny. 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.

1:19:24 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.