The Exchanges

Q the stack today, OpenAI sits in a certain place. How far up the stack is OpenAI going to go? I think it's a brilliant question, but if you're spending a lot of time tuning your rag system, is this a waste of time because OpenAI ultimately thinks they'll own this part of the application layer, or is it not? And how do you answer a founder who has that question?

A The general answer we try to give is we are going to try our hardest. And believe we will succeed at making our models better and better and better. And if you are building a business that patches some current small shortcomings, If we do our job right, then that will not be as important in the future. If, on the other hand, you build a company that benefits from the model getting better and better, if, you know, an oracle told you today that, oh, four was gonna be just absolutely incredible and do all of these things that right now feel impossible, and you were happy about that, then, you know, maybe we're wrong, but at least that's what we're going for. And if instead you say, okay, there's this area where there are many, but you pick one of the many areas where, oh, one preview underperforms, and so I'm gonna patch this and just barely get it to work, then you're sort of assuming that the next turn of the model crank won't be as good as we think it will be. And that is the general philosophical message we try to get out to startups. Like, we, we believe that we are on a pretty, a quite steep trajectory of improvement. And that the current shortcomings of the models today, um, will just be taken care of by future generations. And, you know, I encourage people to be aligned with that.

AI assessment note: “If you are building a business that patches some current small shortcomings”

Q On the model side, Sam, everyone says that, ah, models are depreciating assets. The commoditization of models is so rife How do you respond and think about that? And when you think about the increasing capital intensity to train models, are we actually seeing the reversion of that where it requires so much money that actually very few people can do it?

A It's definitely true that there are depreciating assets. Um, this thing that they're not though worth as much as they cost to train, that seems totally wrong. Um, to say nothing of the fact that there's like a, there's a positive compounding effect as you learn to train these models, you get better at training the next one, but the actual, like, revenue we can make from a model, I think, justifies the investment. To be fair, uh, I don't think that's true for everyone, and there's a lot of, there are probably too many people training very similar models, and if you're a little behind, or if you don't have a Product with the sort of normal rules of business that make that product sticky and valuable, then yeah, maybe you can't. Maybe it's harder to get a return on the investment. We're very fortunate to have ChatGPT and hundreds of millions of people that use our models, and so even if it costs a lot, we get to, like, amortize that cost across a lot of people.

AI assessment note: “revenue we can make from a model, I think, justifies the investment.”

Q incredibly proud moment for me, ah, for certain segments like the one you mentioned there, there would be the potential to steamroll. If you're thinking as a founder today building, where is OpenAI gonna potentially come and steamroll versus where they're not? Also for me as an investor, trying to invest in opportunities that aren't going to get damaged. How should founders and me as an investor think about that?

A There will be many trillions of dollars of Market cap that gets created, new market cap that gets created by using AI to build products and services that were either impossible or quite impractical before, and there is this one set of areas where we're going to try to Make irrelevant, which is, you know, we just want the models to be really, really good, such that you don't have to, like, fight so hard to get them to do what you want to do. But all of this other stuff, which is building these incredible products and services on top of this new technology, We think that just gets better and better. Um, one of the surprises to me early on was, and this is no longer the case, but in like the GPT 3.5 days, it felt like 95% of startups, something like that, wanted to bet against the models getting way better. And so, and they were doing these things where we could already see GPT four coming and we're like, man, it's going to be so good. It's not going to have these problems. If you're building a tool just to get around this one shortcoming of the model, that's gonna become less and less relevant, and we forget how bad the models were a couple of years ago. It hasn't been that long on the calendar, but there were, there were just a lot of things, and so it seemed like these good areas to build a thing, ah, to like, to plug a hole, rather than to build something to go deliver Like, t…

AI assessment note: “If you're building a tool just to get around this one shortcoming of the model”

Q The question is, how do you think about hiring incredibly young, under-thirties as this, like, Trojan horse of youth energy ambition, but less experience, or the much more experienced, I know how to do this, I've done it before?

A I mean, the obvious answer is you can succeed with hiring both classes of people. Like, we have, I was just, like, right before this, I was sending someone a Slack message about There was a guy that we recently hired on one of the teams. I don't know how old he is, but low twenties probably doing just insanely amazing work. And I was like, can we find a lot more people like this? This is just like off the charts brilliant. I don't get how these people can be so good, so young, but it clearly happens. And when you can find those people, they bring amazing fresh perspective energy, whatever else. On the other hand, uh, when you're like designing some of the most complex and massively expensive Computer systems that humanity has ever built, actually like pieces of infrastructure of any sort, then I would not be comfortable taking a bet on someone who is just sort of like starting out, uh, where the stakes are higher. So you want both, uh, and I think what you really want is just like an extremely high talent bar of people at any age and a strategy that said I'm only gonna hire Younger people, or I'm only gonna hire older people, I believe would be misguided. I think it's like somehow just not, it's not quite the framing that resonates with me, but the part of it that does is, and one of the things that I feel most grateful About Y Combinator Four is inexperience does not inherentl…

AI assessment note: “the obvious answer is you can succeed with hiring both classes of people.”

Q the stack today, OpenAI sits in a certain place. How far up the stack is OpenAI going to go? I think it's a brilliant question, but if you're spending a lot of time tuning your rag system, is this a waste of time because OpenAI ultimately thinks they'll own this part of the application layer, or is it not? And how do you answer a founder who has that question?

A The general answer we try to give is we are going to try our hardest. And believe we will succeed at making our models better and better and better. And if you are building a business that patches some current small shortcomings, If we do our job right, then that will not be as important in the future. If, on the other hand, you build a company that benefits from the model getting better and better, if, you know, an oracle told you today that, oh, four was gonna be just absolutely incredible and do all of these things that right now feel impossible, and you were happy about that, then, you know, maybe we're wrong, but at least that's what we're going for. And if instead you say, okay, there's this area where there are many, but you pick one of the many areas where, oh, one preview underperforms, and so I'm gonna patch this and just barely get it to work, then you're sort of assuming that the next turn of the model crank won't be as good as we think it will be. And that is the general philosophical message we try to get out to startups. Like, we, we believe that we are on a pretty, a quite steep trajectory of improvement. And that the current shortcomings of the models today, um, will just be taken care of by future generations. And, you know, I encourage people to be aligned with that.

AI assessment note: “if you are building a business that patches some current small shortcomings”

Q Am I allowed to ask what's your top worry? I'm, I'm in so much, I've got past the stage of being in trouble for this one.

A It's sort of generalized complexity of all we as a whole field are trying to do, and it feels like a, I think it's all gonna work out fine, but it feels like a very complex system. Now, this kind of, like, works fractally at every level, so you can say that's also true, like, inside of OpenAI itself, uh, that's also true inside of any one team, um, but, you know, an example of this, since you were just talking about semiconductors, is you got to balance the power availability with the right networking decisions, with being able to, like, get enough chips in time and whatever risk there's going to be there, um, with the ability to have the research ready to intersect that, so you don't Either, like, be caught totally flat-footed, or have a system that you can't utilize, um, with the right product that is going to use that research to be able to, like, pay the eye-watering cost of that system. So, it's, supply chain makes it sign, sound too much like a pipeline, but, but yeah, the overall ecosystem complexity at every level of, like, the fractal scam is unlike anything I have seen in any industry before. Uh, and some version of that is probably my top worry.

AI assessment note: “It's sort of generalized complexity of all we as a whole field are trying to do”

Q incredibly proud moment for me, ah, for certain segments like the one you mentioned there, there would be the potential to steamroll. If you're thinking as a founder today building, where is OpenAI gonna potentially come and steamroll versus where they're not? Also for me as an investor, trying to invest in opportunities that aren't going to get damaged. How should founders and me as an investor think about that?

A There will be many trillions of dollars of Market cap that gets created, new market cap that gets created by using AI to build products and services that were either impossible or quite impractical before, and there is this one set of areas where we're going to try to Make irrelevant, which is, you know, we just want the models to be really, really good, such that you don't have to, like, fight so hard to get them to do what you want to do. But all of this other stuff, which is building these incredible products and services on top of this new technology, We think that just gets better and better. Um, one of the surprises to me early on was, and this is no longer the case, but in like the GPT 3.5 days, it felt like 95% of startups, something like that, wanted to bet against the models getting way better. And so, and they were doing these things where we could already see GPT four coming and we're like, man, it's going to be so good. It's not going to have these problems. If you're building a tool just to get around this one shortcoming of the model, that's gonna become less and less relevant, and we forget how bad the models were a couple of years ago. It hasn't been that long on the calendar, but there were, there were just a lot of things, and so it seemed like these good areas to build a thing, ah, to like, to plug a hole, rather than to build something to go deliver Like, t…

AI assessment note: “If you're building a tool just to get around this one shortcoming of the model”

Q What do you think people think about agents that actually they get wrong?

A Well, it's more like, I don't, I don't think any of us yet have an intuition for what this is going to be like. You know, we're all gesturing at something that seems important. Maybe I can give the following example. When people talk about An AI agent acting on their behalf. Uh, the, the main examples they seem to give fairly consistently is, oh, you can like ask the agent to go book you a restaurant reservation. Um, and either it can like use open table or it can like call the restaurant. Okay, sure. That's, that's like a mildly annoying thing to have to do. And it maybe like saves you some work. One of the things that I think is interesting as a world where, ah, You can just do things that you wouldn't or couldn't do as a human. So what if, what if instead of calling, uh, one restaurant to make a reservation, my agent would call me like 300 and figure out which one had the best food for me or some special thing available or whatever. And then you would say, well, that's like really annoying if your agent is calling three restaurants. But if, if it's an agent answering each of those 303 hundred places, then no problem. And it can be this like massively parallel thing that a human Can't do. So that's like a trivial example, but there are these like limitations to human bandwidth that maybe these agents won't have. The category I think though is more interesting is not the one t…

AI assessment note: “category I think though is more interesting is not the one that people normally talk about”

Q On the model side, Sam, everyone says that, ah, models are depreciating assets. The commoditization of models is so rife How do you respond and think about that? And when you think about the increasing capital intensity to train models, are we actually seeing the reversion of that where it requires so much money that actually very few people can do it?

A It's definitely true that there are depreciating assets. Um, this thing that they're not though worth as much as they cost to train, that seems totally wrong. Um, to say nothing of the fact that there's like a, there's a positive compounding effect as you learn to train these models, you get better at training the next one, but the actual, like, revenue we can make from a model, I think, justifies the investment. To be fair, uh, I don't think that's true for everyone, and there's a lot of, there are probably too many people training very similar models, and if you're a little behind, or if you don't have a Product with the sort of normal rules of business that make that product sticky and valuable, then yeah, maybe you can't. Maybe it's harder to get a return on the investment. We're very fortunate to have ChatGPT and hundreds of millions of people that use our models, and so even if it costs a lot, we get to, like, amortize that cost across a lot of people.

AI assessment note: “the actual, like, revenue we can make from a model, I think, justifies the investment.”

Q What did you not know that you would have liked more time to learn?

A I mean, I would say, like, what did I know? One of the things that just came to mind out of, like, a rolling list of a hundred is how hard it is, or how much active work it takes to get the company to focus not on how you grow the next 10%, but the next 10 X. And growing the next 10%, it's the same things that worked before will work again. But to go from a company doing, say, like, a billion to ten billion dollars in revenue, Requires a whole lot of change, and it is not the sort of, like, let's do last week what we did this week mindset. And in a world where people don't get time to even get caught up on the basics because growth is just so rapid, I, I badly underappreciated the amount of work it took to be able to, like, keep charging at the next big step forward. While still not neglecting everything else that we have to do. There's a big piece of internal communication around that and how you sort of share information, how you build the structures to like get the company to get good at thinking about 10 X more stuff or bigger stuff or more complex stuff every eight months, 12 months, whatever. Um, There's a big piece in there about planning, about how you balance what has to happen today and next month with the long lead pieces you need in place for to be able to execute in a year or two years with, you know, build out of compute or even, you know, things that are more nor…

AI assessment note: “I badly underappreciated the amount of work it took”

Q The question is, how do you think about hiring incredibly young, under-thirties as this, like, Trojan horse of youth energy ambition, but less experience, or the much more experienced, I know how to do this, I've done it before?

A I mean, the obvious answer is you can succeed with hiring both classes of people. Like, we have, I was just, like, right before this, I was sending someone a Slack message about There was a guy that we recently hired on one of the teams. I don't know how old he is, but low twenties probably doing just insanely amazing work. And I was like, can we find a lot more people like this? This is just like off the charts brilliant. I don't get how these people can be so good, so young, but it clearly happens. And when you can find those people, they bring amazing fresh perspective energy, whatever else. On the other hand, uh, when you're like designing some of the most complex and massively expensive Computer systems that humanity has ever built, actually like pieces of infrastructure of any sort, then I would not be comfortable taking a bet on someone who is just sort of like starting out, uh, where the stakes are higher. So you want both, uh, and I think what you really want is just like an extremely high talent bar of people at any age and a strategy that said I'm only gonna hire Younger people, or I'm only gonna hire older people, I believe would be misguided. I think it's like somehow just not, it's not quite the framing that resonates with me, but the part of it that does is, and one of the things that I feel most grateful About Y Combinator Four is inexperience does not inherentl…

AI assessment note: “the obvious answer is you can succeed with hiring both classes of people.”

Q know, wave to the internet bubble, uh, in terms of, you know, the excitement and the exuberance, and I think the thing that's different is the amount that people are spending. Larry Ellison said that it will cost a hundred billion dollars to enter the foundation model race as a starting point. Do you agree with that statement? And when you saw that, were you like, yeah, that makes sense.

A Uh, no, I think it will cost less than that, but there's an interesting point here, um, which is everybody likes to use previous examples of a technology revolution to talk about, to put a new one into more familiar context, and A, I think that's a bad habit on the whole, and, but I understand why people do it, and B, I think the ones people pick for Analogizing AI are particularly bad. So the internet was obviously quite different than AI, and you brought up this one thing about cost, and whether it costs, like, ten billion or a hundred billion or whatever to be competitive, it was very, like, one of the defining things about the internet revolution was it was actually really easy to get started. Now, another thing that cuts more towards the internet is mostly For many companies, this will just be like a continuation of the internet. It's just like someone else makes these AI models, and you get to use them to build all sorts of great stuff, and it's like a new primitive for building technology. But if you're trying to build the AI itself, that's pretty different. Another example people use is electricity, um, which I think doesn't make sense for a ton of reasons. The one I like the most, caveated by my earlier comment that I don't think people should be doing this, or trying to, like, Use these analogies too seriously is the transistor. It was a new discovery of physics. It h…

AI assessment note: “no, I think it will cost less than that”

Q Do you agree with the saying that it's like one or two decisions a year define a company? Or do you agree with the, you make 10 decisions a day, and actually it's all about the incremental little decisions that add up To the progress of a company. I'm always stuck between both mindsets.

A I very much think it's both. Um, I think there are, one of the things that I loved about being an investor was that job is really a job about one or two decisions a year, or maybe one or two decisions a decade. An operator role is definitely not my natural, this is not my natural place in the world, by the way, but in an effort to, uh, get slightly better at it, one of the things I have learned is that It is true that there are only a handful of strategic decisions. It feels more like one or two a month than one or two a year, but it's not like that many, like big, like here is the, here's the what decisions, but the like, The how decisions are, there are a lot of those. And I think people who claim there are not a lot of those have not tried to run a complex company before, because it would be ridiculous to say that any CEO makes one or two decisions a year or a month. Um, it is really nonstop. But there's a difference between, like, the big, like, we're going to do ChatGPT or we're not going to do ChatGPT, and then the, like, to make that successful along the way, in the spirit of making that one decision a successful one, here are the 10,000 little things you have to do along the way.

AI assessment note: “I very much think it's both.”

Q proceeding to ask most of them , so I'm sorry for this, but we mentioned kind of model improvement there. Like, how do we see the rate of model improvement? Is it like linear? Is it like, does it plateau at points? Obviously now it's accelerated faster than ever in the last whatever time period we want to call that. How do we see that rate of improvement in models?

A It feels very punctuated externally. Which means I think we've done a suboptimal job on one of our core beliefs. We have this idea that iterative deployment, um, is important, and what you don't want is to go build AGI in secret in a lab. This is like the limit case, toil away for a couple of decades, and then push a button, and all at once the world has to, like, contend with AGI. And better than that to us, it seems, is to put, uh, you know, a model out into the world Let people have some time to think about that, react, figure out how they want to use it, what they'd like to do differently, what they'd not like it to do, what guardrail society wants or doesn't want, and then, you know, build up sort of more, um, societal engagement with it, and I think In some sense, one of the most important decisions we ever made was this one. And that includes things like deploying ChatGPT into the world and getting the world to take advanced AI seriously, which we tried to talk about for a long time and didn't really work. And, you know, deploying that really did. But as we think about future models, uh, I think we Underestimated, because we've like lived with these models for so long, because we watch them get better and better little by little, uh, we underestimated how much, even with our strategy of iterative deployment, a lurch forward some of these things would be. So as we think a…

AI assessment note: “It feels very punctuated externally.”

Q Uh, Sam, you mentioned the research driven kind of culture and the importance to retain that. When you bring in a go to market function and sales leaders and wholesale teams, it's very difficult to blend kind of product and sales functions or cultures so efficiently. How do you think about the challenges that one faces?

A I think this is where Brad and I have a great partnership in that we Have different opinions about maybe how to balance any particular decision, and we're, I think, very good at deferring to the other based off of whoever has, like, more context or feels like it will have a more important impact, but we have really deep agreement, I think, in a way that many people in Brad's role wouldn't about the critical Um, focus of making sure that, that we let research drive product and product drive sales. Now that doesn't exclusively mean that, of course, there's gotta be feedback the other direction. And one of the reasons that we love having users now is this is like the most important reward signal you can get for if the model's good or not. It's like, how useful is it really to people like that? That's what matters. But we also know that the best thing we can do to sell more product Is to make the product better, and the best thing we can do to make the product better is to have a better, to have better research, and there's like zero disagreement between us ever on that, and that is really important.

AI assessment note: “making sure that, that we let research drive product and product drive sales”

Q What are the one or two things that you think are most important to you now then?

A There are a lot of AI orgs in the world that can, um, copy what other people do. Uh, like once you know something is possible, once you kind of know the rest shape of it, once you know that people want it, that's not so hard. Um, or it's like somewhat hard. It's really hard to figure out how to do something new for the first time. And to do that consistently, ah, over years, and hopefully, if we're lucky enough, over decades. Building a research org and a product org, and a whole company that puts these things out in the world, because we also innovate on business models and anything else. This culture of repeated innovation, so that we're not just making GPT-V amazingly great, but six, seven, eight, whatever we're gonna call those, we won't keep numbering them like that at that point. Um, making sure that we're set up to do that, Uh, from a thinking about where the researchers can take us, what that means for where the product's got to go, what that means for the whole company has to follow. Um, that's a big one.

AI assessment note: “This culture of repeated innovation... making sure that we're set up to do that”

Q Do you agree with the saying that it's like one or two decisions a year define a company? Or do you agree with the, you make 10 decisions a day, and actually it's all about the incremental little decisions that add up To the progress of a company. I'm always stuck between both mindsets.

A I very much think it's both. Um, I think there are, one of the things that I loved about being an investor was that job is really a job about one or two decisions a year, or maybe one or two decisions a decade. An operator role is definitely not my natural, this is not my natural place in the world, by the way, but in an effort to, uh, get slightly better at it, one of the things I have learned is that It is true that there are only a handful of strategic decisions. It feels more like one or two a month than one or two a year, but it's not like that many, like big, like here is the, here's the what decisions, but the like, The how decisions are, there are a lot of those. And I think people who claim there are not a lot of those have not tried to run a complex company before, because it would be ridiculous to say that any CEO makes one or two decisions a year or a month. Um, it is really nonstop. But there's a difference between, like, the big, like, we're going to do ChatGPT or we're not going to do ChatGPT, and then the, like, to make that successful along the way, in the spirit of making that one decision a successful one, here are the 10,000 little things you have to do along the way.

AI assessment note: “I very much think it's both.”

Q might be a little bit boring, but it's just the commoditization of models, and I've never seen them before where you have like, Mistra one week, so hyped, and then you have, you know, whatever bar the next week, and it's like the transience of different players being preceded in the media as kind of winning, so to speak, is so moving every week. Is this a game of commoditization?

A There was a time when There were, like, more than a hundred car companies in the US, I believe, or at least close to that, and if you go, like, look at some of the old media at the time, it was like, no, there's this better car, no, there's this better one, no, there's this better one, and I think that same thing holds true for most new industries. I think it's fine. I mean, it's probably good, uh, but I don't think that's where the enduring value will be. I think eventually it will shake out. There will be a small number of Providers, just a relatively small number, you know, dozen, something like that, doing models at big scale, and it'll be extremely complex, extremely expensive, and the differentiation, and I hope we all continue to push each other to make the models better, cheaper, faster, and commoditize in that sense, and the long-term differentiation will not be I don't think the base model, like that's just, you know, intelligence is just like some emergent property of matter or something. Uh, the, the long-term differentiation will be the model that's most personalized to you, that has your whole life context, that plugs into everything else you want to do, that's like, Well integrated into your life. Um, but for now, the curve is just so steep that the right thing for us to focus on is just make that base model better and better.

AI assessment note: “I hope we all continue to push each other to make the models better, cheaper, faster, and commoditize”

Q proceeding to ask most of them , so I'm sorry for this, but we mentioned kind of model improvement there. Like, how do we see the rate of model improvement? Is it like linear? Is it like, does it plateau at points? Obviously now it's accelerated faster than ever in the last whatever time period we want to call that. How do we see that rate of improvement in models?

A It feels very punctuated externally. Which means I think we've done a suboptimal job on one of our core beliefs. We have this idea that iterative deployment, um, is important, and what you don't want is to go build AGI in secret in a lab. This is like the limit case, toil away for a couple of decades, and then push a button, and all at once the world has to, like, contend with AGI. And better than that to us, it seems, is to put, uh, you know, a model out into the world Let people have some time to think about that, react, figure out how they want to use it, what they'd like to do differently, what they'd not like it to do, what guardrail society wants or doesn't want, and then, you know, build up sort of more, um, societal engagement with it, and I think In some sense, one of the most important decisions we ever made was this one. And that includes things like deploying ChatGPT into the world and getting the world to take advanced AI seriously, which we tried to talk about for a long time and didn't really work. And, you know, deploying that really did. But as we think about future models, uh, I think we Underestimated, because we've like lived with these models for so long, because we watch them get better and better little by little, uh, we underestimated how much, even with our strategy of iterative deployment, a lurch forward some of these things would be. So as we think a…

AI assessment note: “It feels very punctuated externally... we watch them get better and better little by little”

Q What do you think is the biggest barrier to that happening?

A I think the models are just not smart enough, which sounds like a annoying, low information kind of cop-out answer, but I think it's like deeply fundamentally true. Like the models just aren't smart enough. You fix that one thing, all these other things get better. There will be all these ways that we have to figure out how to integrate tools into people's workflow and, you know, modelability in different areas. Well, Will matter a lot, but if you zoom out, you know, doing scientific research with the help of GPT-II would have seemed fairly laughable. With GPT-IV, people do use it just in very, to help them do science, just in extremely primitive and limited ways. And with GPT-VI, I think people will say, hey, this is like helping me as a general purpose tool in all these ways. And then with GPT-VIII, maybe people are like, you know, this can do some limited, maybe not so limited tasks for me.

AI assessment note: “I think the models are just not smart enough”

Q You said also before that the founders that do the very best have this extreme paranoia, kind of full of existential crises, to quote you. Walk me through this, and what do you mean by that kind of existential crises and paranoia?

A So, it is easy to be the wrong kind of paranoid, like actually paranoid, where you think people are spying on you all the time, and you don't trust your own employees, and that's bad. The kind of paranoia that I meant in that quote is always thinking about What can go wrong with the product or the strategy? And it's just sort of this constant thinking through all of the branches of the tree, and how a strategy could go wrong, or how a product could go wrong, or how a competitor could beat you. And there is just this, I'd say, extreme alertness concern that really good founders have, where they never get too comfortable, even when things are going really well. And they're always aware that they're sort of only as good as their next decision.

AI assessment note: “The kind of paranoia that I meant in that quote is always thinking about”

Q So let's start with you and PG, widely discussed and agreed upon as one of the great leaders of the Valley over the last decade, but I'd love to hear then. You've seen him since batch one. What do you think makes PG the truly special leader that he is?

A I think one of the things that is in short supply in Silicon Valley these days, and certainly required of great leaders, is Conviction around ideas that are right, but not consensus, and the willingness to sort of keep doing the thing that you believe will work. Convince other really good people to come do that with you, even in the face of a lot of other people claiming that it won't work, or that it's bad, or that it makes no sense. And when Y Combinator started itself, it was sort of regarded as this very dumb thing that was never going to yield big companies. Incubators have been tried that didn't work, and Y Combinator was funding all these things that It wasn't going to make sense, and it didn't understand that you needed sort of experienced CEOs, and it was, at the time, these things that are now sort of accepted wisdom of Silicon Valley were super controversial, or people just thought they were wrong.

AI assessment note: “Conviction around ideas that are right, but not consensus”

Q you, you know, you've got particular interest in everything from space rockets to computer brain interfaces to nuclear fission. I'm intrigued as to kind of the vertical expansion that's enacting at YC. You previously said that most people do too many things and you should do, do few things relentlessly. How do you think about that kind of in contrast with the vertical expansion that you're also enacting with YC?

A Well, The thing that I am doing is figuring out how to scale YC. You know, I'm not running a continuity fund. I'm not running the accelerator. I'm not advising all these companies. I'm not running our admissions process. There'd be more things than I could do. So the thing I try to really do is just be thoughtful about how YC develops into this organization that we're trying to build. And a big part of that is we really believe that the startup model applies to a lot of different areas. And the thing that I always say is let's find the ten billion dollar companies The companies that could be ten billion dollar companies, and that that is such a difficult constraint. Those are so rare. We can't have any other constraints. And also, after you do something for a while and it starts to work, that area gets so flooded with other people trying to do the same thing that you've got to continually push into new areas to find these sort of undiscovered returns. I think it's always a good sign we start doing something new and people say, well, that's, that's crazy. There's, that's not going to work. I'd never touch that. That makes me lean into it more. You know, when, when someone says like, oh, that's a crazy idea, or oh, anyone that does that is stupid, I pay extra attention.

AI assessment note: “The thing that I am doing is figuring out how to scale YC.”

Q So let's start with you and PG, widely discussed and agreed upon as one of the great leaders of the Valley over the last decade, but I'd love to hear then. You've seen him since batch one. What do you think makes PG the truly special leader that he is?

A I think one of the things that is in short supply in Silicon Valley these days, and certainly required of great leaders, is Conviction around ideas that are right, but not consensus, and the willingness to sort of keep doing the thing that you believe will work. Convince other really good people to come do that with you, even in the face of a lot of other people claiming that it won't work, or that it's bad, or that it makes no sense. And when Y Combinator started itself, it was sort of regarded as this very dumb thing that was never going to yield big companies. Incubators have been tried that didn't work, and Y Combinator was funding all these things that It wasn't going to make sense, and it didn't understand that you needed sort of experienced CEOs, and it was, at the time, these things that are now sort of accepted wisdom of Silicon Valley were super controversial, or people just thought they were wrong.

AI assessment note: “Conviction around ideas that are right, but not consensus”

Q You said also before that the founders that do the very best have this extreme paranoia, kind of full of existential crises, to quote you. Walk me through this, and what do you mean by that kind of existential crises and paranoia?

A So, it is easy to be the wrong kind of paranoid, like actually paranoid, where you think people are spying on you all the time, and you don't trust your own employees, and that's bad. The kind of paranoia that I meant in that quote is always thinking about What can go wrong with the product or the strategy? And it's just sort of this constant thinking through all of the branches of the tree, and how a strategy could go wrong, or how a product could go wrong, or how a competitor could beat you. And there is just this, I'd say, extreme alertness concern that really good founders have, where they never get too comfortable, even when things are going really well. And they're always aware that they're sort of only as good as their next decision.

AI assessment note: “The kind of paranoia that I meant in that quote is always thinking about What can go wrong”

Q you, you know, you've got particular interest in everything from space rockets to computer brain interfaces to nuclear fission. I'm intrigued as to kind of the vertical expansion that's enacting at YC. You previously said that most people do too many things and you should do, do few things relentlessly. How do you think about that kind of in contrast with the vertical expansion that you're also enacting with YC?

A Well, The thing that I am doing is figuring out how to scale YC. You know, I'm not running a continuity fund. I'm not running the accelerator. I'm not advising all these companies. I'm not running our admissions process. There'd be more things than I could do. So the thing I try to really do is just be thoughtful about how YC develops into this organization that we're trying to build. And a big part of that is we really believe that the startup model applies to a lot of different areas. And the thing that I always say is let's find the ten billion dollar companies The companies that could be ten billion dollar companies, and that that is such a difficult constraint. Those are so rare. We can't have any other constraints. And also, after you do something for a while and it starts to work, that area gets so flooded with other people trying to do the same thing that you've got to continually push into new areas to find these sort of undiscovered returns. I think it's always a good sign we start doing something new and people say, well, that's, that's crazy. There's, that's not going to work. I'd never touch that. That makes me lean into it more. You know, when, when someone says like, oh, that's a crazy idea, or oh, anyone that does that is stupid, I pay extra attention.

AI assessment note: “The thing that I am doing is figuring out how to scale YC.”

Q Can I ask you a really weird one? I had a great quote the other day, and it was, the heaviest things in life are not iron or gold, but unmade decisions. What unmade decision weighs on your mind most?

A It's different every day. Like I don't, there's not one big one. I mean, I guess there are some big ones that, like, about are we gonna bet on this next product or that next product, uh, or are we gonna, like, build our next computer this way or that way that are kind of, like, really high stakes one-way door-ish that, like everybody else, I probably delay for too long. But, but mostly the hard part is every day it feels like there are a few new 51 49 decisions that come up that Kind of make it to me because they were 51 49 in the first place, and then I don't feel like particularly likely I can do better than somebody else would have done, but I kind of have to make them anyway. And it's, it's the volume of them. It is not anyone.

AI assessment note: “there's not one big one. I mean, I guess there are some big ones”

Q Can I ask you a really weird one? I had a great quote the other day, and it was, the heaviest things in life are not iron or gold, but unmade decisions. What unmade decision weighs on your mind most?

A It's different every day. Like I don't, there's not one big one. I mean, I guess there are some big ones that, like, about are we gonna bet on this next product or that next product, uh, or are we gonna, like, build our next computer this way or that way that are kind of, like, really high stakes one-way door-ish that, like everybody else, I probably delay for too long. But, but mostly the hard part is every day it feels like there are a few new 51 49 decisions that come up that Kind of make it to me because they were 51 49 in the first place, and then I don't feel like particularly likely I can do better than somebody else would have done, but I kind of have to make them anyway. And it's, it's the volume of them. It is not anyone.

AI assessment note: “It's different every day. Like I don't, there's not one big one.”

Q Uh, Sam, you mentioned the research driven kind of culture and the importance to retain that. When you bring in a go to market function and sales leaders and wholesale teams, it's very difficult to blend kind of product and sales functions or cultures so efficiently. How do you think about the challenges that one faces?

A I think this is where Brad and I have a great partnership in that we Have different opinions about maybe how to balance any particular decision, and we're, I think, very good at deferring to the other based off of whoever has, like, more context or feels like it will have a more important impact, but we have really deep agreement, I think, in a way that many people in Brad's role wouldn't about the critical Um, focus of making sure that, that we let research drive product and product drive sales. Now that doesn't exclusively mean that, of course, there's gotta be feedback the other direction. And one of the reasons that we love having users now is this is like the most important reward signal you can get for if the model's good or not. It's like, how useful is it really to people like that? That's what matters. But we also know that the best thing we can do to sell more product Is to make the product better, and the best thing we can do to make the product better is to have a better, to have better research, and there's like zero disagreement between us ever on that, and that is really important.

AI assessment note: “we let research drive product and product drive sales”

Q And so I wanted to hear from each of you, like an all-star Mr. and Mrs. Like, what is Brad amazing at that the world doesn't know?

A Look, I think one of the, one sign of like a good partnership, I'm thankful to have this with like a lot of the key people at OpenAI, certainly Brad is like, um, if you can't do each other's job, Maybe Brad could do my job for a week. I certainly could not do Brad's job for a week. Um, and I, I think that ability to divide up as a team, um, and have a very high bandwidth communication channel with each person and all together as a leadership team is super important. Brad is good at a lot of things. Uh, I'll talk about just two here in the interest of time. One is adaptability. Uh, Brad joined, uh, To do finance, obviously, and now does something, I guess it's like in the sphere of finance, but very, very different. Um, we didn't have a business at all, or we didn't have an appreciable business until very recently. And when it became clear that we were going to have a very fast growing business, um, kind of like looked around, I was like, I really need somebody. We gotta, we gotta get someone to do this. And, uh, I kind of like looked around the room and I asked Brad to do it. And it was, and he was just like, okay, I'll figure it out. Like, I'll, you know, Uh, just like, you know, I might need, like, a little bit of time to get up to speed, but this is, you know, I've done, like, business-ish stuff before and can go, like, build all this out. So the willingness to just, like, t…

AI assessment note: “One is adaptability. Uh, Brad joined, uh, To do finance, obviously, and now does something”