Why Sam Altman is against a single superintelligence

Sam Altman, CEO of OpenAI, joins John to break down how models uncovered critical PHP vulnerabilities, why test-time compute unlocks massive agent swarms, and OpenAI's internal experiments proving complex math problems like Navier-Stokes. He also shares why sandboxing superhuman models is an unsolved safety dilemma, how robotics will scale future gigawatt data centers, and why businesses must adapt as personal agents replace traditional web forms and mobile apps. Timestamps (00:00:00) Introduction (00:00:12) PHP week (00:02:37) Internet early days (00:07:33) Cyber Blue (00:08:54) Sandboxing (00:15:47) Math (00:23:12) Agent swarms (00:24:48) Stripe Atlas (00:25:17) Compute (00:37:53) Revenue per gigawatt (00:49:27) Token pricing is fake

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00John Collison?

  We're going off script. We're just having a beer. It's not the first non-Guinness, but it is the first IPA. And so it's a first time for everything. Cheers, cheers.

• 0:12John Collison?

  I want to ask about PHP Week.

• 0:15Speaker 01

  Yeah.

• 0:16John Collison?

  First off, how did it? Why did it take so long to find these bugs? Because it sounds like they were kind of simple.

• 0:23Speaker 01

  Why did it take like the 20 years or why did it take the last two?

• 0:25John Collison?

  No, why did it like, exactly, why did it take a year? Maybe start by explaining the bug or the problem.

• 0:31Speaker 01

  So the problem is that although we have known for a while that these models are getting good at finding vulnerabilities in the wild.

• 0:41Speaker 01

  A, they continue to get better fairly rapidly. And B, people just hadn't really used them to go look at everything out there. There's like a lot of surface area and a lot of code on the internet. And as people have started to do that, I think we're going to find that we are in a a world of many more serious vulnerabilities than we expect, even though many people who have been following this for a while expect a lot.

• 1:06John Collison?

  Yeah, I just feel like a lot of the discussion ends up being around, we need to fix the software. And sure, yeah, we should fix the software. But also, we maybe need to change the architectures, which is harder, but a much more durable fix.

• 1:17Speaker 01

  Like, that seems to me like what needs to happen.

• 1:19John Collison?

  Yeah.

• 1:20Speaker 01

  It would seem really hard, except we have this amazing new tool to do it with. So I kind of think that's what's going to happen. And you're already seeing like a new generation of startups talking about building this like new kind of internet or new kind of application architecture for this world, where you assume that the models get smarter and smarter.

• 1:37John Collison?

  What do you think the global software upgrade looks like? For example, is a PHP codebase just a bad idea? And you should rewrite a nice, clean version in Rust? Or do we upgrade PHP? But yeah, as we, everyone thinks that we need to kind of upgrade all hand-spun software. Just what form will that actually take? Is this huge for Infosys and all those guys? Because it's going to be a lot more software consulting project.

• 2:03Speaker 01

  I mean, people are talking more and more about just like, let's re-architect what it means to be a router or an operating system or an application. And, you know, I've seen some of these new startups saying that they're just going to like make a whole new stack. And that would have seemed insane two years ago. You would have been like, there's no way, there's too much, you can't write all that. And now maybe soon you can. I would bet that we have a completely new sort of model of how we make systems and networks.

• 2:37John Collison?

  What else reminds you of the early internet days?

• 2:39Speaker 01

  The thing that most reminded me of the early internet was, you know, there is, there is this like spirit of just like unbelievable creative unleash

• 2:50Speaker 01

  in people using these tools. You see it in kids particularly. And like, and again, maybe this is just like, because I remember what it was like when I first got a computer with internet and I could like do all these things that adults just couldn't even think to ask for. And like, watching a, you know, eight or 10 year old make a complete video game with some crazy ideas they have that like no adult would really have. And expect that to just be the way that technology is going to work for the rest of their lives. In fact, getting better and better. And just like this constant, I'm going to make useful stuff for me and show it to my friends.

• 3:22John Collison?

  Yes.

• 3:23Speaker 01

  Like, we had to suffer when I was a kid to make a piece of software.

• 3:25John Collison?

  Yes.

• 3:26Speaker 01

  They barely do anything. And now you can have like a full 3D video game, like unbelievable, anything you ever think of in 10 minutes. That reminds me of the early internet, including that just like these people, they're never going to be mind blown by that, because that's always what they expect out of the world. Like any,

• 3:42Speaker 01

  any ideas they can come up with, anything they want to create, anything that they think might be useful, anything that they just need to make their lives better, like the computer's supposed to make it. There was a crazy moment at OpenAI where we had a sev and Codex went down.

• 3:58John Collison?

  And everyone had forgotten how to code.

• 4:00Speaker 01

  I mean, not all the way, but it was like a whoa, what if?

• 4:03John Collison?

  Yes.

• 4:03Speaker 01

  I remember when I was a student and I would like meet the really old guys.

• 4:09Speaker 01

  And they would like, once in a while, you know, pull out their gigantic Xerox Alto disk. Or they would like pull out their like, you know, with like line by line written assembly, or they would pull out their punch cards. And, and I felt like it is incredible to me

• 4:25John Collison?

  Yes.

• 4:25Speaker 01

  that you all bootstrapped us to this point.

• 4:27John Collison?

  Yes.

• 4:27Speaker 01

  Doing it this way.

• 4:28John Collison?

  Yes, no, I agree. And keeping programs in your head, right? You know, programs running, programs running the first time after, you know, hundreds, thousands of lines were written. No, if you lost arts like that. Go back to the cyber side of things. In retrospect, you know, was the cyber RL a mistake? Or did we just have to go there? Because it's a little like the SNL thing of, you know, we urge you to urge us to stop. It's like the models are getting very good at cyber capabilities. Like, yeah, they are. I mean, they are getting very good, but not just totally emergently.

• 5:03Speaker 01

  Look, the reason that SNL thing was so funny is because it was so true. Like, humor like that doesn't stick unless there's an element of like, it's like clearly pointing out a ridiculous but true thing about society or someone. And,

• 5:17Speaker 01

  and the cyber thing, there is a point which is like, okay, you know, sure, the models are really good at cyber. But like, who asked for this? Why are you doing this? Like, one of my big disappointments, criticisms are, or our field or whatever. And I think maybe the cyber, models getting good at cyber is an example of this is, I don't think we have done a great job of explaining to people why we think all of the potential downside of this technology is worth it. Like, the kind of caricature of this is people feel like, we're like, all right, you know, fine. You say the model's going to cure cancer someday, maybe it will. And if so, I guess that's good. And you say it's going to like make stuff cheaper. And okay, maybe I already see that's good and I get access to medical advice or legal advice.

• 6:00Speaker 01

  But, but then you paint this picture of the future where it's like a new industrial revolution. And maybe the last one wasn't that good and this one sounds worse. And I will have no job, no ability to like have control or choice or freedom of my future. And AI is so dangerous that you're going to have to like have a crazy surveillance state so we don't all die.

• 6:22John Collison?

  I'm with you so far.

• 6:24Speaker 01

  This is like a terrible sales pitch.

• 6:25John Collison?

  Yeah.

• 6:26Speaker 01

  And one of the things that I am most proud of about OpenAI is even when it has been hugely unpopular in the AI safety community for us to push for iterative deployment and putting

• 6:40Speaker 01

  these tools in the hands of people. We have really stuck to that principle. And I think this is the only kind of good path. I think people, it's not enough to tell people like, you're going to get the model to do these great things for you. It's, it's got to be like, we are going to enable you to have a better life that you define for yourself to be, to do useful things. You're not going to just sit and, you know, get a basic income and have a bunch of people sitting in a few square miles of San Francisco make all the decisions about the future.

• 7:08John Collison?

  Iterative deployment is this idea that you just have to get the advances out there? And you can't kind of stopper them up inside?

• 7:14Speaker 01

  I mean, not all like, you still have to make them like safe before you put them out there. You can't like cause havoc in the world. But, but the, but the building up a capability overhang is a thing to fight against. That directionally speaking, yes, you want to get this into the world. You want to get this into people's hands and you want to let people use the technology to figure out all these wonderful things.

• 7:32John Collison?

  The exploits get a lot of headlines. The prevented exploits do not. How quickly are people making progress with CyberBlue?

• 7:42Speaker 01

  I don't want to declare victory at all in the PHP thing, because I still don't think we know. And I assume some bad things are going to happen. Relative to my worst fears, if I can like think back 10 days ago, the first day of that went much better than I expected.

• 7:58Speaker 01

  And we heard a bunch of stories from people that were using our models to, I mean, not only build patches and build cyber defense, but like individual companies who were just like, I need to just do a quick scan deployment. Like, I'm going to take my thing down for an hour. I'm going to use this thing and I want to put it back up because my application's offline. Now I feel like I really jinxed it and I shouldn't have said that. But the first day did go a lot better than I thought. And I think people have been able to use that. The, the thing that I've heard a lot of cyber experts talk about is

• 8:26Speaker 01

  they just no longer are confident that we have the right assumptions baked into our threat model. And they're like, look, if, if everything is what we think and hardware exploits are as hard as we think they are, then yes, we can see how this is going to work. But if AI is going to discover really deep novel bugs in hardware that require physics we didn't understand, but all of a sudden means some assumption doesn't hold, then I think people aren't sure what happens.

• 8:53John Collison?

  One thing that always strikes me is that when I talk to very knowledgeable people inside OpenAI and similar companies, there is a pessimism that sandboxing is possible, basically. And that would not have been my mental model going in. It would have been this, yeah, you have to build a really tight sandbox. And we haven't kind of hardened them historically enough compared to the current capabilities. But ultimately, you can have a maximum security prison for the model to work in and keep it inside there. It seems like you guys have kind of a pessimism that sandboxing models is possible.

• 9:35Speaker 01

  My instinct, well, first of all, I think we knew much better than we're doing now and we should. I think we clearly, that is the main thing to focus on right now, is a better security model. If you imagine a theoretical AI that is like a million times smarter than the smartest security researcher, would you like really be willing to trust that sandbox? Especially if the model can like discover new science that people don't already?

• 10:01John Collison?

  But you can use, you can use the models to construct the sandboxes.

• 10:03Speaker 01

  I think you can, well, so this gets, let me come back to that.

• 10:06John Collison?

  Yeah.

• 10:08Speaker 01

  I think sandboxing works to a point.

• 10:11Speaker 01

  Then I think monitoring works to a further point. Where we can look at what the models are doing, not only in their chain of thought, but we can do activation monitoring and try to really understand like, are the models, are we happy that the models are not deceiving us? And then

• 10:26Speaker 01

  eventually, I think you need to believe with huge confidence that the models are truly aligned and are going to value and try faithfully to do, to follow their actual, the intent of the people operating them. And that they're going to value sort of people being in control of the future forever. And if you had a model that you could fully trust the alignment, that you were absolutely certain of the alignment of, then yes, I think you can have that to help you build better and better sandboxes. But like, if you really let yourself project forward into the like, what does it mean to have something a million times smarter than we are?

• 11:05Speaker 01

  I don't know, like, I think that just gets strange.

• 11:11John Collison?

  Because the alignment becomes hard.

• 11:12Speaker 01

  Because you need to really, really trust the alignment to trust that these sandboxes stay ahead of capabilities.

• 11:20John Collison?

  Can you not give models competing incentives and give them incentives to defect from each other?

• 11:24Speaker 01

  I think this is a promising direction.

• 11:28Speaker 01

  I suspect that this idea, I hate the word agent swarms, let's call them agent teams.

• 11:35John Collison?

  But it seems to be taking over.

• 11:37Speaker 01

  I suspect that teams of agents have very interesting scaling law properties, but also very interesting alignment properties.

• 11:46Speaker 01

  And I think that's like a direction we should push in more, not just from the capabilities, but from a safety perspective, too.

• 11:53John Collison?

  I noticed when you talk about agent teams, not swarms.

• 11:57Speaker 01

  A swarm is such a scary.

• 11:59John Collison?

  I know, it's a locus, yeah. Not good, not good branding, in keeping with many historical AI industry traditions.

• 12:06Speaker 01

  We are so good at this.

• 12:08John Collison?

  We'll get back to branding, but I noticed that there seems to be a big gap in what the leading AI users are doing and kind of standard corporate use in the use of swarms, where like all the advanced math, sorry, teams, all the advanced math uses agent teams, all the advanced hacking uses agent teams, and then like common or garden corporate stuff is not using big swarms. Why has that not, because we've had this for many months.

• 12:39Speaker 01

  Not to the same degree, like the one way to look at this problem or this question is how much compute can we productively throw at a single problem? And, you know, a few years ago, we forget how fast this has all gone, but a few years ago, it was like impossible to throw more than a dollar of compute at a problem and have it help you at all.

• 13:00John Collison?

  Why?

• 13:02Speaker 01

  You could have GPT-4 like grind out for you, you know, like 100,000 tokens or something, and it was going to answer your question, and it certainly was not going to be able to go right productively, like write you a piece of software.

• 13:16Speaker 01

  We just, we had not figured out how to use a lot of test time compute to get you a better result. And then we figured out chain of thought and the model's good thing card about a problem. And now, and then like, people really started using a lot of compute. And now you, you can throw a whole team at something and you can do that, you know, 10,000 times in parallel.

• 13:36John Collison?

  You need chain of thought and test time compute before you can get to swarms, because you're kind of a useful swarm with a bunch of GPT-3s.

• 13:42Speaker 01

  But so the effect of this, I think this is not being well followed, paid attention to in the world, but the incredible exponential of how much

• 13:55Speaker 01

  compute you can throw at a single problem and get better results is one thing. The delta between like what most people are using AI for and what the people are using AI the most productively for has just continued to spread and spread and spread.

• 14:08John Collison?

  How do you allocate compute budgets now? Because presumably, even though researchers are paid a lot, it's eclipsed by compute. And so capital allocation at OpenAI is sort of compute allocation. And it's hard to a priori know how much it will cost to go do something. It's actually also hard to just like control the, like, you can't quite prompt an agent where it's like, go do a really good job of this and don't spend more than $10,000.

• 14:40Speaker 01

  You can.

• 14:41John Collison?

  Okay.

• 14:42Speaker 01

  Yeah, if you say do the best job you can with $10,000 of compute, like, yes, it will understand that. I mean, there's like a, there's another kind of like hard question for society about what does it mean when you have two companies competing? And one says you can spend 10,000 and one says you can spend 100,000 on the same problem. Like, it's clear what's going to happen there, but there's like a question about how society thinks about

• 15:04Speaker 01

  compute allocation and wealth and things like that, but we have our own version of this inside of OpenAI. We allocate, you know, a lot of compute to internal productivity. And then people get very wide latitude to make their own decisions unless they're really being abusive.

• 15:23John Collison?

  But you just encourage people to like do useful stuff, and if they spend $10,000 in compute, it's like, oh, interesting.

• 15:28Speaker 01

  Correct. We try never to reward people for using a lot of compute, sure, like no token leaderboards. It's actually slightly bad to be high on the token leaderboard, certainly on a promotion. But, but it's like, it's a skeptical default, and if you can show that you're doing something important, you will have nothing but appreciation.

• 15:47John Collison?

  But at a certain point, this will need to change, as, because you're describing the increasing orders of magnitude. And right now, you have a researcher who does some very interesting math for $10 million. And you're like, ah, little pricey, but cool math. And then $100 million, and you know, when they're doing the billion dollar math, you're like, look, I'm really happy that you're making so much math progress, but I really need you to cool the jets with the math.

• 16:09Speaker 01

  We did let it get out of control for like one week on the math, intentionally so, because you have to be like.

• 16:13John Collison?

  And it worked.

• 16:14Speaker 01

  Well, you don't know how, you don't know how far it's going to go until you try. And I think it's important to have, when there's a new paradigm, I think it's important to let people be fairly unconstrained. So when we did the Navier-Stokes proof, we understood absolutely nothing about the scaling laws of teams of agents. Are you supposed to like?

• 16:35John Collison?

  The very recent Navier-Stokes.

• 16:36Speaker 01

  Yes. Are you supposed to give them like the smartest model? Are you supposed to give them like many copies, fewer copies, longer rollouts, shorter rollouts? We did not understand that at all. And now we understand it a little bit, I think we can still understand it like much better. But like has happened in every other new paradigm we've discovered, I am confident that we will find our, you know, 10x improvements per year at how we allocate dollars or computer flops, whatever you want to measure, into useful things. So I think this is a thing that our company is particularly good at, but also like fits an important property of the field, which is let people explore at the limit of what is possible

• 17:17Speaker 01

  and then trust that the incredible optimization pressure will figure out how to do this in an efficient way. I, and, you know, directionally speaking, I would rather us be more willing to let people use crazy amounts of compute and let users have extremely high-end plans to figure out the edge of the possible and then figure out and make that efficient at a mass scale.

• 17:43John Collison?

  But when did the optimization function come in?

• 17:45Speaker 01

  Well, I think we're already doing it. I mean, like, I,

• 17:49Speaker 01

  I bet we could redo Navier-Stokes at 1/100 of the cost a year from when it was done. So we did it in like.

• 17:57John Collison?

  Not just knowing how the proof works, but like just with better approaches.

• 18:00Speaker 01

  Yes, smarter models and better approaches.

• 18:03John Collison?

  How much did all the math cost?

• 18:04Speaker 01

  Order 10 million per like, for like Navier-Stokes, I think.

• 18:08John Collison?

  Okay.

• 18:08Speaker 01

  Maybe a little more.

• 18:09John Collison?

  10 million-ish, that order of magnitude, yeah, yeah, but for one proof, and so times 700?

• 18:16Speaker 01

  Well, no, no, because then we kept getting better so fast, like we understood more, like.

• 18:18John Collison?

  Got it.

• 18:19Speaker 01

  We spent way less than that on the average one of those.

• 18:22John Collison?

  But like hundreds of millions of dollars of math.

• 18:24Speaker 01

  Yeah, yeah, that kind of work.

• 18:26John Collison?

  Interesting.

• 18:26Speaker 01

  But that include like, that includes like much more other discovery than what we have published.

• 18:31John Collison?

  Yes.

• 18:32Speaker 01

  So it was like, maybe we spent that much.

• 18:33John Collison?

  We have more than the 700 proofs. When did you get those?

• 18:37Speaker 01

  We are trying to learn how to, this is like a real tension. I think on one hand, very clearly, and I feel unequivocal on this point, humanity deserves fundamental knowledge, and we should not be sitting on a major fundamental science discovery.

• 18:57Speaker 01

  Like, the world deserves that and gets to know, and I think that outweighs another very important equity, which is

• 19:08Speaker 01

  if you are a scientist, your career and, you know, a significant amount of your enjoyment, fulfillment in life comes from discovering science. And so if we're going to continue to do this for math or begin to do this for other fields, then I think we more than owe it to the people that work in those fields to figure out, you know, what is the solution space where the world gets to benefit from the knowledge? And they also get the best version of the future that they can have. We are working through that right now.

• 19:39John Collison?

  I did not, I don't know, I think it may be too contrite. I did not interpret that moment as particularly bad for AI, where you had, I think everyone saw the mathematicians' union up in arms. And I think everyone else says, yeah, I mean, have you paid attention to 2026?

• 19:55Speaker 01

  As I said, I don't think it was bad for AI, and I think, I think on net, the world deserves to understand how the universe works. But I do have empathy for the math, like, I think we need to do this, and we should do this. I think like if we can go understand how to, you know,

• 20:12Speaker 01

  a theory of all of physics or how to cure a bunch of diseases, like clearly, that is knowledge the world should have. But I would like to figure out a way to do this.

• 20:23Speaker 01

  That is as sensitive to the communities that have driven these fields forward for centuries as we can. And it's just, it's a hard thing.

• 20:31John Collison?

  I wasn't that worried about the math. I thought it was great.

• 20:33Speaker 01

  But you agree that mathematicians are not that happy?

• 20:35John Collison?

  Yeah, it's fine. Like, the Longshoreman Union weren't happy about the porous automation. Remember the guy?

• 20:44Speaker 01

  No, no, no, I do. Why do I feel differently about this? There was one weekend when we got a new model and in, I'm actually probably not supposed to say this, but I will since we're drinking.

• 20:53John Collison?

  Exactly, we're half pint in, yeah. Perfect time.

• 20:55Speaker 01

  The name of the channel was model name dash hold my beer. And people were just throwing that at the hardest problems they could. And people just like didn't think the model was going to do these things. And we had this collective experience of people's like hardest research problems from their academic career, the model was just knocking down one after the other, many different fields. It was incredible to watch. And kind of two nights in a row, I was up most of the night because I was so excited that I couldn't go to bed. It was like, many of those problems that we published, many other things. And I was tremendously excited and happy. But I also had like much more sadness and melancholy and like a feeling of like

• 21:38Speaker 01

  something ending than I was expecting. So, again, I agree, this is the right thing on behalf of the world. And also, I agree that this is just like the shape of reality and nothing that anyone can stop or should stop. But I do have like real.

• 21:57John Collison?

  I think it's a wistfulness and nostalgia you can feel.

• 22:00Speaker 01

  It's more than nostalgia.

• 22:01John Collison?

  Wistfulness.

• 22:02Speaker 01

  Yeah.

• 22:03John Collison?

  Like, my dad describes how, you know, growing up in Ireland in the 1950s, he observed people around him still using horses productively in industry and.

• 22:13Speaker 01

  In the 50s?

• 22:13John Collison?

  Yeah. And, I mean, not, obviously, they were also tractors, but you know, it was being phased out. And, but it still for sure happened. And observing that this was a technology that humans had been using for thousands of years, and he was witnessing the very tail end of it. And yeah, I think there's always that feeling of witnessing the closing of a chapter that leads to some wistfulness, but it's okay.

• 22:41Speaker 01

  It's like, it's people's lives and it's something that has brought them great joy. And, you know, maybe most, most, I mean, certainly, most mathematicians never solve a leading prize problem, but the idea that you might be able to, and the idea that there is stuff out there that we don't understand yet that through great effort and, you know, pushing aside other things in your life and dedicating yourself to, you can still solve and understand, like, there's something beautiful about that.

• 23:09John Collison?

  What else are you excited about from a safety point of view?

• 23:12Speaker 01

  We talked about this a little bit earlier, but I want to like emphasize it again. I, I'll just call it swarms since that seems like the term right now. I do think the properties of swarms match a lot of the properties of the resilience of society or the robustness of society that, that I kind of, I think we're just missing from the like single rollout, single brained version of the world. Like, for all of the deep flaws of society, it is unbelievably robust. And this would not happen without like a lot of people keeping each other in check and collaborating in kind of these values of pluralism and democracy. And I had been sort of sad that that was gonna trend away in AI.

• 23:57Speaker 01

  And now I, like, I'm excited about this as a direction.

• 24:00John Collison?

  You're saying basically it shouldn't lead in a centralized direction because you will have many agents and agents pushing against each other in a dynamic equilibrium. And, you know, offense, defense is a good example of this.

• 24:13Speaker 01

  I would much rather intuitively have

• 24:18Speaker 01

  a trillion AIs of like slightly superhuman intelligence join society than one AI of one trillion IQ join society.

• 24:27John Collison?

  Yeah. And you think we're headed in that direction of the multiple?

• 24:31Speaker 01

  All of these things are things we have to push for. Like, all of these things are like, we have this incredible technology.

• 24:37Speaker 01

  It can be pointed and deployed in many ways. I think there are a lot of people who may have a different opinion about where it should, how it should be used and what is the best version of the world. But this is the one that I think is right.

• 24:51John Collison?

  Let's talk about starting a business. Thanks to coding models, your weekend side hustle can quickly become a serious undertaking. And we've been busy here at Stripe making that process as simple as possible. Stripe Atlas helps you with everything you need to start a business, like company formation, tax ideas, founder equity, and all the other bits that usually require someone in a suit. So if you've got a company in the making, take a look at stripe.com/athlus.

• 25:17John Collison?

  How's compute shortage going?

• 25:21Speaker 01

  I mean, there are all these ways where I think things

• 25:27Speaker 01

  at various times along the curve are going to get somewhat out of hand. And there will be moments, like with any new thing of this importance, there are going to be moments of irrational exuberance, there are going to be moments of bust, there are going to be moments of constraint and glut and up and down. But I think the trend is toward a, we always just want more compute. Like, the demand for sufficiently capable models at a sufficiently low price, I think that trend only goes one way. And we're doing a lot of things to try to be prepared for this. Our, we made a chip that is an incredible performance at like, you know, units of intelligence performer, units of intelligence per watt. We are learning how to build our own data centers.

• 26:12Speaker 01

  We have a robotics program that we're going to point at making lots of data centers. People talk a lot about.

• 26:18John Collison?

  The real RSI.

• 26:20Speaker 01

  Yeah, people talk a lot about RSI as like a AI that does AI research. There's very little, relatively little talk about like self-replicating data centers. You know, this idea that you can use AI to control robots and run fabs and build turbines or solar panels and, you know, build data centers. But I think we are, I think we have that like within grasp. But, and so we'll try to make a lot of really cheap compute.

• 26:44John Collison?

  But aren't data centers already not labor bottlenecked? They're permitting bottlenecked and component bottlenecked.

• 26:52Speaker 01

  Can we have more components for robots?

• 26:54John Collison?

  And.

• 26:55Speaker 01

  Eventually. Yeah, yeah, yeah. And all of this is like, how we solve this over a long period of time.

• 27:00John Collison?

  Like, my observation of the compute stuff is they say the solution to high prices is high prices. That's like, okay, step one, we have the high prices for memory and GPUs and everything. And the solutions should be here any moment now. How are we doing on the solutions front?

• 27:15Speaker 01

  Bad.

• 27:16John Collison?

  And I'm not saying this is OpenAI's job. I feel like this is the invisible hand's job, right? Like, memory's gotten really expensive and so new memory makers are going to flood the market. Any moment now.

• 27:26Speaker 01

  I think the invisible hand is failing right now. I mean, the invisible hand, it has like a checkered track record at best. But I think it's like not quite working here.

• 27:38John Collison?

  Why not? Like, is it just histories or?

• 27:40Speaker 01

  I think it's, I think the like specifics of each area matter here. But in the, since you started with the memory guys, like, you know, for a very long time, if you were a producer of DRAM, you would invest $20 billion of CapEx to make $19 billion. And you would go through these like booms and busts, like maybe some years you would make $21 billion, some years you would make $18 billion. But like, until right now, I don't think, I don't think making memory has ever been a great business over a long period of time. And then all of a sudden, you know, these guys are each going to make like $100 billion of like free cash flow in a year. Unbelievable.

• 28:19Speaker 01

  But, but I think they all like remember their history and how many times they almost went bankrupt by making too many fabs and being promised this time it was different.

• 28:26John Collison?

  Which leads them to be conservative in their CapEx.

• 28:29Speaker 01

  Clearly.

• 28:30John Collison?

  It feels like what you should expect to see is that every part of the supply chain, both new startups coming along, and maybe more relevantly, people with relevant experience getting into. So, you know, Intel makes chips, they can probably make, they can get back into the memory game as is tradition.

• 28:57Speaker 01

  As is tradition.

• 28:57John Collison?

  Switch back.

• 28:59Speaker 01

  Oh, I think there are like all of these ways where the supply chain, it's just like super non-obvious.

• 29:06Speaker 01

  An example of this is people.

• 29:11Speaker 01

  A lot of people want local inference and somehow they assume that like the problem is solved now because we're going to have open source models. We're going to be able to run them on our computers. And so we're not going to need all this stuff in the data center. But from like a global allocation perspective, actually, it's quite terrible to have everybody like have a bunch of the world's limited supply of DRAMs sitting in their own computer and being used 5% of the time. Like you want maximum utilization in the cloud if you're trying to like deal with this very supply constraint environment. And somehow everyone's, many people's intuition is the opposite of that.

• 29:44John Collison?

  Where else have we seen the market respond to high prices, not just in memory, but across the supply chain?

• 29:51Speaker 01

  Turbines? Batteries? Concrete?

• 29:55John Collison?

  Sorry, the market is responding there in in high prices. I'm seeing where, are you seeing any interesting market responses in terms of high prices pulling in innovation?

• 30:04Speaker 01

  Components for, some components for robots. And not like the humanoid, like, you know, kind of like fake demos, but like the industrial robots that can just make stuff faster.

• 30:12John Collison?

  I see, with this new actuator makers and that kind of stuff.

• 30:15Speaker 01

  Yeah, honestly, I mean, maybe if OpenAI sells our chip, maybe that would be an example.

• 30:20John Collison?

  Okay, yeah, I want to get to your chip because that was going to be my example. I think the the solution to high prices and short, you know, limited demand, or sorry, limited supply, excuse me, is you make your own chip. Presumably, you're using a lot of AI to make the chip. I would have thought that's hard because I would have thought that chip design is one of those places limited by the public tokens you can get, where there aren't actually that many blog posts on the public internet on how to build a, you know, leading edge two nanometer chip.

• 30:51Speaker 01

  Well, in terms of like how to do the lithography, yeah, I think there's like one company on earth that seems to know how to do that. In terms of how to do the chip design, like to build the file that you eventually send it tape out, that is well understood and that is a great problem for AI. And given that the world is going to be like limited in its silicon capacity, we all, we have like a shared interest. One of the things that I'm proud of about OpenAI, and maybe it sounds obvious, but it was maybe like not as obvious at the time, is we have a culture where people, even if they've done something the old fashioned way for a long time, are are super willing to like try using AI to do it better.

• 31:35John Collison?

  What are other examples of that?

• 31:37Speaker 01

  AI research, software engineering, but also increasing like procurement or things that people think of like very outside the realm of what most people are using AI to do.

• 31:48John Collison?

  How does AI-driven AI research actually work? Like the math proofs where you say, or the finding the PHP vulnerability where you say, please find me some AI advances?

• 31:58Speaker 01

  So I think we don't know yet how it truly is going to work. We know today how to say like, find me a 5% computer efficiency win.

• 32:05John Collison?

  But it's that high level. Like I could be an incredibly successful AI researcher now.

• 32:11Speaker 01

  I don't know, incredibly, but yeah, if you were looking for a new job, you can come do it. What I don't think anyone knows how to do yet is to say like, go discover a new architecture that is as big of a single E board as the transformer was. And I think that's where things are heading.

• 32:25John Collison?

  And what's missing there?

• 32:26Speaker 01

  Smarter models, smarter models that are smarter in this dimension.

• 32:30John Collison?

  The high level planning and conceptual reasoning.

• 32:32Speaker 01

  I think if you apply the same pipeline that we applied to get our models to be so good at math to AI research, like it will probably work now. Per an early part of this conversation, I kind of think that at least for a while, maybe for a long while, we are going to have to agree to some limits about how smartly that these single models go. But a lack of progress will come from that decision, not from the AI's inability to do good research.

• 32:57John Collison?

  Concretely, if I'm using AI to develop improved AI, because we should expect GPT-7 to be really good now because you have GPT-6 and whatever, you know, unreleased version of it to develop it. And so my expectations are sky high. But I come into work, you know, with my backpack and I come into the OpenAI office in San Francisco and I sit down at my desk, like, am I saying, find me more intelligence in this model? Am I giving a real direction of research? How directed does it get? What does the iteration loop look like? Just what does RSI, because you are, you have a front row seat to RSI. And I'm curious what it looks like.

• 33:42Speaker 01

  One of the sort of interesting things that people talk a lot about OpenAI, which sounds very obvious, that sounds like a good answer, just like, well, build me, build me RSI or whatever, is what are the questions that you're supposed to ask of GPT-7? Probably it's not enough to just say like, make me RSI. And also, like, it's probably not enough to say like, buy me a 5% computer efficiency win. I think this is like one of the most interesting debates going on right now. And the answer is not just like what question to ask it, but it's things like agent swarms. Like what are we, how are we going to use these models in a way that like actually maximizes what we can get out of them? I think we're going to learn a lot about that in the next like three months.

• 34:26John Collison?

  Okay, but it sounds like you guys do not have a codified OpenAI standard way of doing an AI-enabled AI development. It's still lots of experimentation yourself. Yeah. When we're talking about the token budgets, it just strikes me that this is one of the big unresolved questions for AI use in enterprises generally, is how, you know, we don't have hard budgets for individual engineers. You don't get cut off at a certain point of like, sorry, you've spent all your arcade tokens. Maybe we should, but we don't. And again, I think we all believe in the scale up of AI, but if you take that seriously, it means the

• 35:08John Collison?

  dollars of compute or AI per person will go up. And so I feel like we're pretty soon to now actually just running into this question of how you actually decide how much spending power to give individual people inside the organization.

• 35:24Speaker 01

  There was time when a lot of companies had this like a thing in their Google Calendar which showed the like approximate cost of each meeting in terms of like everybody's comp that had to be there. And I think that was always still an underestimate of it. And there was this observation that like Paul Graham made it at one point, you know, and many other committed to, which is that there was like, there are people at companies that have incredible power to save like, you know, $17 on the snack budget of the office every week. And yet you can have these meetings for like unbelievable amounts of money that are adding no value and that there's kind of like no equivalent of the ability to stop that.

• 36:03Speaker 01

  I think that most use, okay, so I'm like really talking my book here and I'm very biased, but I genuinely believe it. That use of AI is like not a company's biggest spend problem relative to other things that happen. And even if it's like somewhat inefficient, it is not where I would focus.

• 36:23John Collison?

  Although, hang on, that's not scaling law pilled enough, like, I'm not, okay, two years from now.

• 36:28Speaker 01

  You want to go send all Stripes revenue, all the Stripes profit on OpenAI services and then complain about this? I'd be delighted to have a conversation.

• 36:33John Collison?

  But we're on a path to that, like, you know.

• 36:35Speaker 01

  Great.

• 36:35John Collison?

  Yeah, we now spend, you know, 1% a shorter magnitude of Stripes OpEx is tokens. That is obviously.

• 36:43Speaker 01

  That feels too low.

• 36:44John Collison?

  Yeah, so we're kind of fixing it and it'll get to 5% and 10% and 20%. And then you have a real question.

• 36:52Speaker 01

  Do you have an intuition on what the right numbers? Like what it should be today?

• 36:56John Collison?

  What it should be today is maybe tricky, but steady state, isn't it going to be, depending on the kind of company, isn't it going to be tens of percentage points of their OpEx?

• 37:16Speaker 01

  I think for many companies, the answer to that is clearly yes. And and I think there will come a time when the right thing to do is to really optimize that. And also, like, another thing I'm proud of about OpenAI is we've tried to really drive the cost per unit of intelligence down and down and down. I think we've been like a good spokesperson and good champion of this. But still, I think the problem, even at current costs, there are more companies, the problem of not using enough AI rather than using too much AI.

• 37:53John Collison?

  Do the dollars per gigawatt of leading edge AI go up or down?

• 37:57Speaker 01

  That should go down.

• 37:58John Collison?

  Down?

• 37:59Speaker 01

  That should go down.

• 38:00John Collison?

  Why? Because it's gone up to this point, right?

• 38:05Speaker 01

  I mean, the reason it could go up is that if AI just gets increases in value per what it can do per watt so quickly and there's such a supply shortage that people are willing to spend much more.

• 38:15John Collison?

  But you would view that as like compute not keeping up?

• 38:17Speaker 01

  But I would view that as compute not keeping up. And if I look at the like fundamental cost of turning a standard to intelligence, that should go down. Energy, the cost of energy should go down. The cost of chips should go down, the rest of the cost of the supply chain should go down. If it doesn't go down, I'll be very sad. But for it to go down, given how much value there is in this, supply's got to go way up.

• 38:38John Collison?

  Okay, so you would like to see a compute glut, where we start producing a lot more memory, a lot more data centers, a lot more of all the inputs. And as a result, anyone who owns a computer for whatever reason can have one.

• 38:50Speaker 01

  People give different numbers for the kind of the cost per gigawatt measured in like the five year total cost of ownership right now, but like, let's say, let's pick $70 billion as a number. I think if that number were.

• 39:02John Collison?

  That's a cost to install.

• 39:04Speaker 01

  And run it for five years.

• 39:05John Collison?

  Yep, okay.

• 39:06Speaker 01

  I think if that number were like 20 instead of 70, we should be much happier. And and we should still demand to get, you know, 10x better performance out of it with smarter models every year.

• 39:17John Collison?

  Presumably, you don't think like 2 is a good number, like something's gone wrong if we're seeing 2 because it would be surprising if we were. Or is 2 better than 20?

• 39:27Speaker 01

  Why is it not? Like, I don't think that will happen. But why do we not want that number to be as low as possible?

• 39:33John Collison?

  I don't know, this is your theory. You have a view that generally deflationary technology is good.

• 39:36Speaker 01

  I do believe in that. Yeah, I'm pro-deflation in a sense, pro-deflationary technology.

• 39:40John Collison?

  So how are we going to get there? What's going to happen between now?

• 39:43Speaker 01

  Well, $2 a gigawatt, I can't imagine.

• 39:45John Collison?

  Okay, what's going to happen between now and $20 a gigawatt?

• 39:47Speaker 01

  $20 billion. $20 a gigawatt would be really great.

• 39:49John Collison?

  Okay, yeah, yeah.

• 39:50Speaker 01

  Even to get down to 20, you're getting into the realm of energy's got to get cheaper. But certainly, like, all of the gross margin stacking that happens between melting sand, or the equivalent of melting sand and getting the token back, a lot of that's got to get compressed.

• 40:05John Collison?

  You could just get there with eliminating profits? Or do you need to also make the energy cheaper? I don't know the math off the top of my head.

• 40:12Speaker 01

  I mean, if you assume like all materials were free and no one made a profit, maybe you could get there in the current energy envelope, but only just.

• 40:17John Collison?

  I see. Personal agents are the most exciting thing happening in AI right now. One of the, there's many exciting things happening in AI, but one of the most exciting things, you guys launched Dots a few weeks back at this stage and a bunch of other people also launching in the field. Was? Will be true of the winning personal agent.

• 40:43Speaker 01

  It's like intelligence, context, tools, it's a limited number of things and that someone is going. I think no one has yet figured out the like, the thing that people like. People like these, but this is not something that people are like, I cannot imagine my life without this thing.

• 41:01Speaker 01

  In a way that when people talk about coding agents or even chat GPT, they do say. And someone, I think someone soon is going to figure out what that looks like. And you talked about one of the reasons why this doesn't quite work yet. But I think this is like a small number of ideas away from a super mainstream thing.

• 41:18John Collison?

  Oh, interesting. You think we have not had the Macintosh moment for personal agents?

• 41:22Speaker 01

  I think we have not.

• 41:24John Collison?

  Do you have any sense of what the shape of it will be?

• 41:26Speaker 01

  I, I think it'll look a lot like the current, you know, dots or news or Grokbot or whatever, but it'll just work. Like it'll just work, it'll be reliable, you'll like never have, you'll just like give it more complex tasks and it'll do all of those things.

• 41:40John Collison?

  Computer use dynamic is very interesting to me because not everyone is embracing the new agentic future with open arms. But it's not totally within their control because you can just have an agent using a computer. Where do you think that dynamic goes, this sort of cat and mouse?

• 42:04Speaker 01

  Like rough instinct is the world will.

• 42:08Speaker 01

  And people are going to fight this in both directions, but eventually converge on something like you're allowed to like tell an AI to use your computer on your behalf, including software.

• 42:18John Collison?

  People always talk about that, but isn't it a bit like the semantic web? It's like a really nice conceptual idea.

• 42:23Speaker 01

  I was gonna say, like, I don't think this is like so black and white. I don't think I have like a super high degree of confidence of how this is gonna play out. As a consumer, I feel like somewhat annoyed if I can't

• 42:33John Collison?

  Yes.

• 42:35Speaker 01

  have an AI automate tasks on my own computer. Like I want it to feel like it's my agent doing what I ask, but then it gets like, I don't, I don't have to like go do something else. Yeah.

• 42:43John Collison?

  Yeah, I note a very interesting dynamic right now, which is

• 42:51John Collison?

  every company is really eager to add AI to their products in a very specific way. Which is, there is a tab or pop-up widgets or something that gives you a messaging type UI. And it's often surprisingly capable. And so you can say, hey, I actually want to return my most recent order. And then you're kind of surprisingly impressed when it's actually is able to find your most recent order. And you know, it's a bit janky and they can do all those things. However, no consumer really wants this UI because consumers have strong,

• 43:36John Collison?

  not immovable, but strong preferences for their existing primary chat app, the kind of top of wallet app that they open when they want to go to do something. And what they want is to be able to say, hey, go find my recent order on this business and refund you or something like that. And I noticed that businesses are really reluctant to allow people to use those businesses inside another AI app. Because like, oh, we're gonna lose control and our motive, whatever. And it just strikes me as fine because your defensibility probably did not come from your ability to put together HTML. Your defensibility came from the brand loyalty and the product selection.

• 44:21John Collison?

  And the new UI is that agentic interaction inside chat GPT or Gemini or whatever it is. But all businesses disagree with me here and they're very good at business, so maybe they're right.

• 44:32Speaker 01

  I mean, again, I, I have like a bias here, but I so directionally agree with you. Like I have such an extreme preference to buy something from Amazon versus some random e-commerce vendor that I don't know if I can return to, or I don't know if I can like trust to deliver it quickly or something like. If I ask chat GPT to buy me something and I can choose Amazon versus some other thing, like 10 out of 10 times I click Amazon, pay a higher price, I have just like such.

• 45:01John Collison?

  Yes.

• 45:01Speaker 01

  And it's nothing to do with like my experience of clicking on amazon.com.

• 45:05John Collison?

  Exactly.

• 45:06Speaker 01

  And now, I think you could say we have not done a good enough job offering partners a platform inside of chat GPT if they don't want this to happen. But, and again, maybe I'm really biased here. I think Amazon's like, you know, great. I think they're in a great place of commerce, but it's like much more agentic.

• 45:23John Collison?

  Yeah, Amazon in particular has this amazing defensibility via the fulfillment network. And so you can use Amazon kind of from anywhere.

• 45:30Speaker 01

  I super trust that.

• 45:31John Collison?

  Yeah. Now, I'm just really struck by this dynamic, like my new thinking, in particular with regards to agentic commerce, but also anything is anytime there is a web form field that you as a human are filling out, these days that is below your pay grade. Because you are probably not inventing that information for the first time. Like we've all filled out the medical forms where it's like, oh, your family history or whatever. And that's so frustrating because how many times have you filled out your dad's and your mom's medical history and things like that? It just, it feels so wasteful, but the million versions of that, you are never inventing new information for a web form.

• 46:07Speaker 01

  I would be so happy to never do like fake work at a computer again.

• 46:12John Collison?

  Yeah. And my argument is that any web form is basically fake.

• 46:15Speaker 01

  I agree, but, but like the, the, I mean, I was like one of the very earliest adopters that I knew of of those like plugins that put in your like name and address like very clumsily. You know, because I was just like, this is, I cannot do this again. But I kind of feel like that using like most of the web at this point.

• 46:33John Collison?

  Yeah, I just find this phenomenon really interesting where businesses right now are really attached to their HTML and mobile apps as sources of business defensibility. And it just feels very, like, either they're wrong and it's not a source of defensibility or it may be, but it's transitory and they just need to get over it.

• 46:54Speaker 01

  The last example of this that I think we really witnessed was like the app stores. The iPhone app store specifically. And I remember every startup was like, or many startups, like most of the YC batch was like, I'm gonna build an iPhone app. This is like new distribution. This is awesome. I can do a new thing I couldn't do before. And many new startups got created and all of the big companies felt like they had to make one, but they kind of didn't want to. Because they're like, oh, and then I'll like hand my, you know, relationship with my consumer over to Apple. And it was great for me.

• 47:23John Collison?

  It was just a new UI.

• 47:24Speaker 01

  Yeah.

• 47:25John Collison?

  Okay, I'm gonna go more rapid fire. People complain about benchmarking. Is that a real concern? Are OpenAI models benchmarked? Do you, are you actually Rick Rubin-ing it where you just feel like the vibe of this model is right and you don't care what Sweetbench gives us?

• 47:39Speaker 01

  Uh, maybe there are like two things we should really care about models being good at. One is like, you know, scientific discovery and doing things that people cannot do. And the other is like being enjoyable to work with. And I don't think we have great benchmarks for being enjoyable to work with. I'm actually not sure we have great benchmarks for like the smartest scientific scientist either. But we endeavor not to benchmark. And I, I, I kind of, I keep thinking that benchmarking is like over as a phenomenon and people keep doing it.

• 48:14John Collison?

  So are you better than other labs at not getting sucked in by the siren call of benchmarking?

• 48:20Speaker 01

  I believe we are better at that.

• 48:22John Collison?

  It feels like.

• 48:23Speaker 01

  I don't think we've always been better than that in history. I believe we're better than that right now.

• 48:27John Collison?

  It feels like the industry is still figuring out the being nice to work with, where there was the glazing era. And more recently, there is the stroppy, ornery era of models that were anti-glazing and as a result tried to disagree with you and nitpick at you on every minor thing. The current ones are pretty good, actually.

• 48:44Speaker 01

  It's gotten, I think the latest models are much better than we've had in a long time.

• 48:49John Collison?

  But it also feels like there's probably higher peaks to ascend.

• 48:52Speaker 01

  Emmett Shear said to me years ago that what he wanted was a really thoughtful and challenging, but still respectful senior coworker. And I have still not heard a better like brief encapsulation than that.

• 49:06John Collison?

  Yes. After working with him for many years, what's your grand unified theory of Greg Brockman?

• 49:10Speaker 01

  I think Greg Brockman is like one of the kind in the best sense.

• 49:13John Collison?

  He's singular on the Greg Brockman scale and you just kind of have to be there to understand what that is.

• 49:16Speaker 01

  But I think it's, it's like a combination of like, yeah, intensity, like monster productivity, extreme dedication and like focus on succeeding at the mission. It's very cool.

• 49:27John Collison?

  You hate tokens as a metric for anything really. Pricing or just kind of, it's not the standard unit of AI and yet we found ourselves there.

• 49:41Speaker 01

  You know, we did.

• 49:43John Collison?

  And you might say outcomes, but I don't think it can be outcomes because outcomes is too different across different use cases. It's what's the right standardized unit?

• 49:51Speaker 01

  We almost made a very, what would have been bad call to say we're not gonna price our next model in tokens at all. And, and you're just gonna like give it a prompt and we're gonna give you a price back. And we're gonna do it. And it's gonna be like AWS, you just do something and we send you a bill. But the, the customer response on that is so negative. They're like, I need to know what I'm paying for somehow.

• 50:15John Collison?

  Yes.

• 50:15Speaker 01

  And I asked a few people, like, would you rather pay more for the same output, but have it priced in some number of consistent tokens? Or something like, even if we like inflated the tokens? And I go, I'd much rather inflate, you inflate the tokens and pay more because I know there's like a real thing there. So I think you are right that people don't just want us to like give them a price because they, they feel like over the long term, they're gonna get screwed somehow.

• 50:41Speaker 01

  In the same way that maybe people would like rather pay for CPU hours rather than just like getting their thing done and even if it was less expensive.

• 50:48John Collison?

  You know, the way like nanometer, the chip gate length, yeah, the nanometers has very little relation to reality these days. Will we get something like that in tokens where you guys should start saying, here is an intelligence equivalent token like, I can't believe it's not butter.

• 51:02Speaker 01

  Of like the smallest feature. I, it's like made up and I hope not.

• 51:08John Collison?

  You think that would be bad?

• 51:08Speaker 01

  It off, I mean, it offends me that we like use the word nanometers for something that has nothing to do with nanometers.

• 51:14John Collison?

  I know, it's very silly.

• 51:16Speaker 01

  So.

• 51:17John Collison?

  You'd not like that?

• 51:17Speaker 01

  I, I would rather not make up a fake unit. Tokens are a kind of fake unit. I mean, they're real in some sense, but like they're not a good unit.

• 51:25John Collison?

  They're fake from a business value point of view.

• 51:28John Collison?

  But isn't the only, okay, isn't the only defense of tokens, that if you're wasting tokens, we should fix that pretty soon. Like we shouldn't be streaming, Claude Shannon would not approve of us just streaming noise rather than signal. And so there will be a, there's a floor of the information theoretic most useful tokens on both sides and we'll kind of come back to that.

• 51:50Speaker 01

  Yeah, I mean, right now you have to think about like whether you should want like the 10 cent per million tokens out of a small model or $10 per token, tokens out of a big model and then whether, like how long the chain of thought is likely to be. And like, that's just like cognitive overhead that I don't think actually adds any value.

• 52:10John Collison?

  Is the chain of thought included in the price?

• 52:12Speaker 01

  In what we charge, most people do.

• 52:13John Collison?

  Okay.

• 52:15Speaker 01

  By the way, the Guinness was not as bad. It's definitely not as good as an IPA, but not as bad.

• 52:19John Collison?

  It's good.

• 52:20Speaker 01

  Yeah. I mean, an IPA is great, but it's, it's pretty good.

• 52:23John Collison?

  We like Guinness.

• 52:24Speaker 01

  Yeah.

• 52:25John Collison?

  At least in this podcast. Thank you.