Sam Altman in conversation with Patrick Collison

All right, I think we have some Codex fans in the audience.

Love to hear that.

Um, how's the week going?

So fun. Um, it's a busy week, but I'm happy to be here. This is a unexpected surprise.

Well, thank you for joining us. We

appreciate it. Um so we opened uh this morning by saying that it is uh we've kind of arbitrarily decided uh that the singularity started on January 1st and

thus today is day 119. Uh what do you think of that?

It does feel like we are somehow in the takeoff. Day 119 feels like a reasonable

takeoff. Day 119 feels like a reasonable enough guess. Uh yeah, I won't fight it.

enough guess. Uh yeah, I won't fight it.

Did you did you feel like did did you feel this? So we've started to see a

feel this? So we've started to see a bunch of our metrics inflect as of um I mean late last year beginning of this year I mean things were kind of doing

well but somehow they really just the the shapes of the curves changed. They

really went parabolic. Is that matched in what you guys see? Was there some trajectory change? Like why are we

trajectory change? Like why are we seeing this?

I do think the models got really good especially for coding but really good in general starting

late last year very early this year and at least in my own experience of using this technology and seeing what other people are doing with it. It and also

this sense that every week is now a little bit different than the week before like a lot happens very fast. um

it seemed to all correlate with the models hitting some threshold.

And why did they like why did coding models suddenly start to click over the last you know couple months? Like was

there a was there a research trick? Was

it just got enough code data in the pre-tri like what why why did why did it suddenly start to work?

Yeah, it's a great question. We we

wondered a lot about like why kind of several people crossed that threshold at the same time. Um

I I I'm sure it's a number of factors, but model intelligence, just the raw kind of reasoning horsepower, um enough of a feedback loop of people using it for code to figure out where it was good

and where you needed to improve it. Uh

enough data. I think it was like all of these things. And then also there was

these things. And then also there was like a like many other endeavors, once you know something's possible, it's much easier to go do it with vigor.

And so it seems like codeex is kind of having a moment right now.

Yeah, it really it crossed some subjective threshold for me with the latest app updates and 5.5. And this is also like it's quite hard to say why

right now and not a little bit sooner.

Why not the next model? But one of the things I have learned about the history of all of the things we put out is it is very hard to say why this

particular thing was the thing that worked. And this goes back to chat GPT.

worked. And this goes back to chat GPT.

Like why was GPT 3.5 the thing that got over the threshold where most people went from saying not that impressive to like going to change the world and why not one model earlier or later? I really

can't explain it. um you just kind of feel it. And the I've had two inflection

feel it. And the I've had two inflection points with codeex. One was kind of with GPT 5.2 and then a really big one in the last few weeks where it's like, okay, this is

going to be the primary interface to a computer for me.

And is everyone using it for coding or are you starting to see usage diffuse into other domains?

I think the most adamant users are still using it for coding. Um, but there there's been just this like tidal wave of people coming into Codeex recently and one of and I'm like really trying to

understand like what has happened that this is causing this and the the depth of what people are using it for or starting to use it for has

surprised me. So certainly our ambition

surprised me. So certainly our ambition is for it not just to be about coding but to be about all the work you do in front of a computer. Uh, and

you know, I would say we're maybe like 10% of the way there for the non-coding stuff, but now that we see what's happening, now we have a real user base sort of using it in in these other ways, I think we'll get good at it very fast.

What what do you think will be the next domain that subjectively feels like it has this big unlock after coding? Like

is it going to be spreadsheets? Will it

be uh, you know, performance reviews?

What's what's it going to be? I so I think there will be a lot of first of all I do think coding is a little bit special and these models are a great fit for coding. The world needs so much more

for coding. The world needs so much more code than currently gets written. Um

there may be no other domain that is quite like coding. Uh but there will be a lot of others that I think are close.

But the the next kind of coding like thing that I think will happen is not any specific domain but the realization of how much time people waste trying to

use a computer and the idea that you can do a huge percentage of your day in a very different way. And you know, maybe maybe you don't realize like how much

time you spend clicking between messaging apps and copying and pasting stuff and like, you know, responding to very boring things that you could clearly automate once. Um, but the degree to which most people will realize

they can like sit back and watch an AI do most of their like drudgery um is going to surprise people. And

in my own experience trying to work that way actually gives me much more like enjoyment of work. I didn't realize how much like the little stuff drags me down, gets me like out of a sort of happy flow state or whatever. Um, so the

subjective like quality of life improvement is huge. Are you an openclaw user?

I am. Um, I open users here.

We have some uh good news for you all coming. Um I

coming. Um I you can just tell them um OpenClaw has been one of my biggest

like this is magic AGI moments ever in the field. Uh I I remember the first

the field. Uh I I remember the first time someone told me about it. They were

trying to explain it and I was like okay like those all that all sounds cool but you know I can like make a lot of that work. And then it was a real reminder

work. And then it was a real reminder how when when the models cross some threshold and also the product designer gets a handful of critical ideas really right, it's like a much more magical

experience than it than it sounds like.

I find um I I've been a um an attempting openclaw evangelist uh and you know trying to describe it um that experience

you just recounted uh to others and I find it a difficult experience to communicate. I mean it sounds kind of

communicate. I mean it sounds kind of prosaic, right? It's a it's a a stateful

prosaic, right? It's a it's a a stateful chat GPT session that can also make some use of tools and so forth. Um how like what do you use your open cloud for?

If we scrolled your message thread, what how do we see? So this is like a very embarrassing thing to admit. The thing I always try first, perfect place.

Um, the thing I always try first with a new like a new kind of AI system of any sort is to I'm like a home automation nerd and so to like try to build a better home automation interface system

because it's all it never works like it never is good.

Um, and OpenClaw was the first time that I was able to get like a setup that I was happy with. Um,

I also built a messaging app that I had always wanted to work. I've since

switched it to something I built with codecs, but OpenCloud was the first time I was able to like I'm sure like you it like feel like drowning in messages and it's like this very unpleasant task to wake up in the morning have to go through all this stuff. So I was like

all right I'm finally going to be able to automate this and that was like again should have been doable with previous systems. Hard to explain like what it's like when it all actually just

works and you trust that it's going to work. Um, I was uh testing the uh the

work. Um, I was uh testing the uh the new link uh uh CLI that we uh that we just launched today uh in preparation for launch and I asked my agent to uh

you know it it means you can uh you can easily get a singleuse card they could use on any business. Um and so uh I asked my agent to go and buy itself a

gift just anything on the internet uh for under $20. Um, and it uh it it chose to buy itself uh an HTTP design from uh from Gumroad.

Wow.

Yeah. There's all of the there's all this stuff that like feels no matter intellect no matter how much how convinced you are intellectually that this is like not a real thing wanting a real gift for itself. And no matter how

much you're convinced like okay this is like a weird emergent behavior and I'm not supposed to read into this. Um there

are these things that uh feel a little strange. I

um we're going to have a party for GPT 5.5 and I wasn't quite sure like we're going to you know invite people who are big users and whatever and I wasn't quite sure what to do. And so on like kind of whim I was like um I'm going to

ask 5.5 like xxx high what it would like for a party for itself this morning. Um and

this morning. Um and I did and it was this sort of like beautiful set of things including like

you know I like here's what I would want for like the flow of the party. Here's

what I would not want. You know you should do it on May 5th. That would be funny. Um I would like you know like

funny. Um I would like you know like only a short little toast and I want it not to be by me but by the people that built it. I would like like a you know

built it. I would like like a you know big central suggestion for 5.6 six and I would like you to like feed them all into me and I'll like you know make sure we work on that and now there's real moral pressure on you to

well we're going to do it um but it was a strange thing.

So I want to ask you about open eye itself. Um you know a lot of crazy open

itself. Um you know a lot of crazy open I mean it's it's now I am aware it's an 11year-old organization now somehow feels like so much longer than

that but yes I I get just over 10 I guess. Yeah. Yeah. Okay. 10 10 years.

guess. Yeah. Yeah. Okay. 10 10 years.

Um, a long 10 years. Um, I can't like remember pre-open AI life that well at this point. It feels like it's been so

this point. It feels like it's been so long. But yes, you you have a

long. But yes, you you have a singularity looking forwards, but also a singularity looking backwards. Um, so

what's um what's the craziest OpenAI story that's never been told?

I mean it sound it sounds so prosaic relative to like the crazy drama that's happened but

there was this period after we had finished training oh there I guess there's two that are kind of similar but the one that just came to mind was after we had finished training

GPT4 there about eight months before we released And so there was this 8-month period inside of OpenAI where we were all using

this thing. We kind of knew that it was

this thing. We kind of knew that it was dramatically better and different and going to unlock a bunch of things in the world. Um,

world. Um, and no one outside the company or almost no one outside the company knew about it. And and there was this we'd like

it. And and there was this we'd like walk the hall sometimes. be like, you know, are we engaging in like collective psychosis? Like are like do we just have

psychosis? Like are like do we just have we like gotten like, you know, totally whipped each other into this frenzy? And

and there was there wasn't there was like no feedback to keep us in check or sane from the outside world.

And it doesn't sound that weird relative to like, you know, crazy board drama or

Elon trial or something like that, but living through it was an unbelievably strange time.

Um, what's the um what's the Sam Almanagement style? If I'm if I'm working for you

style? If I'm if I'm working for you directly or maybe indirectly I'm you know leading some product or or something just what does that look like?

Um I'm definitely not a hands-on manager. I

I'm very much the of the of this style that you get great people um kind of give them like a very

high level thing to point at. um and try to let stuff just happen. Um

I I think there have been kind of two main phases of open eye and we're heading into a third and the first one was like we were we were only a research company. We were trying

to figure out how we were going to build AGI at a time when it sounded completely completely crazy. uh and we really had

completely crazy. uh and we really had no idea what to do. And then there was a second phase where in addition to continuing to do that, we had to figure out how to build a product company. Um

now we have to in addition to both of those two things, figure out how to build this like mega mega scale token factory for the world. Like like I I

think of what we're doing is building sort of a new utility. Um and people are going to want to use a lot of tokens, a lot of intelligence in all sorts of

ways. We need to make that as smart, as

ways. We need to make that as smart, as cheap, as abundant, as easy to use as possible. Um, and that will require, I

possible. Um, and that will require, I think, pretty deep full stack integration and a massive, massive infrastructure buildout. Um,

infrastructure buildout. Um, the thing that I didn't really appreciate between the phase one phase 2 shift was how much my management style had to change. You know, running a

research lab and running a like a product company are two extremely different things. And I suspect this

different things. And I suspect this third phase is going to be very different yet again. And so I've been reflecting on, you know, if we're going to really go do this, how I'll have to

change. Um,

change. Um, and I think it's not going to be like a natural fit for my management style. Uh,

so I either have to find someone a few people great to hire or I have to figure out how to do things in a different way or uh I have to like build an AI that can manage this new thing. When I

interviewed Jensen here uh two years ago, he told me and you know his several thousand closest friends uh about his 60 direct reports. Do you have any super

direct reports. Do you have any super weird not that that I shouldn't call his practices weird, but uh do you have any unusual practices like that? I think the closest thing that I have to anything

like that is I I probably talk to like via Slack or whatever or text like a few hundred people at the company a day.

um very quick like you know one two messages whatever not done by an agent like I actually do it and the the context I get from that

sometimes is like very helpful in these diffuse ways I find there's an interesting watershed uh of pre-Slack organizations post Slack organizations

and they're truly quite different totally um I like many other people I hate Slack but I can't imagine having to like still communicate via

email or whatever we used to.

That's roughly where Stripe is.

Yeah.

Yeah. Um

so, um okay, I want to talk about um I mean you kind of just elliptically referenced it. Um there's uh a view that

referenced it. Um there's uh a view that the AI labs are going to uh progress up the stack gobbling up the value chain

voraciously. all these things uh that

voraciously. all these things uh that are certainly within the software sector but perhaps even other sectors and that there'll be this incredible positive

feedback loop and runaway and um kind of hegemonic force that we should all be getting very concerned about. Um

what's your view? I think some of them do want that. Um we don't. I one of the things I've always admired about Stripe is it is like very clear that Stripe is

aligned with its customers. you know, we make more revenue. We charge our customers more. Um, thank you, by the

customers more. Um, thank you, by the way. The partnership with Stripe when

way. The partnership with Stripe when CHBT launched was extremely critical.

And I don't think anyone else could have scaled that quickly, but we scaled. We

pay you more money. It's like very aligned and we're all happy. And you

just provide like a layer of infrastructure for the internet.

Internet gets bigger, you're happy, your user happy. It's clear what the

user happy. It's clear what the alignment is. Um,

alignment is. Um, I don't know exactly how to do this yet, but I would like to get to a model for OpenAI that is similar. Like, I would like us to be an infrastructure

provider. I'd be happy for us to be like

provider. I'd be happy for us to be like a forever low margin as long as we can be huge and growing fast business. And I

would like us to supply kind of an intelligence meter. I don't know what quite to call it that companies can buy that they can you know use to automate

things accelerate things inside their company they can use to build products people can buy it people can take it with them and we find ways to really align ourselves with the success of all

the entire gigantic distributed um economic engine of the world I believe that will work um I believe that switching costs of AI are it's going to

be hard to have like huge margins and AI. Anyway, it's like you've seen

AI. Anyway, it's like you've seen recently how easy it is, many people have seen to switch from our competitor's coding product to ours. Um,

this is actually a consequence of AI getting smarter. It gets easier to do

getting smarter. It gets easier to do things like this. It gets easier to just say like, "Hey, agent, go do this thing for me." Um, but if we can provide a

for me." Um, but if we can provide a utility and people build on top of that utility uh, and you know, we think of ourselves as that kind of a company. I

think that can be quite powerful and very aligned. Well, we're we're happy to

very aligned. Well, we're we're happy to share lots of tips and tricks for doing a a low margin business. Um

but um so many people um um have been um either implicitly critical or in

certain cases explicitly critical uh of OpenAI for um for procuring so much compute. Uh and I think not the codeex

compute. Uh and I think not the codeex users, right? Exactly. So um you I think were

right? Exactly. So um you I think were quite noteworthy for as early as I exactly but you know on the order of two or three years ago um stipulating what

at the time sounded like preposterous figures uh with respect to the magnitude of the buildout that would be required.

And obviously the preposterousness of those figures now looks um less uh less uh tenuous by uh by the day. Um

thoughts on compute capex the buildout.

Um yeah it's going to take a lot of money.

Uh it I think this will be clearly at this point the most expensive infrastructure project that the world has ever undertaken.

um the revenue is ramping to meet it, so people feel better about that. Also, the

efficiency gains that we've all been finding are incredible. So, we're going to get way more out of each GPU than I thought we were going to. Uh but as has

often been remarked, like the demand goes up more than linearly as you drop the price of each kind of unit of intelligence, particularly if you can drop the price and the sort of speed

with which you get it back. Uh so,

this this question now of like what is enough I don't have a good answer to I don't in some sense I think demand for

intelligence at a low enough price um is effectively uncapped now I was going to say we're not but maybe we are like you know we're not going to build a Dyson sphere and then just like

cover it with data centers but maybe we do space data entries good luck with that.

Um, I don't even think he's that serious about it.

Um, I don't think that we're in a capex bubble. I'm not an expert in this. This

bubble. I'm not an expert in this. This

is not, you know, Stripe's business, but just the figures I see relative to the magnitude of demand. It looks reasonable to me.

If we were in a capex bubble in the future, how would we tell?

People love to proclaim bubbles.

And I I can't articulate why, but intellectually I kind of get it. Like

it's it does feel fun and it feels smart. Journalists in particular love to

smart. Journalists in particular love to um you know talk about bubbles. So there

there's like ample desire to write about this when anything looks a little bit silly. And clearly sometimes it's right

silly. And clearly sometimes it's right like there there clearly are bubbles.

But how you can discern between the amount of times someone calls a bubble and the amount of time you're actually in a bubble I have never figured out how to do. In my previous career I was an

do. In my previous career I was an investor. So, I was like quite

investor. So, I was like quite interested in trying to see if I could come up with some sort of framework for this and figure out when you're supposed to deploy capital or not. And I never

was able to figure it out. Um the like I went back and I read what smart people had read had said at different points in history and I was like, "Oh, they called it exactly right." But then I read a little more and they said it like 10

more times than the 10 previous years. I

don't know. I don't have an answer.

Economists are the people who have called eight of the last three recessions But they're so happy when they're right.

So um you know a lot of I mean your business OpenAI um depends in a very significant

way on super talented people. Um and the difference as I understand it between uh the uh the um you know 20th most talented person versus the fifth most

talented person versus the most talented person might be quite large andite quite consequential. And then these super

consequential. And then these super talented or effective people you know in they're not in every case super easy to work with. No.

work with. No.

Um and you look and some of them are wonderful people and some of them are the most fantastic collaborators and some of them are very iconoclastic and strong willed and they get easily or you

know whatever just like the full spectrum of of the human condition. Um

but I guess I'm curious in a domain that's so sensitive to this efficacy and skill and talent and so forth kind of intersected with all the foibless of humans as they exist. How do you think

about this? Like do you guys tolerate

about this? Like do you guys tolerate primadanas? Do you tolerate them more

primadanas? Do you tolerate them more than you used to, less than you used to?

Do you try to manage them in a special way? How do you think about managing

way? How do you think about managing elite skill here?

Someone was working on this book, OpenAI, and said to me like, I I think I think I figured out the thing that you you sort of were really great at and kind of did uniquely well in making

OpenAI happen. And I was like, I would

OpenAI happen. And I was like, I would love to hear. I have no idea what the next sentence is going to be like. I

could not predict the next sentence. Um,

and they said, you know, you figured out how to get a lot of people who all thought they were the only capable or most capable

person uh, and everything had to go their way to work together long enough to like figure out the breakthroughs and that, you know, that was the that was the magic of OpenAI and Okay. So, so

what's the trick? Uh, a a lot of pain.

Um, I I think we we had very even when people didn't like each other and even when people thought they were much smarter than other people

or had a better approach than other people, um, we had a few deeply shared convictions. uh like we we did kind of

convictions. uh like we we did kind of collectively believe in scale and concentrating resources and that we were

going to do this one thing and that we thought getting this right was important enough that people were going to like put aside various personal conflicts.

Um, one of the most unusual things about open is at the time we trained GPT3, the vast majority of our compute at

at the whole organization was going into this one single research program and we would talk to, you know, people that we were like trying to recruit from Deep Mind at the time and they would say,

"That's insane. It's going to create

"That's insane. It's going to create this terrible culture.

We love music.

Um anyway, they would say like we're we're talking about managing unusual personalities.

Yeah.

Um they they would say you have to like divide your compute equally otherwise you'll have this like very toxic competitive culture and you know this thing and that thing and and we would

just take the approach that like we're going to bet with conviction on this.

It's not going to feel totally equal but this is the right thing and we do think we know the thing. We do think we know the direction. we we we really want to

the direction. we we we really want to go in and they would say well you know you might be wrong like we have to do those other things you have to have this research program and

and having a culture where we said we're going to we're going to have conviction and do this uh and ignore the distractions um was great.

We hope this is a memorable sessions for all of you. Um,

so John and I have been doing this thing at Stripe for, you know, quite a while now. We started out in 2010. Uh, you and

now. We started out in 2010. Uh, you and Greg started out in 2015. And to your point, you've um ventured through many trials and tribulations and

stratospheric successes and you know all the rest. Um that was a nice little

the rest. Um that was a nice little gloss over but go ahead. Um

thoughts on I mean it's it's not easy to work successfully with a co-founder for now more than a decade uh and for things even after a decade uh to seemingly work

as well as they did uh you know from the from the beginning. just thoughts on thoughts on that partnership. Why has it worked? Um, and how have you guys made

worked? Um, and how have you guys made it such a success?

Obviously, you and John knew each other for longer than Greg and I did. Um, but

Greg and I did know each other for a long time before OpenAI and I think having the shared history really helps.

One of the things that I had observed at Y cominator was that one of the biggest predictors of success was had the co-founders known each other for a

long time or at least relative to their lives for a long time and the teams that came together like you know 7 days before applying to YC on a co-founder matching site or whatever that was that

didn't work too often. It's not

impossible. I think there are one or two cases where it did work but it was rare.

Uh so we had known each other for a while and we had kind of had a sense of sort of shared values and history and

ecosystem and kind of we were clear on what we wanted to do. Um and

and I think we had this like deep mutual respect and complimentary skill set that has just worked really well. Um

the I'm extremely grateful. I think having to go through a any startup experience but particularly like a intense one without a co-founder you have like a

deep connection trust to is is really hard. I've watched people do it but it's

hard. I've watched people do it but it's very hard. So um I am you know extremely

very hard. So um I am you know extremely grateful that we've gotten to do this together.

on a adjacent topic um you know we talking about open AI but then there's this entire ecosystem um of companies and startups and uh and enterprises that

are you know building on the platform.

It's obviously an interesting moment in startups um given on the one hand the ability to build products and generate

revenue at seemingly unprecedented rates and certainly we see this in the stripe data like the number of businesses reaching thresholds um you know meaningful thresholds is far faster than

it ever has been before. Um you are one of the most prolific and successful startup investors ever. You of course ran Y Combinator. Have the traits that

make founders successful changed in this era or is it sort of the same thing it's always been?

Uh there was a time when we used to make fun of the idea guy. There were these people that wanted to start a company and they'd say like I have the best

idea. I'm not going to tell you what it

idea. I'm not going to tell you what it is. I have the best idea. I just need a

is. I have the best idea. I just need a coder to like build it for me and then I'm going to be in great shape. And we

would like make fun of these people.

They weren't that successful. Um the

and it was like kind of always personally annoying to me because it's like it would be like saying like I have a great idea for a song and I just need that guy with the

guitar to like make it for me. Um

and and so it kind of it didn't it didn't work like you know NYC had a version of this which is teams without nontechnical founders are difficult to get work all

of a sudden it's like the revenge of the idea guys which is actually awesome for the world like I'm happy I'm here for it for sure um but

for a long time I think the most important ingredient that I looked for YC looked for that kind of this part of our industry looked for on

a founding team with technical talent and that's still very important but now people who just really deeply understand their users and can't code at all I like

want to fund those people and that's a that's a big turnaround.

How does one think about startup investing these days? Because on the one hand you have a couple of years potentially to AGI or ASI or the

Singularity, who knows what and then you have investing time horizons or funds with 10ear time horizons.

How does that all fit together? Does it?

Uh I think to do anything at this point on a 10ear time horizon requires a real suspension of disbelief. Uh, and yet that's probably the right way to live

your life. Like I don't think it works

your life. Like I don't think it works to say there's this singularity in three years or five years, whatever. We can't

see past it and so we're going to do nothing or we're just going to like give up or we're going to go crazy or whatever. You like you have to live as

whatever. You like you have to live as if stuff's just going to keep going in an understandable way for a long time.

How how far ahead does OpenAI plan?

Uh I mean we signed like 20-year power and land agreements and for the product um

I think we have like a clear vision of what things can look like in two years and then it gets much hazier after that.

So there was in the relatively recent past uh a narrative that um you know GPT rappers and companies of that ilk were

um were I guess as the peorative suggests undifferentiated and flimsy and be swept away by a rising tide of model improvements. Whereas now it feels like

improvements. Whereas now it feels like that narrative in some sense is flipping somewhat for now instead of talking about rappers we talk about harnesses and harnesses are seen as having this

you know significant heft and uh and importance and I guess I'm curious for your view on this and how you view uh businesses for which AI is a critical

enabling component and their prospect of durability.

I I have kind of had uh the same view all the way through which is you as a business want to be on the side of hoping that AI gets smarter.

Um and whether so so like in the early model days if you were the GPT rapper um and you were like patching some kind of weakness in the current model that was clearly going to get better with the next model. The

next model was much better you were kind of sad. If you were doing something that

of sad. If you were doing something that got better like you know you're making any of the wonderful like services that people were building with the models that benefited from intelligence you you

would be happier. I think the same thing in the world of harness is like I I kind of think the right way to think about this is like data center model harness like that whole thing is

just this one cluster out of which comes this like very usable intelligence. Um

but there are so many things to go build where you're just like happier for that whole cluster to get better and better and better. Um, and then if you're kind

and better. Um, and then if you're kind of secretly hoping it doesn't because you're patching some weakness in that, probably like the next model crank turn somewhere in that stack is just going to

solve it. When you look at the

solve it. When you look at the organizations that are making the most effective use of AI today, like if you um I mean you meet OpenAI customers constantly, large and small, if you

think about the top three that have impressed you the most or the one that's impressed you the most, what specifically are they doing that's different? Like everyone here knows yes

different? Like everyone here knows yes AI big deal we should make enthusiastic use of it etc. But what specifically differentiates those which in your opinion are employing it most

effectively?

Uh a a few different directions there.

um friend of both of ours, uh Toby Lucky of Shopify was the first CEO I knew that just said like we are going to be all in on AI and the way we run our company and

he got himself got his hands dirty just building like AI automation of everything and made his team do it.

Um and you know said we're just going to figure out how we take all these things that are bad and make them good with AI.

And it was not like you know a token leaderboard. It was not some other kind

leaderboard. It was not some other kind of like gamified hackable thing. It was just like the CEO

hackable thing. It was just like the CEO of the company said, "We are now going to put AI into everything we do and I'm going to like not be happy with you, I guess, or we're not, you know, we're not

going to allow it if you're not doing that." So, that energy has now been done

that." So, that energy has now been done by other people. But when um when the CEO of a company just says like we're going to automate ourselves, accelerate ourselves, however they phrase it,

internally um and then really holds people to it and ideally does it themselves. Um

themselves. Um that has worked very well. Uh, I think we're going to try this new experiment where we start

sending like an FTE to work with like hands-on just whatever the CEO of a company needs, work with them to automate their job. Um, as much of it as they can and that I think will have like

if you just do it for the leader of a company, there's like a nice fractal effect throughout the company. So that

works and we'll try to help companies do that. Um,

that. Um, a second thing is being like uncomfortably permissive with data access. Um, there's huge reasons not to

access. Um, there's huge reasons not to do this and I'm not this is like I'm stopping short of recommendation. You

just ask what I've seen from the most effective companies. um this is easier

effective companies. um this is easier for small startups than companies that have a lot of sensitive data and a lot of um you know like process and compliance in

place but saying like you know what we are going to record our meetings we are going to like let this AI have access to our codebase we are going to let this have access to every Slack message every email everything and every employee at

the company is going to get to use it that way like it is amazing watching these two or three startups and AI doing everything work.

Um, and I don't know how the world is going to decide the trade-offs on data privacy versus AI efficiency.

And I think there's like some regulation that's going to have to change for that.

Uh, but it's so powerful.

Um, Tempo is a new blockchain uh that uh Stripe incubated with Paradigm uh which obviously we're partners uh with uh with

you guys on um and uh it the mainet just launched but the project launched uh back uh last uh last summer. So it's a

relatively new and uh and small team a couple dozen people. The Tempo team um

set up a um a harness um a tool in their uh in their Slack installation for orchestrating pretty much everything at the company. Everything. You can just

the company. Everything. You can just ask any task. Go and read these Google Docs. Uh turn those into a bunch of um

Docs. Uh turn those into a bunch of um uh a bunch of linear tasks. Then go

write a pull request to you know implement them. uh then go deploy them

implement them. uh then go deploy them and use our log analysis tool to test the deployment actually worked and the agent will happily go and employ tool use across all of this and it's extremely trippy watching a whole organ

I mean a small organization but an organization of people do everything in a single slack channel and I don't think that would scale to stripe it was the first time I had the experience you're describing which is it's clearly

incredible for them I don't quite see how to transpose it for us but this is really something it's really something to watch um the and I find that a lot of people just

can't this is where there's a big overhang. They have not yet been able to

overhang. They have not yet been able to wrap their heads around the fact that you can just kind of ask us anything and it'll probably happen. Um I I myself still find myself like not trusting

quite enough that it's going to be possible. I don't know exactly how it's

possible. I don't know exactly how it's going to transpose to bigger companies.

It does feel like we're missing kind of like one more abstraction there. Um like

how humans and AIS are going to interface at massive scale. The

advantage that these smaller companies have is like it's just the AIS. They

don't have to figure out the interface with all the all the people. Um, but

we'll figure it out.

Open source AI, where's it going? Does

it have a future for sure? Uh,

right now people clearly want smarter, faster, cheaper frontier intelligence and most of the demand is there. But

there is also a lot of demand for open source and I expect that to increase relatively over time.

So, I want to talk a little a little bit about science. Um, because I know it's

about science. Um, because I know it's something you're very excited about. Um,

but it's also uh relevant to uh to something I spend some of my time on uh which is uh which is the Arc Institute.

Um and OpenAI is in fact a foundation um a nonprofit and recently um recently made

a grant uh to the Ark Institute. Uh and

so maybe we'll get to the details of that uh in a second, but first you just want to speak a little bit to AI as applied to science, you know, what you're seeing and just how you think

about grant making generally uh in the context of the uh open eye foundation.

So generally on on the AI and science question, I I hope that this will be the most important contribution um of

AI of this technology to human quality of life over time and that if we can start to discover new science at a much faster rate which can be like new materials or cures to diseases or any

number of other things like I believe that to a first order approximation life gets better because we understand science better stuff with it and distribute it to people.

But really now with 5.5, the models have gotten smart enough that excellent scientists are saying, I am able to figure out better ideas. The models are

able to make some small but important discoveries. Um, and the pace of science

discoveries. Um, and the pace of science is going to increase. Eventually we'll

have you know automated labs and robots and you know building who knows what. Uh

and we'll be able to do science much faster. But if we can start doing like a

faster. But if we can start doing like a decade of science of what it would have taken us in the old world in a year the compounding effect there and what we'll be able to do and discover will will

just be extremely great. So uh I think this is going to be incredible and this will be one of the big areas of focus of the of the open AI foundation is uh

basically like money and expertise and technology to accelerate science and trusting that that will you know flow to the world in wonderful ways and this is going to be a big foundation.

Yeah, this will I think it is it's one of the biggest maybe it's the biggest I think it will be the biggest foundation um in the world. Uh

and yeah, so we're really focused on science and then AI resilience um like helping the world through this transition uh with this new technology in it. Um

in it. Um but you know, we were thrilled to get to support ARC. I think it's clearly the

support ARC. I think it's clearly the best sort of AI and bio effort and if we can make even a small contribution to

with this technology and with the capital and the foundation to helping uh make people healthier, treat diseases, this whole cluster of what we can do as

we get better at understanding biology.

Um we will be very thrilled. I thought

that was going to take longer and looking at the incredible work the Arc Foundation is doing. Uh, and now I think it maybe won't be won't be that far off. So, you know, in a podcast um

far off. So, you know, in a podcast um midway through you might hear a little interstitial ad. Um, this is your

interstitial ad. Um, this is your interstitial ad for the Arc Institute.

Um, there is um there's an Arc Institute booth downstairs. Uh, and you might

booth downstairs. Uh, and you might wonder why there's uh what what it's doing at the um at the Internet Economy Conference. Um and uh the uh the context

Conference. Um and uh the uh the context here is so the the ARC Institute is an organization we started four years ago.

Um and its goal is to uh produce hopefully the first cure for a complex disease in humans. So a complex disease is uh is one that involves some genetic factors and some environmental factors.

So you can think of most cancers, most autoimmune disease, most neurodeenerative disease uh for example as being a complex diseases in this kind of specific sense. And humanity has

never cured a complex disease. Not one.

We've cured lots of infectious diseases.

We know how to screen for monogenic diseases for this one genetic mutation that um that undergurs it. We've never

cured a complex condition. Um so we started our institute with this is the goal. Alzheimer's uh is the first

goal. Alzheimer's uh is the first complex disease uh that we're uh that we're targeting. Um and our hope is that

we're targeting. Um and our hope is that with both new genome engineering technologies like crisper and then of the amazing advances in AI uh that uh that we'll be able to make some

hopefully meaningful progress and it's only four years old uh but the early results uh are very encouraging. The ARC

Institute is currently looking for a CTO.

Um uh we had uh one CTO uh do a sbatical at ARC uh last year. uh his was it last year before? It was last year. Um and

year before? It was last year. Um and

his name was Greg Brockman. Um and he did some great stuff. Uh he helped us train Evo 2, which is the largest biology foundation model uh ever

trained. Um but we're looking for a

trained. Um but we're looking for a full-time CTO. Um uh and uh the um uh we

full-time CTO. Um uh and uh the um uh we thought that well perhaps that person might be in this audience and if not their friend might be in this audience.

So, if you know somebody for whom that sounds of interest, go check out the Ark stand downstairs and that's your interstitial ad.

I think it's so much better that you labeled that as the interstitial ad rather than just doing it. That was

good.

Um, okay. Well, we haven't talked about

okay. Well, we haven't talked about Stripe. So um

Stripe. So um you were the um you were the I think it was second investor in Stripe.

Was YC the first?

You and Paul at the same time in the same kitchen in fact. Uh so maybe you were first in fact. I don't remember in which order the checks were handed over.

Oh yes, this was then his like PaloAlto kitchen.

Yeah, that's right. Um so you know we were sort of you know we were two pimply teenagers uh proposing building this financial services institution. It

sounded like a bit of a a ludicrous proposition. Why? Why did you invest?

proposition. Why? Why did you invest?

Um, honestly, I I he didn't team me up for this. Uh,

for this. Uh, you and John were two of the most I I'd seen a lot of founders and you and John were two of the most impressive founders of any age, but certainly of pimply

teenager age um that I had ever met. I

clearly was right about that. Um, and

I had like known of you and I had heard Paul talk about you and I had like known of you on the internet. Um,

and I was struck that you were solving a problem for yourself. Um, and that it was sort of like it fit a trend that I

thought was going to be big in the world. um like I kind of really believed that commerce was going to move online in a huge way and there were going to be lots of startups and both of those suggested that this would you know

could be really big but um I don't know I was like a believer that if you can find really smart and still am if you can find like really smart founders and a market that's going to be big you

should just invest and that's kind of it and based on your um well just based on your general perspective in the world um but then also OpenAI's use of Stripe and

what you've seen from that like what you see OpenAI needing and needing in the future and building the business and so forth. What's your advice for Stripe on

forth. What's your advice for Stripe on navigating this um the AI era before that? Just to check my memory, the thing

that? Just to check my memory, the thing you were building the re the reason you realized you needed payments was you had like built this iPhone app to download all of Wikipedia cuz you were going somewhere crazy offline and it was like

hard for you to take payments for it and that was like Yeah, you were like hard. Yeah. Okay. I

hadn't thought about that in a long time. Um

time. Um I I think my my advice more towards like Stripe as a company itself.

Yeah. Going forward

um like it's it's a crazy time in the world. Uh there's a lot changing a lot a

world. Uh there's a lot changing a lot a lot happening. Um and again you've seen

lot happening. Um and again you've seen Stripe from the perspective of a customer. So what um what what are your

customer. So what um what what are your complaints and feature requests?

Well, it's more like I want to I now I have now that I'm thinking that we should be thinking more like a stripe style model. I have a bunch of question.

style model. I have a bunch of question.

I probably have more to learn from you than you you do from me here. Um cuz I remember like when investors would outsmart themselves saying like, "Oh, Stripe's going to be a commodity. When

everybody gets big, they're going to just like build their own thing." And it turns out that like yes, there are multiple payments providers. And yet in practice, if you make a great product, if you make like a great thing, people

are still going to need to take money in the future. and they're probably just

the future. and they're probably just going to like stick with you if you're like a good reasonable ecosystem partner and don't do crazy things. Um, and I think that's going to keep working. I I

actually like I think every company clearly does need to get more efficient and with AI and they will do that but this this whole

mindset that like every company is going to go away and everything is going to be completely different and like maybe it'll be agents that need to handle

payments between each other instead of you know consumers and merchants but like clearly money is going to have to move somehow and my my advice vice would

be like adopt AI especially internally use AI to build better products but don't assume that the like entire socioeconomic system completely reconfigures I think the world has

gotten a little bit delusional about this you don't have to answer this but if I probably will um if we index OpenAI headcount to 100

today what do you think open I had counted in 5 years.

Um I would love it to be 200. I think

that see like I I we can clearly get phenomenally more efficient than we are now with these tools.

I probably we can't keep it at a 2x uh just because we need to like if we're going to have to go you know learn how to build data centers and robots and all these other kind of things like it's

just a lot of stuff to do. Um but you know we're maybe I think we're a lot more efficient than the large tech companies per per person and I I think we can turn that up even

more over time.

Um, apart from AI, as we look out over the next decade, just what are the technologies and areas that excite you?

Um, sort of AI at the kind of model and product layer. Um, I'm obsessed now with

product layer. Um, I'm obsessed now with data center infrastructure.

Like relevant I think there's so much cool stuff to do there. Uh I think there will be amazing new technologies like at the physical layer. Uh energy robots

those are made like that stack is probably what I think about the most. Um

it does seem like the world is finally making a little progress on um brain machine interfaces. I'm excited about

machine interfaces. I'm excited about that. Uh I am excited about

that. Uh I am excited about well I'm both afraid of and I'm excited about progress in biotech. Um, but you know, I'm certainly hopeful that that

can get much better quickly and and I think defensive biotech is about to become very important unfortunately. Um,

I'm excited about new kinds of computer interfaces. I think we are in a insane

interfaces. I think we are in a insane area right now where we're stuck with these kind of old devices and old

operating systems and we have this magic new enabling technology and it feels like Codeex is amazing and what it but it feels like very broken to

be like telling this thing to use my computer and then it's like clicking around and there's all the stuff that was made for a person but doesn't really make sense for like an AI to go use.

Like we can do so much better there. I

think there's a whole new internet protocol to make too. Um, so those things.

When do you think we'll have the world's first profitable nuclear fusion reactor?

Depends how far electricity prices get pushed by data center demand. Maybe

sooner than we thought. Um,

I'll guess in the next 5 years.

It's a bold prediction. I hope so.

Just going to hope we're on a roll here.

Um, hypersonic commercial air travel.

Don't follow it as closely. Um,

I hypersonic meaning like Mach 4. Sure.

A good chunk of time.

I don't know more than 10 years.

Okay.

Maybe a little less but something like that.

Um are there any domains of science and technology that are not in the discourse in a significant way that you think will be accelerated a lot by AI and will have broad kind of social consequence and

impact? Yeah. The one that I think just

impact? Yeah. The one that I think just does not get enough attention is material science. Um,

material science. Um, it's like a very it's not a cool thing and I think people underestimate how much of the world is materials like how

much of what we depend on and how much progress AI can make like it's such a beautifully AI AI shaped problem just getting new catalysts totally that I

expect very rapid progress there and that it'll impact all of our lives in a very positive way and it gets like very little attention.

Last question.

um it feels that in some sense at least some version of AI is kind of inevitable. There's lots of people

inevitable. There's lots of people rushing towards it and building it out and creating the data centers and uh and all the rest. So there's this kind of

sense of determinacy. Um, how do you hope that your specific involvement changes the trajectory for the world

relative to some other counterfactual?

Um, I believe in democratization and personal agency and access and that everybody deserves a really great life.

Um, the most controversial decision we made in the history of OpenAI was um, what we now call iterative deployment.

But a lot of the thinking at the time that we released chat GBT was this was insanely dangerous uh to do. You can't

do this. Only this like small set of people who have been thinking about AI safety can know what's coming. It's an

infazard to tell the world and it's all too dangerous to ever release. We have

to keep this locked up and we'll, you know, in our ivory tower, we will discover these wonderful things and we'll share the fruits with the world, but we'll have the AI and we'll control

it. And that sat very poorly with me.

it. And that sat very poorly with me.

And I thought then and I believe now that it is extremely important that we avoid that kind of power concentration and that we build this for the world and the world gets to use it in lots of

ways. Some of which will be good, not

ways. Some of which will be good, not all of which will be good. Um but that the by enabling people to explore this very wide opportunity space in front of us, messy at times though it will be and

obviously we'll put guard rails on it for reasonable safety. um we will give the world a gift but the world will give the world will build a much bigger gift on top of it for all of us and that if

you don't enable people with this technology and if you tried to keep it locked up which again now this may sound obvious but this like was the sort of rough consensus plan of people working

on it before we came along um I think that would have been really bad.

I am a believer in entrepreneurship and innovation and and that people do people are mostly good and mostly do amazing

things with tools. Um so by uh I think my single biggest contribution uh has been will be whatever pushing for this to be a democratized technology that people get to use and build on.

Sam Alman thank you very much.

Thank you very much.