OpenAI’s Sam Altman on Building the ‘Core AI Subscription’ for Your Life

Our next guest needs no introduction, so I'm not going to bother introducing him Sam Alman. I will just say Sam is now

Sam Alman. I will just say Sam is now three for three and joining us to share his thoughts at the three AI events that we've had, which we really appreciate.

So, I just want to say thank you for being here. This was our first office.

being here. This was our first office.

That's right. Oh, that's right. Say that

again. Yeah, this was this was our first office. So, it's nice to be back. Let's

office. So, it's nice to be back. Let's

go back to the first office here. He

started in 2016.

2016. We just had Jensen here who said that he delivered the first GGX1 system over here. He did. Yeah. It's amazing

over here. He did. Yeah. It's amazing

how small that thing looks now. Oh.

Versus what? Well, the current boxes are still huge. But yeah, it was a fun

still huge. But yeah, it was a fun throwback. How heavy was it? That was

throwback. How heavy was it? That was

still when you could kind of like lift one yourself.

He said it was about 70 lbs. Yeah. I

mean, it was heavy, but you could carry it. So, um, did you imagine that you'd

it. So, um, did you imagine that you'd be here today in 2016?

Uh, no. It was like, uh, we were sitting over there and there were, you know, 14 of us or something and you were hacking on this new system. I mean, even that was like a we were sitting around like

looking at whiteboards trying to talk about what we should do. Like this was a ve it's almost impossible to sort of overstate how

much we were like a research lab with no with with a very strong belief and direction and conviction but no real kind of like action plan. I mean, not only was like the idea of a company or a

product sort of unimaginable, the spec like LLMs as an idea were still very far off. And so, trying to play video games.

off. And so, trying to play video games.

Trying to play video games. Are you

still trying to play video games? Now

we're pretty good at that.

Um, all right. So, um, it took you another six years for the first consumer product to come out, which is ChatGpt.

along the way. How did you sort of think about milestones to get something to that level as like an accident of history? The the uh the first consumer

history? The the uh the first consumer product was not Chacht. That's right. It

was Dolly. Um the first product was uh the API. So we had built, you know, we kind

API. So we had built, you know, we kind of went through a few different things.

We we were we had a a few directions that we really wanted to bet on.

Eventually, as I mentioned, we said "Well, we got to build a system to see if it's working. And we're not just writing research papers. So, we're going to see if we can, you know, play a video game. Well, we're going to see if we can

game. Well, we're going to see if we can do a robot hand. We're going to see if we can do a few other things. And at

some point in there, uh, one person and then initially and then eventually a team got excited about trying to do unsupervised learning and to build language models. And that led to GPT1

language models. And that led to GPT1 and then GPT2 and and and by the time of GBT3, we both thought we had something that was kind of cool, but we couldn't

figure out what to do with it. Um, and

also we realized we needed a lot more money to keep scaling. You know, we had done GBD3, we wanted to go to GPT4. We

were heading into the world of billion-dollar models. It's like hard to

billion-dollar models. It's like hard to do those as a pure science experiment unless you're like a particle accelerator or something. Um, even then it's hard. So, we started thinking

it's hard. So, we started thinking okay, we we both need to figure out how this can become a business that can sustain the investment that it requires.

And also like we have a sense that this is heading towards something actually useful and we had put GPT2 out as model weights and not that much had happened.

Um one of the things that I had just observed about uh companies products in general is if you do an API it usually works somehow

on the upside. This was like true across many many YC companies. And also that if you make something much easier to use there's usually a huge benefit to that.

So we're like, well, it's kind of hard to run these models, they're getting big, we'll go write some software, do a really good job of running them. And and

also we'll then rather than build a product cuz we couldn't figure out what to build. Um we will hope that somebody

to build. Um we will hope that somebody else finds something to build. And so I forget exactly when, but maybe it was like June of 2020. Um, we put out GPT3

in the API and the world didn't care, but sort of Silicon Valley did. They're like, "Oh this is kind of cool. This is pointing at something." And there was this weird

at something." And there was this weird thing where like we got almost no attention from most of the world. And

some startup founders uh were like, "Oh this is really cool." Or like I mean some of them are like this is AGI. Um

the only people that built real businesses with the GPT3 API that I can remember were these company a few companies that did like copyrightiting as a service. That was

kind of the only thing GPT3 was over the economic threshold on.

Um, but one thing we did notice which eventually led to ChatGpt is even though people couldn't build a lot of great businesses with the GPT3 API, people love to talk to it in the playground.

And it was terrible at chat. We had not at that point figured out how to do RHF to make it easy to chat with, but people loved to do it anyway.

And in some sense, that was the kind of only killer use other than copyrighting of the API product that led us to eventually build chat GPT. By the time

Chat GPT 3.5 came out, there were maybe like eight categories instead of one category where you could build a business with the API. Um but but our conviction that like people just want to

talk to the model had gotten really strong. So we had done Dolly and Dolly

strong. So we had done Dolly and Dolly was doing okay but we knew we kind of wanted to build especially along with um the finetuning we were able to do we knew we wanted to build this model this

product to let you talk to the model and it launched in 2022 or something uh I think yeah about six years when the

first November 30th 2022 yeah so there was a lot of work leading up to that and 2022 launched today it has over 500 million people who talk to it on a

weekly basis. Yeah. All right. All

weekly basis. Yeah. All right. All

right. So, um, by the way, uh, get ready for some audience questions because that's what that was Sam's request. Um

you've been here for three every single one of the ascents as Pat mentioned and there's been some lots of ups and downs but seems like the last 6 months he's just been shipping, shipping, shipping.

We shipped a lot of thought stuff and it's amazing to see the product velocity, the shipping velocity continue to increase. So this is like multi- sort

to increase. So this is like multi- sort of part question. How have you gotten a large company to like increase product velocity over time? I I I think a

mistake that a lot of companies make is they get big and they don't do any they don't do more things. So they just like get bigger because you're supposed to get bigger and they still ship the same amount of product. And that's when like

the molasses really takes hold. I I like I am a big believer that you want everyone to be busy. You want teams to be small. you want like to do a lot of

be small. you want like to do a lot of things relative to the number of people you have otherwise you just have like 40 people in every meeting and huge fights over who gets like what tiny part of the

product. Um there there was that there

product. Um there there was that there was this like old observation of business that like a a a good executive is a busy executive because you don't people like muddling around. Um

but I I think it's like a good I you know at our company and many other companies like researchers engineers, product people, they drive almost all the value and you want those

people to be busy and high impact. So if

you're going to grow you better do a lot more things otherwise you kind of just have a lot of people sitting in your room fighting or meeting or talking about whatever. Um so we try to have you

about whatever. Um so we try to have you know relatively small numbers of people with huge amounts of responsibility. Um

and the way to make that work is to do a lot of things.

And also like we have to do a lot of things like that the to go kind of I think I think we we really do now have

an opportunity to go build one of these important internet platforms. Um but to do that like if we really are going to

be people's like personalized AI that they use across many different services and you know over their life and across all of these different all these different like kind

of main categories and all the smaller ones that we need to uh figure out how to enable then that's just a lot of stuff to go build. Anything you're

particularly proud of that you've launched in the last six months?

I mean, the models are so good now.

Like, they they still have areas to get better, of course, and we're working on that fast, but like I think at this point, Chad GBT is a very good product because the model's

very good. I mean, there's other stuff

very good. I mean, there's other stuff that matters, too, but the I am like I'm amazed that one model can do so many things so well. You're building small

models and large models. You're doing a lot of things as you said. So, how do this audience stay out of uh your ways

and not be roadkill?

Um, I mean, like I I I think the way to model us is we want to build we want to be people's like core AI subscription and way to use that thing. Some of that

will be like what you do inside of Chad GPT.

Um, we'll have a couple of other kind of like really key parts of that subscription.

But mostly we will hopefully build this smarter and smarter model. We'll have

these surfaces like future devices, future things that are sort of similar to operating systems, whatever. Um, and then you know

systems, whatever. Um, and then you know we want we have not yet figured out exactly I think what the sort of

API or SDK or whatever you want to call it is to like really be our platform.

But we will. It may take us a few tries but we will. Um, and I hope that that enables like just an unbelievable amount of wealth creation in the world and other people to build onto that. But

yeah, we're going to go for like the core AI subscription and the model and then um the kind of core services and there will be a ton of other stuff to build. Okay. So, don't be the Core AI

build. Okay. So, don't be the Core AI subscription, but you can do everything else.

We're going to try. I mean, if you can make a better Corei subscription offering than us, go ahead. That'd be

great. Okay. Um, it's rumored that you're raising $40 billion or something like that at $340 billion valuation.

It's rumors that it's I don't know if I think we announced it. We're okay. Well

if I just want to make sure that you announced it. Um, what's the what's your

announced it. Um, what's the what's your scale of ambition from there? From here

we're going to like try to make great models and ship good products and there's no master plan beyond that. Like

we're gonna I I I think like Sure. No, I

I I I mean there's I see plenty of open eye people in the audience. They can

vouch for that. Like we don't we don't sit there and have like I I am a big believer that you can kind of like do the things in front of you, but if you like try to work backwards from like kind of we have this crazy complex

thing.

Um that doesn't usually work as well.

Like the the we we know that we need tons of AI infrastructure. like we know we need to go build out massive amounts of like AI factory volume. Um we know that we need to keep making models

better. We know that we need to like

better. We know that we need to like build a great top of the stack like kind of consumer product and all the pieces that go into that.

But we pride ourselves on being like nimble and adjusting tactics as the world adjusts. And so the products

world adjusts. And so the products um you know the products that we're going to build next year we're probably not even thinking about right now.

And we believe we can build uh a set of products that people really really love.

Um and we have like unwavering confidence in that and we believe we can build great models. I' I've actually never felt more optimistic about our research road map than I do right now.

Um what's on the research road map?

Really smart models.

Um but but in terms of like the steps in front of us, we kind of take those one or two at a time. So you believe in working forwards, not necessarily working backwards. I have heard some

working backwards. I have heard some people talk about these brilliant strategies of how they're this is where they're going to go and they're going to work backwards and you know this is take over the world and this is the thing before that and this is that and this is that and this is that and this is that

and here's where we are today. I have

never seen those people like really massively succeed. Got it. Who who has a

massively succeed. Got it. Who who has a question? There's a mic coming your way

question? There's a mic coming your way being thrown.

Um, what do you think the larger companies are getting wrong about transforming their organizations to be more AI native in terms of both using the tooling as well as producing products? It's been, you know, it's

products? It's been, you know, it's smaller companies are clearly just beating the crap out of out of larger ones when it comes to innovation here. I I think this basically happens

here. I I think this basically happens every major tech revolution. Um, there's

nothing to me surprising about it. The

thing that they're getting wrong is the same thing they always get wrong, which is like people get incredibly stuck in their ways. Organizations get incredibly

their ways. Organizations get incredibly stuck in their ways. If things are changing a lot every quarter or two and you have like an information security council that meets once a year to decide what

applications you're going to allow and what it means to like put data into a system, like it it's just it's so painful to watch what happens here. But

like, you know, this is this is creative destruction. This is why startups win.

destruction. This is why startups win.

this is like how the industry moves forward. Um I am I'd say I feel like

forward. Um I am I'd say I feel like disappointed but not surprised at the rate that big companies are willing to do this. Um they

do this. Um they will my kind of prediction would be that there's another like couple of years of fighting pretending like this isn't going to reshape everything and then there's like a capitulation and a last

minute scramble and it's sort of too late and in general startups just sort of like blow past people doing it the old way. Um, I mean this happens to

old way. Um, I mean this happens to people too like watching watching like a you know someone who

started maybe you like talk to an average 20-year-old and watch how they use chat GBT and then you go talk to like an average 35-year-old and how they they use it or some other service and

like the difference is unbelievable. It

reminds me of like, you know, when the smartphone came out and like every kid was able to use it super well and older people just like took like three years to figure out how to do basic stuff. And

then of course people integrate, but but the the sort of like generational divide on AI tools right now is crazy and I think companies are just another symptom of that.

Anybody else have a question?

Just to follow up on that, um, what are the cool use cases that you're seeing young people using with Touch EPT that might surprise us?

They really do use it like an operating system.

um they have like complex ways to set it up to connect it to like a bunch of files and they have like fairly complex prompts memorized in their head or like you know in something where they paste

in and out and um the I mean that stuff I think is all cool and impressive and there's this other thing where like they don't really make life decisions without asking like

chbt what they should do. Um, and it has like the full context on every person in their life and what they've talked about and you know that like the memory thing has been a real change there. But but

yeah I think gross oversimplification, but like older people use Chachi PT as a Google replacement. Maybe people in their 20s

replacement. Maybe people in their 20s and 30s use it as like a life advisor something and then like people in college use it as an operating

system.

How do you use it inside of OpenAI?

Um, I mean it writes a lot of our code.

How much? I don't know the number. And

also when people say the number I think is always this very dumb thing cuz like you said Microsoft code is 30 20 30% written measuring by lines of code is just such an insane way to like I don't

I I maybe the meaningful thing I could maybe the thing I could say is it's writing meaningful code like it's writing I don't know how much but it's like writing the the the parts that actually matter. That's That's

actually matter. That's That's interesting. Next question. Hey Sam, mic

interesting. Next question. Hey Sam, mic going away. Is this okay? Hey, Sam. Uh

going away. Is this okay? Hey, Sam. Uh

I thought it was interesting that the answer to Alfred's question about where you guys want to go is focus mostly around consumer and being the core subscription and also most of your

revenue comes from consumer subscriptions. Why keep the API in 10

subscriptions. Why keep the API in 10 years? I really hope that all of this

years? I really hope that all of this merges into one thing. like you should be able to sign in with OpenAI to other services. Other services should have an

services. Other services should have an incredible SDK to like take over the chat GBT um UI at some point. But like

to the degree that you are going to have a personalized AI that knows you, that has your information, that knows what you want to share later, and you know has all this context on you, you'll want to be able to use that in a lot of

places. Now, I agree that the current

places. Now, I agree that the current version of the API is very far off that vision, but I think we can get there.

Uh yeah, I maybe have a follow-up question to that one. You kind of took mine. Um but like a lot of us who are

mine. Um but like a lot of us who are building application layer companies, we want to like use those building blocks those different API components, maybe the deep research API which is not a release thing but could be uh and and

build stuff with them like is that going to be a priority like enabling that platform for us? How should we think about that? Yeah

about that? Yeah I I think I hope something in between those that there is sort of like a new protocol on the level of HTTP for the

future of the internet where things get federated and broken down into like much smaller components and agents are like constantly exposing and using different

tools and authentication, payment, data transfer, it's all like built in at this level that everybody trusts. everything

talk to everything.

And I I don't quite think we know what that looks like, but it's like coming out of the fog. Um, and as we get a better sense for that, again, it'll probably take us like a few iterations toward that to get there. But that's

kind of where I would like to see things go.

Hey, Sam. Uh, back here. Uh, my name is Roy. I'm curious, uh, the AI would

Roy. I'm curious, uh, the AI would obviously do better with more input data. Is there any thought to feeding

data. Is there any thought to feeding sensor data, uh, and what type of sensor data, whether it's temperature, uh, you know, things in the physical world that

you could feed in that it could better understand reality. People do that a

understand reality. People do that a lot. uh people like put that into you

lot. uh people like put that into you know people have whatever they build things where they just put sensor data into like an API and like an 03 API call or whatever and for some use cases it does work super well. Um I'd say that

the latest models seem to do a good job with this and they used to not. Uh so

we'll probably bake it in more explicitly at some point but there's already like a lot happening there.

Hi Sam. Uh I was really excited to play with the voice model in the playground and so I have two questions. The first

is how important uh is voice to open AAI in terms of like stack ranking for infrastructure and can you share a little bit about how you think it'll show up in the product and chat GBT the

core thing.

I think voice is extremely important.

Honestly, we just we have not made a good enough voice product yet. That's

fine. Like it took us a while to make a good enough text model too. Um, we will crack that code eventually and when we do, um, I think a lot of people are

going to want to use voice interaction a lot more. I I am super when we first

lot more. I I am super when we first launched our current voice mode, the thing that was most interesting to me was it was a new stream on top of like the touch interface and I you could talk

and be like clicking around on your phone at the same time. And I continue to think there's something amazing to do about like voice plus guey interaction

that we have not cracked. But before

that, we'll just make voice really great. And when we do, I think there's a

great. And when we do, I think there's a not only is it cool with existing devices, but I I sort of think voice will enable a totally new class of devices if you can make it feel like

truly human level voice.

Similar question. Similar question about coding. I'm curious, is coding just

coding. I'm curious, is coding just another vertical application or is it more central to the future of open AI?

That one's more central to the future of open AI. Um, coding I think will be

open AI. Um, coding I think will be how these models kind of right now if you ask CHP a response

you get text back, maybe you get an image. Um, you would like to get a whole

image. Um, you would like to get a whole program back. You would like, you know

program back. You would like, you know custom rendered code for every response or at least I would. um you would like the ability for these models to go make things happen in the world and writing code I think will be very central to how

you like actuate the world and call a bunch of APIs or whatever. So I I would say coding will be more in a central category. We'll obviously

expose it through our API on our platform as well. Um but you know chat GBT should be excellent at writing code.

So we're going to move from the world of assistants to agents to basically applications all the way through.

I I I think it'll feel yeah like very continuous but yes u so you have conviction in the road map about smarter models awesome I have this

mental model there's some ingredients like more data bigger data centers a transformer architecture test time compute what's like an underrated ingredient or something that's going to

be part of that mix that like maybe isn't in the mental model of most of Um, I mean that's kind of the each of those things are really hard and you know obviously

like the highest leverage thing is still big algorithmic breakthroughs and I think there still probably are some 10 x's or 100 x's left not very many but

even one or two is a big deal.

Um but you know yeah it's kind of like algorithms, data, compute those are sort of the big ingredients.

Uh hi uh so my question is you run one of the best ML teams in the world. Uh, how

do you balance between uh letting smart people like Issa chase uh deeply research or something else that seems exciting versus going top down and being like we're going to build this, we're going to make it happen. We don't know

if it'll work. There are some projects that require so much coordination that there has to be a little bit of like top down quarterbacking, but I think most people try to do way too much of that.

I I I mean this is like there's probably other ways to run good AI research or good research labs in general, but when we started OpenAI, we spent a lot of

time trying to understand uh what a well-run research lab looks like. And you had to go really far back

like. And you had to go really far back in the past. In fact, almost everyone that could like help advise us on this was dead. Um it had been like a long

was dead. Um it had been like a long time since there had been good good research labs.

And you know people ask us a lot like why why does open AI like repeatedly innovate and why do the other AI labs like sort of copy or why do like biolab

x not do good work and biolab y does do good work or whatever. And we sort of keep saying like here's the principles we've observed. Here's how we learned

we've observed. Here's how we learned them. Here's what we looked at in the

them. Here's what we looked at in the past. And then everybody says great um

past. And then everybody says great um but I'm going to go do the other thing.

We said that's fine. Like you came to us for advice. Like you do what you want.

for advice. Like you do what you want.

Um but I find it remarkable how much these few principles that we've tried to run our research lab on which we did not invent. We shamelessly copied from other

invent. We shamelessly copied from other good research labs in history um have worked for us. And then people who have had some smart reason about why they were going to do something else, it didn't work.

Um, so it seems to me that uh these large models uh one of the really fascinating things as like a lover of knowledge about them is that they potentially embody and allow us to

answer these like amazing long-standing questions in the humanities about cyclical changes and artistic uh interesting things or even like uh you know to what extent systematic prejudice

and other sorts of things are really happening in society and can we sort of detect these and I'm uh very subtle things which we we could never really do more than hypothesize before. And I'm

wondering whether OpenAI has a thought about or even a roadmap for working with academic researchers say to help unlock some of these new things we could learn

for the first time in the humanities and in the social sciences. We do um yeah I mean it's amazing to see what people are doing there. We do have academic

doing there. We do have academic research programs where we partner and you know do some custom work but mostly people just say like I want access to the model or maybe I want access to the

base model and I think we're really good at that. Uh one of the kind of cool

at that. Uh one of the kind of cool things about what we do is so much of our incentive structure is pushed towards making the models as smart and cheap and widely accessible as possible

that that serves academics and the really the whole world very well. So

you know, we we have we do some custom partnerships, but we often find that what researchers or users really want is just for us to make the general model

better across the board. Um, and so we we try to focus, you know, kind of 90% of our thrust vector on that.

I'm curious how you're thinking about customization. So, you mentioned the

customization. So, you mentioned the federated like sign in with OpenAI bringing your memories, your context.

I'm just curious if you think customization and like these different post- training on like application specific things is a band-aid for or trying to make the core models better and how you're thinking about that.

I mean in some sense I think the like platonic ideal state is uh a very tiny reasoning model with a trillion tokens of context that you put your whole life

into. The model never retrains. The

into. The model never retrains. The

weights never customize. But that thing can like reason across your whole context and do it efficiently. And every

conversation you've ever had in your life, every book you've ever read, every email you've ever read, um, every everything you've ever looked at is in there, plus connected all your data from other sources. And, you know, your life

other sources. And, you know, your life just keeps appending to the context and your company just does the same thing for all your company's data. Um, we

can't get there today. Uh but but I I think of kind of like anything else as a a compromise off that platonic ideal and that is how I would eventually I hope we

do customization. One last question in

do customization. One last question in the back. Hi Sam, thanks for your time.

the back. Hi Sam, thanks for your time.

Where do you think most of the value creation we come from in the next 12 months? Would it be maybe advanced

months? Would it be maybe advanced memory capabilities or maybe security or protocols that allow agents to do more stuff and interact with the real world?

Um I mean in some sense the value will continue to come from really three things like building out more infrastructure, smarter models and building the kind of scaffolding to

integrate this stuff into society. And

if you push on those, I think the rest will sort itself out. Um, at at a higher level of detail, I kind of think 2025 will be a

year of sort of agents doing work.

Coding in particular, I would expect to be a dominant category. I think there'll be a few others too. Um, next year is a year where I would expect more like

uh sort of AI discovering new stuff and maybe we have AIs make some very large scientific discoveries or assist humans in doing that. And you know, I'm I am

kind of a believer that most of the sort of real sustainable economic growth in human history comes from once you've like kind of spread out and colonized the earth, most of it comes from just

better scientific knowledge and and then implementing that for the world. And

then 27 I I would guess is the year where like that all moves from the sort of intellectual realm to the physical world and robots go from a curiosity to like a serious economic creator of

value. But that was like an off the top

value. But that was like an off the top of my head kind of guess right now. Can

I close with a few quick questions?

Great. One of which is um chat uh GPT5.

Is that going to be just all smarter than all of us here?

Um I mean if you think you're like way smarter than 03 then maybe you have a little bit of a ways to go but 03 is already pretty smart.

Um two personal questions. Last time you were here, you had just come off a blip with open AI. Uh given some perspective

now in distance, you got any advice for founders here about resilience endurance strength?

Um it gets easier over time. I think

like you will face a lot of adversity in your journey as a founder and the the kind of challenges get harder and higher stakes but the

emotional toll gets easier as you kind of go through more bad things. So it's

uh you know in some sense like it does it yeah even though like abstractly the challenges get bigger and harder the your ability to deal with

them the sort of resilience you build up gets easier like with each one you you kind of go through.

Um and then I like I I think the the hardest thing about the big challenges that come as a founder is not the moment when they happen. Uh like a lot of

things go wrong in the history of a company.

Um in the acute thing you can kind of like you know you get a lot of support you can function a lot of adrenaline.

like that's, you know, you're kind of like e even the really big stuff like your company runs out of money and fails, like a lot of people will come and support you. Um, and you kind of get through it and go on to the new thing.

The thing that I think is harder to sort of manage your own psychology through is the sort of like fallout after. Um, and

I think if there's, you know, people focus a lot about how to work in that one moment during the crisis. And the

really valuable thing to learn is how you like pick up the pieces. There's

much less talk about that. I think

there's I've never actually found something good to point founders to to go read about, you know, not how you deal with the real crisis on day zero or day one or day two, but on day 60 as

you're just trying to like rebuild after it. Um, and that's that's the area that

it. Um, and that's that's the area that I think you can like practice and get better at. Thank you, Sam. Yeah, you're

better at. Thank you, Sam. Yeah, you're

officially still on paternity leave. I

know. So, thank you for coming in and speaking with us. Appreciate it. Thank

you.