LIVE: Sam Altman of OpenAI on Building the ‘Core AI Subscription’ for Your Life

0:00Hi, and welcome to Training Data. We are mixing it up for this week's episode and dropping a conversation that was filmed live at Sequoia's annual AI conference in San Francisco with OpenAI Founder and CEO Sam Oldman. Sam is interviewed by our partner Alfred Lin. We hope you enjoyed this special conversation with Sam about the Genesis story of ChatGbT, his predictions for agents in scientific discovery and robotics, and more. And stay tuned for a few more special AI -assent drops on our podcast feed later this week. Our next guest needs no introduction, so I'm not going to bother introducing him.

0:36Sam Altman, I will just say Sam is now three for three and joining us to share his thoughts at the three AI sense that we've had, which we really appreciate. So I just want to say thank you for being here. That's right. Oh, that's right. Say that again. Yeah, this was our first 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 GG X1 system over here. He did. It's amazing how small that thing looks now. Oh, versus what? Well, the current boxes are so huge, but yeah, it was a fun throwback.

1:15How heavy was it? That was still when you could kind of like lift one yourself. He said it was about 70 pounds. Yeah. I mean, it was heavy, but you could carry it. So, did you imagine that you'd be here today in 2016? No, it was like we were sitting over there, there were 14 of us or something. And you were hacking on this new system? I mean, even that was like, we were sitting around looking at whiteboards trying to talk about what we should do. like this was a, it's almost impossible to sort of overstate how much we were like a research lab with no, with a very strong belief in direction and conviction, but no real kind of like action plan.

2:02I mean, not only was like the idea of a company or a product sort of unimaginable, the specific, like LLMs as an idea were still very far off. And so we were trying to play video games. Trying to play video games. You still trying to play video games? Now we're pretty good at that. All right, so it took you another six years for the first consumer product to come out with just chat GPT along the way How did you sort of think about milestones to get something to that level as like an accident of history that the first Consumer product was not chat GPT. That's right. It was dolly. The first product was the API So we had built, you know, we kind of went through a few different things.

2:46We were, we had 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 do a robot hand, we're going to see if we can do a few other things. And at some point in there, 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 GPT -1 and then GPT -2. And by the time of GPT -3, we both thought we had something that was kind of cool, but we couldn't figure out what to do with it.

3:23And also, we realized we needed a lot more money to keep scaling. We had done GPT -3. We wanted to go to GPT -4. We were heading into the world of billion -dollar models. It's hard to do those as a peer science experiment unless you're like a particle accelerator or something. and even then it's hard. So we started thinking, okay, 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 GP2 out as model weights and not that much had happened. One of the things that I had just observed about companies, products in general is, If you do an API, it usually works somehow on the upside.

4:08This was like true across many, many YC companies. Also, if you make something much easier to use, there's usually a huge benefit to that. We're like, well, it's kind of hard to run these models that get in big. We'll go write some software to a really good job of running them. Also, we'll then rather than build a product because we couldn't figure out what to build. We will hope that somebody else finds something to build. And so I forget exactly when, but maybe it was June of 2020. We put out GPD 3 in the API. And the world didn't care, but sort of Silicon Valley did. They're like, oh, this is kind of cool.

4:45This is pointing at something. And there was this weird thing where we got almost no attention from most of the world. And some startup founders were like, oh, this is really cool. Or like, I mean, some of them are like this, the AGI. The only people that built real businesses with the GPT -3 API that I can remember were these company, a few companies that did like copyrightiness of service. That was kind of the only thing GPT -3 was over the economic threshold on.

5:15But one thing we did notice, which eventually they did to chat GPT, is even though people couldn't build a lot of great businesses with the GPT -3 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 RLHF to make it easy to chat with, but people love 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 GPD. By the time chat GPD 3 .5 came out, there were maybe like eight categories instead of one category where you could build a business with API, but our conviction that people just want to talk to the model had gotten really strong.

6:00So we had done Dali and Dali was doing okay, but we knew we kind of wanted to build, especially along with the fine tune we were able to do. We knew we wanted to build this model, this product that you talked to the model. And it launched in 2022? I think, yes. That's six years from when the first... November 30th 2022. Yeah. So there's 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 right. So by the way, get ready for some audience questions because that's what that was sans request. You've been here for every single one of the census Pat mentioned.

6:43And there's been lots of ups and downs. But it seems like the last six months, It's just been shipping, shipping, shipping. It shipped a lot of stuff, and it's amazing to see the product velocity, the shipping velocity continue to increase. So this is like multi sort of part question. How have you gotten a large company to increase product velocity over time? I think a mistake that a lot of companies make is they get big and they don't do more things. So they just 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.

7:20I am a big believer that you want everyone to be busy. You want teams to be small. You want 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. There was this old observation of business that like a good executive is a busy executive because you don't people like modeling around. But I think it's like a good, 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.

7:59So 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 the room fighting or meeting or talking about whatever. So we try to have relatively small numbers of people with huge amounts of responsibility. and the way to make that work is to do a lot of things. And also, we have to do a lot of things. To go kind of, I think we really do now have an opportunity to go build one of these important internet platforms. But to do that, if we really are going to be people's personalized AI that they use across many different services and over their life and across all of these different, all of these different main categories and all the smaller ones that we need to figure out and able, then that's just a lot of stuff to go build.

8:54Anything you're particularly proud of that you've launched in the last six months? I mean, the models are so good now. They still have areas to get better, of course, and we're working on that fast, but like, I think at this point, Chatchy BT is a very good product because the model is very good. There's other stuff that matters too, but I am 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 your ways and not be roadkill? I mean, 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.

9:46Some of that will be like what you do inside of chat GPT. 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. 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. And I hope that that enables just an unbelievable amount of wealth creation in the world and other people to build onto that.

10:34But yeah, we're going to go for the Core AI subscription and the model, and then the core surfaces, and there will be a ton of other stuff to build. 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 Core AI subscription offering Then I'll go ahead, that'd be great. OK. It's rumored that you're raising $40 billion or something like that $340 billion valuation on its rumors. I don't know if we announced it. OK. I just want to make sure that you announced it. What's your scale of ambition from there? From here. We're going to try to make great models and ship good products.

11:16And there's no master plan beyond that. But like we're gonna, I think like, no, I, I, I, I, I, I mean, there's, I see plenty of open air people in the audience, they can vouch for it. Like we don't, we don't sit there and have, like, 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. That doesn't usually work as well. Like the, the, 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. We know that we need to keep making models better.

11:54We 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 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 a set of products that people really, really love. And we have like unwavering confidence in that. And we believe we can build great models. I've actually never felt more optimistic about our research roadmap than I do right now.

12:33What's on the research roadmap? Really smart models.

12:38But in terms of like the steps in front of us, we kind of take those one or two at a time. So you believe them working forwards not necessarily working backwards? I have heard some people talk about these brilliant strategies of how they're this is where they're gonna go and they're gonna 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 has a question? There's a mic coming your way being thrown

13:14What 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 smaller companies are clearly just beating the crap out of larger ones when it comes to innovation here. I think this basically happens every major tech revolution. There's nothing to be 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 stuck in their ways.

13:48If 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 are going to allow and what what it means to put data into a system. It's so painful to watch what happens here. But this is creative destruction. This is why startups win. This is how the industry moves forward. I feel like it disappointed but not surprised at the rate that big companies are willing to do this. They will, my prediction would be that there's another couple of years of fighting, pretending like this isn't going to reshape everything.

14:27and 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. 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 use it or some other service. And the difference is unbelievable. It reminds me of when the smartphone came out and every kid was able to use it super well and older people just took three years to figure out how to do basic stuff.

15:11And then of course people integrate. But this 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, what are the cool use case that you're seeing young people using with touchypt that might surprise us?

15:33They really do use it like an operating system. 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 header like you know in something where they paste in and out.

15:51The, I mean that stuff I think is all cool and impressive and there's this other thing where they don't really make life decisions without asking, like chat GBT, what they should do. And it has the full context on every person in their life and what they've talked about. The memory thing has been a real change there. But yeah, I think it grows oversimplification, but older people use chat GBT as a Google 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 OpenAO? Um, I mean it writes a lot of our code.

16:34How much? I don't know the number. And also when people say the number, I think it's always this very dumb thing because like, Microsoft code is 30, 20, 30 % rich by lines of code is just such an insane way to like, I don't, Maybe the meaningful thing I could say is, it's writing meaningful code. I don't know how much, but it's like writing the parts that actually matter. That's interesting. Next question. Hey Sam. Mike, going over. Okay, hey Sam. 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.

17:13And also most of your revenue comes from consumer subscriptions, why keep the API in 10 years? I really hope that all of this mergers into one thing. Like, you should be able to sign in with OpenAI to other services. Other services should have an incredible SDK to take over the chat GPT UI at some point. But 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 has all this context on you. So you'll want to be able to use that in a lot of places. Now I agree that the current version of the API is very far off that vision, but I think we can get there.

17:55Yeah, maybe I have a follow up question to that when you kind of took mine. But like a lot of us who are building application layer companies, we want to like use those building boxes, different API components. Maybe the deep research API, which is not a release thing, but could be, and build stuff with them. Is that going to be a priority, enabling that platform for us? How should we think about that? I think I hope something in between those, that there is a new protocol on the level of HTTP for the future of the internet, where things get federated and broken down into much smaller components, and agents are constantly exposing and using different tools and authentication, payment, data transfer.

18:40It's all built in at this level that everybody trusts, everything you talk to, everything.

18:48And I don't quite think we know what that looks like, but it's like coming out of the fog. And as we get a better sense for that, again, it 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.

19:08Hey Sam, back here. My name is Roy. I'm curious. The AI would obviously do better with more input data. Is there any thought to feeding sensor data and what type of sensor data, whether it's temperature, you know, things in the physical world that you could feed in that it could better understand reality? people do that a lot. 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 O3 API caller whatever. And for some use cases, it does work super well. I'd say that the latest models seem to do a good job with this and they use to not.

19:48So we'll probably bake it in more explicitly at some point, but there's already a lot happening there. Hi Sam. I was really excited to play with the voice model in the playground. And so I have two questions. The first is how important is voice to open AI 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 good enough text model too. We will crack that code eventually.

20:26And when we do, I think a lot of people are going to want to use voice interaction a lot more. 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 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 gooey interaction that we have not cracked. But before that, we'll just make voice really great. And when we do, I think there's not only, is it cool with existing devices, but I sort of think voice will enable a totally new class of devices if you can make it feel like truly human level voice.

21:12Some more questions about coding. I'm curious, is coding just another vertical application or is it more central to the future of OpenAI? That one's more central to the future of OpenAI. I think coding will be how these models, kind of, right now if you ask Cheshire PT, a response, you get text back, maybe you get an image. You would like to get a whole program back. You would like custom rendered code for every response, or at least I would. 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, like actuate the world and call a bunch of APIs or whatever.

21:53So I would say coding will be more in a central category, well obviously expose it through API on our platform as well. But, you know, chat GPT should be excellent at writing code. So we're gonna move from the world of assistance to agents to basically applications all the way through. I think it'll feel, yeah, it's like very continuous, but yes. So you have conviction in the roadmap 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 maybe isn't in the mental model of most of us?

22:51I mean, that's kind of the, each of those things are really hard. And obviously, the highest leverage thing is still big algorithmic breakthroughs. And I think there still probably are some 10Xs or 100Xs left, not very many, but even one or two is a big deal. But yeah, it's kind of like algorithms, data, compute, those are sort of the big ingredients. questions. Hi. So my question is you run one of the best ML teams in the world. How do you balance between letting smart people like Isa, cheese, deep research or something else that seems exciting, where is it 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 will work.

23:39There are some projects that require so much coordination that there has to be a little a little bit of top down quarterbacking, but I think most people try to do way too much of that. 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 what a well run research lab looks like. And you had to go really far back in the past. In fact, almost everyone that could help it rises on this was dead. It had been a long time since there had been good research labs. And people ask us a lot, why does open AI repeatedly innovate and why do the other AI labs sort of copy or why do BioLab X not do good work and BioLab Y does do good work or whatever.

24:35And we keep saying, here's the principles we've observed, here's how we learned them, Here's what we looked at in the past. And then everybody says great, but I'm going to go do the other thing. That's fine. Like you came to us for advice. Like you do what you want. 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 good research labs in history have worked for us. And then people who have had some smart reason about why they were going to do something else that didn't work.

25:09So it seems to me that these large models, one of the really fascinating things, there's like a lover of knowledge about them, is that they potentially embody and allow us to answer these like amazing longstanding questions in the humanities about cyclical changes in artistic, interesting things or even like, you know, to what extent, to systematic prejudice and other sorts of things are really happening in society and can we sort of detect these? And I'm very subtle things, which 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.