[0:00:00] That's the sound of a butt being worked clean off. Exactly the kind of worked off butt Duluth Trading builds underwear for. Underwear that works. Duluth Trading. Underwear for folks who work their butts off. We have been developing incredible new technology for a long time. And each time it's happened, people have had these conversations. The industrial revolution comes along. Machines take all of our jobs. What does this mean? Computer revolution comes along. What does it mean?
[0:00:31] And the answer, at least in terms of what it means to be human, is not very much. This is Andrew Ross Sorkin with The New York Times, and you're listening to interviews from our annual Dealbook Summit live event, recorded on December 4th in New York City.
[0:00:51] Thank you so much, Sam. Thanks for having me. For being here. We are super appreciative. Sam, of course, is the co-founder and CEO of OpenAI. As I said earlier, one of the most consequential companies of our time. In just the past few years, since the launch of ChatGPT, Sam and his team have brought AI to the mainstream. He's fundamentally really altered how we work, communicate, innovate, and now think about what our future is even going to be like. His company is now worth about $157 billion.
[0:01:19] And I should say, OpenAI didn't just spark an arms race. It really redefined the possibilities of technology. And as he pushes towards artificial general intelligence, we're forced to grapple with some very big questions about the future of work, about the balance of power, about ethical questions, and so many other things. And we're going to delve into all of that with Sam. So thank you, the man at the center of it all. Thanks for having me. So here's where we'll start. Two years ago,
[0:01:48] Almost to this week, basically, you launched ChatGPT. And I think it's fair to say that when you press the button, since then, all hell has broken loose. It does feel like that. It has changed the global conversation. It has changed fundraising, priorities, the move of resources. Other technology companies have shifted the way they're doing things. There have been lawsuits. All of it.
[0:02:14] And so I'm very curious if you could just take us back two years ago, literally almost, as I said, to the day or to the week, and what you thought was going to happen when you pressed the button. So in the abstract, we always knew there would be some moment where, for whatever reason, the world would go from not getting it to getting it. And that all of a sudden, it would be clear that this technology was working. Internally, at that point, it seemed very obvious to us that...
[0:02:42] Language models were going to keep scaling. They were going to do all these useful things. And why it happened at the launch of ChatGPT, I mean, I have some theories, but why it happened then and not when we launched the GPT-3 and the API earlier, or why it didn't happen until GPT-4 launched several months later, like why that exact moment? I think there is some like chanciness to when it actually catches fire. We had observed, though, with the API,
[0:03:12] And GPT-3 that, you know, GPT-3 was a little bit early and it didn't work for that many things. But in some sense, one of the killer use cases of it was developers in the playground, which is where you could like test ideas quickly before you implement the API. They love just talking to them all. They would just sit there and like talk to GPT-3 conversationally about whatever. And that was kind of the main thing people were using it for. So we said, well, if that's what people want, we can make it much easier to use. You don't have to sign up for a developer account and do all these other things. And we can sort of tune it.
[0:03:42] To be good at conversations. And so we said, okay, let's make this as a product. Now, we'd actually been planning to launch it with GPT-4. We finished GPT-4 in like August of 2022. But we held it for a while. But did you expect all of this to happen? Again, at some point, we thought there would be a big moment in the world. Did we know it was going to be right then with the launch of ChatGPT? Certainly not. So here we are now, two years later. And the next big question.
[0:04:10] I can't even believe we're just asking it two years later. And you actually just wrote about it. You said that it is possible that we will have superintelligence in a few thousand days from now. Two years ago, did you think that? And when you say a few thousand days, I can do various versions of math on what that means. Well, I mean, there is wide uncertainty. And there's also different definitions of what superintelligence is.
[0:04:40] And yes, two years ago, we thought we were on a pretty steep curve. I mean, we started OpenAI because we thought this was possible and maybe not that far away. And it has gone well, but we hoped it was going to go well. Like we believed in deep learning as this unbelievable new discovery of humanity. And we thought it could go here and we had a duty to get it here and to make it go well and to broadly share the benefits.
[0:05:07] Like what? Agents are the thing everyone is talking about, I think, for good reason.
[0:05:38] This idea that you can give an AI system a pretty complicated task, like a kind of task you'd give to a very smart human that takes a while to go off and do and use a bunch of tools and create something of value. That's the kind of thing I'd expect next year. And that's like a huge deal. We talk about that like, oh, you know, this thing is going to happen. But if that works as well as we hope it does, that can really transform things. Okay, so what are we supposed to think? There's been a whole bunch of headlines in the past couple of weeks.
[0:06:07] Where people have talked actually about maybe that the scaling laws of AI actually are slowing down. And you put out this, I don't know if it was cryptic or it was very direct. You put out a tweet where you wrote, there is no wall. I don't see how it could be any less cryptic than that. Like, that was an attempt not to be cryptic at all and say, hey, this is going to keep going. In fact, hopefully this will put it to rest at least for a while.
[0:06:35] I've always been struck by how much people love to speculate on, is there a wall? Is scaling going to keep going? Rather than just look at the curve of progress and say, maybe I shouldn't bet against an exponential like that. In any case, we have a bunch of new great stuff. So we're doing this kind of fun thing for the holidays. We're doing 12 days of OpenAI starting tomorrow. We'll either launch something or do a demo every day for the next 12 weekdays. And without saying too much about...
[0:07:05] The thing for tomorrow. It is continued progress. And give us a little bit more of a hint then if it's going to start tomorrow. Start today, right now. I don't want to spoil your surprise. It's not that long to wait. I'm not good with that. I want to know. One of the big questions about AI right now, though, is, and this goes to the scaling question, is...
[0:07:32] How much data and what it takes to actually continue to scale? How much of this is about just pure processing power and therefore capital, right? Versus data, where you get that data, this idea of synthetic data, digital data that we're going to recreate data on top of data. What does it actually take to scale at this point? When people even talk about some of the limiting factors, I think that there's some question embedded in that about...
[0:08:01] Which piece is going to get there? So the three key inputs, and there are others too, but the three key ones are compute, data, and algorithms. And you can trade them off sometimes. Like if you have much better data, you could use less compute. Or if you have a lot more compute, you could use that to create synthetic data and make more data. For a while, there was incredible easy games in front of us with compute. More recently, we've had a lot of algorithmic games. But you want to push on all of those inputs at the same time.
[0:08:32] And at different times, there may be differential returns to one or the other. But I kind of think of all three. But for you, compute, I assume, is the biggie. And if you don't have the compute, all the others don't matter. I mean, in some sense, algorithmic progress is the biggie. The biggest one. Maybe you can double the size of a computer. Maybe over time, with crazy amounts of capital, you can 10x the size of a computer. Once in a while, these 1,000x gains come along algorithmically.
[0:09:02] The Transformer is like a good recent example of that. That's rare, but that is the biggest gain when it happens. So right now, there's an arms race, it feels like, for processing power, right? There's an arms race for fundraising, for processing power. There's an arms race to build and get more compute. You have a relationship with Microsoft, and you rely on them right now. I think...
[0:09:29] There's a race on all of these things. The compute one is like the most fun, the most theatrical, the biggest numbers. It gets the most attention. It is really important. I don't want to dismiss it. But there's certainly also huge amounts of effort going into who can come up with the best algorithmic ideas, who can secure new data sources. So I think it is on all three fronts.
[0:09:50] You do have this relationship with Microsoft. It has been described as, at times, it's been described as a tech bromance in the best way possible. And then I was just reading Alex Heath, who wrote, it's clear that Microsoft and OpenAI are disentangling. I don't think we're disentangling. Look, I will not pretend that there are no misalignments or challenges. Obviously, there are some. But on the whole, I think it's been a tremendously positive thing for both companies and excited to do much, much more together.
[0:10:20] But long term, do you think you need to have your own processing power that's yours as opposed to relying on somebody else? No. And in fact, I mean, I think we need to ensure that we get enough compute of the kind we want that we can rely on and all of that. And there may be reasons we have like some very crazy ideas about things we'd like to build that are, you know, like high risk, high reward.
[0:10:52] We certainly don't need to have like OpenAI get really good at building computer, like massive scale data centers. In fact, one of the things, and maybe this is a consequence of kind of just like how I grew up, but I like grew up as one of those like startup founders of the generation where it was like, oh, before that period in time, everybody had to build their own hardware and you had these colos and data centers. And it was this thing you had to do. And then all of a sudden AWS changed the scene.
[0:11:22] And so I've always like thought that way. I'm open to us, you know, saying, okay, we are going to have to build stuff ourselves or maybe vertical integration turns out to be more important over time. But it's really nice to get to focus on like research and product and do what we're good at. But one of the reasons I ask is it creates this sort of friend, enemy, frenemy situation, I imagine, where you have your product, you know, I was actually using it to...
[0:11:51] Check out a couple of things just yesterday. How'd it do? Great. Happy to hear it. By the way, we're going to have Sundar here later. The search product is fascinating, what you're doing. It is really cool. I have to say, it is my favorite product we've launched. Thank you for saying that. It's my favorite product we've launched in a long time. I use it so much and I'm so happy with it. It has completely changed my usage of the internet. But where I was going with this is, you have this relationship where you have
[0:12:20] Microsoft that's using your product in co-pilot. You now have Apple using your product. And then you have your own version of a sort of native product that you sell commercially yourself. At some point, do the interests no longer become aligned? And I say that because there's been some reporting, including in the New York Times, that suggested there has been frustrations about how much processing power and access you're getting, different requirements of things and services that...
[0:12:49] Folks at OpenAI are supposed to use that are Microsoft products, and maybe they don't want to use. I'm just trying to understand all that. There's definitely, at various times, real compute crunches. Again, our product has scaled. Two years ago, we had almost no business. Now, we have more than 300 million weekly active users. Users are sending more than a billion messages a day to ChatGPT. We have a very large developer business. I think it's like 1.3 million US developers, many more in the world.
[0:13:18] We need lots of compute more than we projected. And that's just been like an unusual thing in the history of business to scale that quickly. And there's been tension on that. I have not heard people like upset about using Microsoft services though. You know, we have a big platform business. We have a big first party business. Many other companies manage both of those things. And we have like things that we're really good at. Microsoft has things they're really good at. Again, there's not no tension, but on the whole, like.
[0:13:49] I think our incentives are pretty aligned. Sounds like there's a little bit of tension. For sure. On the artificial general intelligence piece, though, because you always said part of your deal with them is that if you ever get there, right, that then the deal could technically be called off. It sounds like you might be getting close. We've also said that our intention is to treat AGI as a mile marker along the way. We have left ourselves some flexibility because we don't know what will happen.
[0:14:21] My guess is we will hit AGI sooner than most people in the world think, and it will matter much less. And a lot of the safety concerns that we and others expressed actually don't come at the AGI moment. It's like AGI can get built. The world goes on mostly the same way. The economy moves faster. Things grow faster. But then there is a long continuation from sort of what we call AGI to what we call superintelligence. And that's when we should get word. I mean, I remember when you...
[0:14:48] Even in the past couple of years, you were in Washington and other places telling us that we should be nervous. Well, so first of all, even at the AGI moment, I think there are things to be nervous about. I expect the economic disruption to take a little longer than people think because there's a lot of inertia in society, but then to be more intense than people think. So the first couple of years, maybe not that much changes. And then maybe a lot changes in the economy. I don't...
[0:15:17] I'm not a believer in the no work. I think we'll always find things to do. And with every major technological revolution, there's a lot of job turnover. But I would bet we will have never seen it go this fast. So I think there's things like that to worry about in the relatively near term. But not these super dangerous things. That stuff, I think, is further out than a system that can, you know... Do you have any faith that the government or somebody is going to figure out how to avoid that? I have faith that...
[0:15:47] Researchers will figure out how to avoid that. I mean, there are going to be some... Let me caveat that a little bit. I think there's a set of technical problems that the smartest people in the world are going to work on. And I'm a little bit too optimistic by nature, but I assume that they're going to figure that out. And we're working super hard. Others are working super hard. I think there's a lot of work in front of us, but I also think we have this magic... Not magic. We have this incredible...
[0:16:18] Piece of science called deep learning that can help us solve these very hard problems. I assume we'll get that right. The societal issues around not just AGI, not the thing that we think will come in a few years that can do a bunch of jobs and create a lot of economic value, but like true super intelligence. The system that is not just smarter than you or smarter than me, but smarter than all of us put together. Just unbelievable capability. Even if we can make that technically safe, which I assume we'll figure out.
[0:16:48] We are going to have to have some faith in our governments. There are going to have to be some policy issues around that. There is going to have to be global coordination to a degree that I assume will rise to the occasion, but seems challenging. Okay, so here's the safety question. And we've all read the headlines. You've been in the headlines over the past year or two years. There have been so many people, both inside OpenAI, who've either left or spoken out or other things where they say, these guys are not focused on safety enough, right? You have to trust questions.
[0:17:19] Just speak to it so we understand. When people say that there's not enough focus on safety, you think what? What is that? What is not happening that is supposed to be happening? Well, I would point to our track record. I would say, you know, we've talked about we got this like big product out there. This is a very new technology moving very quickly. It started off as something that we did not know how we were going to align very well.
[0:17:48] It is now generally considered by most of society to be acceptably safe and acceptably robust. Safety, that's always like a contract negotiated by a whole bunch of stakeholders. It's maybe hard to exactly define what it means for Child GPT to be safe. But the reason I ask is when we hear different people who have left the organization and they've gone public, they take to Twitter and they say, these folks are not focused on safety. Is that about...
[0:18:14] Is that about processing power? Is that about your attention? There are definitely people who think that ChatGPT is not sufficiently safe, that it allows things that it shouldn't or doesn't do things that it should. There are people who will push on parts of that. There are people who will say, you know, fine, ChatGPT, maybe that's like a little bit safer than we thought it was going to be. But what's the plan for the next system? What are you going to do about that? And also there are people who would say like,
[0:18:42] Putting out a system like this at all is unsafe because OpenAI accelerated a race in the world. And that gives us less time to work on safety. We believe, and this is an opinionated stance, that this idea of iterative deployment is really important. We've got to put these systems out into the world. Society and the technology have to co-evolve. You have to start while the stakes are lower. You have to understand how people are going to use this and what it doesn't work for, what it does.
[0:19:11] There are other people who say, you know, sure, there's some benefits there, but it's not worth the costs. We'll be right back.
[0:19:37] Your sausage McMuffin with egg didn't change. Your receipt did. The sausage McMuffin with egg extra value meal includes a hash brown and a small coffee for just $5. Only at McDonald's for a limited time. Price and participation may vary. Promotion pricing may be lower than meal pricing.
[0:19:52] You know, Kevin, sometimes I feel like computers are the only thing around here that are getting smarter. Yeah, my model hasn't had an update in a while. But that changes now. We think that we can smarten these listeners up about all things related to this crazy new future. Welcome to Machine Gods, a new NPR podcast that will be your guide to the AI future. Twice a week, we're going to help you get smarter about AI so you know what's real, what's hype, and what you should probably install on your phone. Listen to Machine Gods from NPR on the NPR app.
[0:20:21] And maybe it goes back to the issue of processing power. And I was going to mention Elon Musk is building his own processing power labs effectively to power XAI. And one of the reasons I was asking whether you needed your own is because it seems like everybody is getting their own. We were just talking before to Ken about...
[0:20:55] Elon Musk. Elon Musk has a sort of unusual relationship in your life over the years, both as a founder and somebody who's suing the company now. How much do you see XAI as a competitor? We talk about Google, obviously, all the time. And we talk about Anthropic and, to some degree, Microsoft. And I don't know where you think Amazon is. We'll talk to Jeff Bezos this afternoon. But how do you see that? They're a competitor.
[0:21:22] Do you think of them as one of your biggest competitors? It appears that they are coming on super strong. Yeah, I assume they'll be a really serious competitor. Did you expect that? Yes. You did? Yes. And what are they doing? And how have they been able to literally spin up a company that feels, I'm not claiming it's comparable to what you're doing, but getting super close super fast? Meaning, what is that? Well, a lot of the models of the frontier are all pretty close now.
[0:21:50] But certainly tremendous respect for how quickly they built that data center. Is that, by the way, do you fear that all this becomes commoditized as a result? I mean, if it seems like, I'm not saying anybody can do it, but it just seems like, what is your advantage? And by the way, what is your disadvantage? I think the technology itself, to the degree you believe deep learning is just like, sort of like a law of physics, and that we discovered an important new piece of science.
[0:22:21] Then, yes, that part will be done by lots of people. Everybody has their analogy for what AI is like. You know, Sasundar is going to be here. He talks about it like electricity. A whole bunch of others. People talk about the Industrial Revolution. Some people talk about the Renaissance. The one I like is the transistor. There was a thing, a scientific discovery, that a few companies discovered first that transformed our society.
[0:22:51] That scaled unbelievably well. When people talk about scaling laws, the best analogy for that, I think, is Moore's Law. And that came to be used by a lot of companies in all of the world. Like, look at all the things in this room with transistors. But we don't think about that as a transistor device, and we don't think about Google as a transistor company, even though they wouldn't be here without one. And I kind of think that's what's going to happen with AI. There will be shockingly capable models.
[0:23:20] Widely available, used for everything. And it will be inconceivable to people in the future that the devices, the products and services they use are not really smart. And they won't all call themselves AI companies or AI products. And in that sense, in some sense, the AI itself, the kind of the reasoning engine nature of it will become commoditized. And that's fine. That's good. Like science should be.
[0:23:50] Just diffused throughout society. And that's why we're focused on building things like ChaiGPT. One of the other questions, and it is Elon related, is you are now moving towards having a for-profit component part of your business. Right now, it's a not-for-profit. And he, as everybody knows, if you read the papers, is suing you over this and many other things. I actually just want to ask you on a very personal level.
[0:24:20] Forgetting about even the merits of the case, as a guy who has spent a lot of time with him and you guys founded this thing together, how do you feel personally today about all of this? This is tremendously sad. I grew up with Elon as like a mega hero. I thought what Elon was doing was absolutely incredible for the world. And I'm still, of course, I mean, I have different feelings about him now, but I'm still glad he exists. And not just because...
[0:24:50] No, I mean that genuinely. Not just because I think his companies are awesome, which I do think, but because I think he, at a time when most of the world was not thinking very ambitiously, he pushed a lot of people, me included, to think much more ambitiously. And I'm grateful is the wrong kind of word, but I'm thankful. I'm positive about that.
[0:25:22] We started OpenAI together and then at some point he totally lost faith in OpenAI and decided to go his own way. And that's fine too, but I think of Elon as a builder and someone who, like, you know, known thing about Elon, he really cares about being the guy. But I think of him as someone that if he's not, that just competes in the market and in technology and whatever else and doesn't resort to lawfare.
[0:25:53] And, you know, whatever the stated complaint is, what I believe is he's just like he's a competitor and we're doing well. And that's sad to see. I got to ask you this. There was an article in The Wall Street Journal. You were in it. Jeff Bezos was in it. Some of the other people who are his main competitors were in it. And there was speculation in the article that folks like yourself were worried about his influence, not just his influence in.
[0:26:20] The technology area, but given his close relationship now with the president-elect. I'm actually not. I believe that pretty strongly, I may turn out to be wrong, but I believe pretty strongly that Elon will do the right thing and that Americans, it would be profoundly un-American to use political power to the degree that Elon has it to hurt your competitors and advantage your own businesses.
[0:26:51] And I don't think people would tolerate that. I don't think Elon would do it. It would go, again, lots of things not to like about him, but it would go so deeply against the values I believe he holds very dear to himself that I'm not that worried about it. Okay, one last related question on that, and then I actually want to talk a different tech question about the company, which is one of the things that he alleged is that he thinks that you guys are now so big that there's not competition and that you're trying to prevent competition from happening.
[0:27:21] By preventing potential funders of, for example, XAI or arguably others from both funding you and funding them at the same time. Incorrect. What we said, which is very standard, and what we say is if you invest in us and want to invest in any of our competitors, that's totally fine. We stop your information rights. You can still do it. But like is a very standard term for other companies of our scale. You know, we're going to stop like telling you our research roadmap.
[0:27:49] And everybody is like, that seems reasonable. But again, it's like, you want to go invest in them? Go invest. Let me ask you a different question just about the company. You did start it as a not-for-profit. It was a research outfit originally. Now, I think there's a view that it has to shift to be some kind of profit-oriented company. Does it have to? First of all, I like to talk about the reason we started as a non-profit because I think it sort of speaks to the answer, too. When we started, we did not...
[0:28:18] Have any idea that we were going to be a product company? We also did not know that the amount of capital we needed would turn out to be so huge. If we did know those things, we would have picked a different structure. It's hard to overstate, because it wasn't that long ago on the calendar, but it's hard to overstate how different 2016 was. This was years before we had the research that led to language models. This was four and a half years before we launched our first product.
[0:28:48] This was like six and a half years before we launched ChatGPT. All we knew that we wanted to do was do some AI research and that we thought AGI and super intelligence eventually would be this important thing to the world. And we wanted to like somehow do something that would be good. At the time, we were working on like writing papers, new RL algorithms, new theories, a way to play video games, a robotic hand.
[0:29:17] And it was not clear that there would ever be a product or revenue stream at all. And it wasn't clear that we were going to need one because it wasn't clear we needed so much money. It became clear after GPT-1 and some other work that we were going to need to scale. And we started, and also Elon decided to stop funding us as a nonprofit, and we couldn't get somebody else to do it, that we needed to find a way to make a capped profit. Now, we wanted to keep going with a lot of the things that we think are good about a nonprofit.
[0:29:46] And so we had this subsidiary, Cap the Profit, that worked for a while. And in some senses, still does kind of work. And in some other senses, is like straining the theory of what a nonprofit-controlled org can be and the amount of capital we need at this next stage. So we are, and have been for a while, looking at some changes. Nothing is decided. It is, as you can imagine, very complicated to figure this out. And the board is hard at work on it.
[0:30:17] In any configuration, like the nonprofit doesn't go away. Like one thing that the board has looked at, for example, is a PBC that a nonprofit owns a huge chunk of and then figures out how to use that amount of wealth for the purpose of the nonprofit. There's other ideas too. Yeah. One of the things I've always been fascinated by is you never took equity in the company. I think you think about that more than I do. I'm surprised. You get paid $76,000, I looked. And that's it.
[0:30:46] If the company does have one of these moments, there's an expectation you will get some equity. There was much of press about that. And there's been like some investor pressure. Look, it is weird that I didn't get equity. Do you want it? You didn't want it. No, I didn't want it. If I could go back in time, I would have taken it just some little bit just to never have to answer this question. And no matter how many times I try to explain to people, like, I am...
[0:31:19] I have the most interesting, coolest job in the world. This is like my retirement dream way to spend my time after like what was a pretty good career. And people can work on like art projects and not get paid for that. And no one thinks it's weird or whatever. It just does not come across. So I wish I had taken some. I don't imagine I would work any harder or less hard. There would be like, I think something clearer about the alignment.
[0:31:47] Would have had with investors or whatever else, and it definitely would have been easier to raise money. There are plenty of investors who have not invested because I didn't take equity. Really? That's interesting. They will not invest because you don't have equity. It's come up a few times, yeah. So now the company is $157 billion. What do you think is a good number for yourself if it happens? Again, I think you think about it more. We're going to run out of time, so I have to ask a couple other quick questions.
[0:32:13] One is, and at full disclosure, we should say, it just so happens. Can I say one more thing on that point before you go? I know I'll be quick. This is my childhood dream job. Not every day, not day to day. I would rather not be beating my head against the wall all the time. But getting to work on AGI and getting to sit in the room with the smartest researchers in the world and go on this crazy adventure, that is what I always wanted to do. That was my literal childhood dream. And that is...
[0:32:44] All of the weirdness aside, I think it should at least be understandable that that is worth more to me than any additional money. Fair enough. Thank you. Let me ask you this, and I would be remiss if I didn't mention it. It just so happens that the New York Times company has a lawsuit against OpenAI and Microsoft over training on content. And there's a lot of content creators and other folks in this room who make their living off of content.
[0:33:18] And I'm just so curious, and I know you can't speak to the lawsuit itself, but for those folks who are in this room who've written books or written articles or made movies or what have you, that have had their information trained on, whether it was on the open web or the closed web or on YouTube or what have you, how we should feel. I think we do need a new...
[0:33:48] Deal standard protocol, whatever you want to call it, for how creators are going to get rewarded. I very much believe in the right to learn or whatever you want to call it. And if an AI reads something, a physics textbook and learn physics, it can use that for other things like a human can. I think those parts of copyright law and fair use really need to keep applying. But I think there's additional things that we're starting to explore and others are where, you know, a particular passion of mine has always been.
[0:34:18] Can we figure out how to do micropayments where if you generate a story in the style of Andrew Ross Sorkin, you can opt into that for your name and likeness and style to be used and get paid for it? There's many other ideas, too. I think the discussion on fair use or not is at the wrong level. Of course, we very much believe you need one of these right-to-learn approaches. But the part I really agree with is we need to find new economic models where creators can have new revenue streams.
[0:34:48] On the New York Times, look, I don't believe in showing up in someone else's house as a guest and being rude. But I will say, I think the New York Times is on the wrong side of history in many ways. We can discuss and debate that. And we'll do that, I think, in court. Look forward to seeing you. My final question for you actually has almost nothing to do with technology, but maybe it does, which is you have a very exciting thing that's happening in your life.
[0:35:16] Personally, in the next year, would you have your child? Yeah. And for all parents out there and would-be parents who are thinking about AI and what this is going to do to our life and even what our own meaning is as people, if the machine actually can do things that we really can't, what we're supposed to do, how we're supposed to think about it, what it does to our own dignity, what do you think if...
[0:35:45] When you have your child and you're thinking and talking to that child, what do you think you're going to be telling them about this new world we're about to enter? There's nothing like, I mean, I'm sure having a kid will be even stronger, but just even like really like getting ready to have a kid. There's nothing like, at least for me, there's been nothing that has made AGI seem so irrelevant. And like the level of excitement I have about AGI, I thought it was pretty high, but it's so much.
[0:36:15] It really puts into perspective what actually matters. And I think that's a specific lens on the Gemma phenomenon, which is we have been developing incredible new technology for a long time. And each time it's happened, people have had these conversations. What does this mean? The Industrial Revolution comes along. Some machines take all of our jobs. What does this mean? Computer revolution comes along. Computers take a bunch of current jobs. What does it mean? And the answer?
[0:36:45] At least in terms of what it means to be human, it's not very much. The economy will grow. The kinds of jobs people will do will change. And people will care way more and love their kids way more than they care about AGI and anything else that any technology can deliver. The deep human drives are so powerful and have been here for so long. Evolutionary drift is pretty slow.
[0:37:16] I think in some sense, my kids will grow up in a super different world. And in some other sense, it'll be exactly the same. Sam Altman, everybody. Thank you so, so much for the conversation. Thank you. Dealbook Summit is a production of the New York Times. This episode was produced by Evan Roberts and edited by Sarah Kessler. Mixing by Kelly Piclo. Original music by Daniel Powell. The rest of the Dealbook events team includes Julie Zahn, Hillary Kuhn, Angela Austin, Haley Hess.
[0:37:49] Dana Prokowski, Matt Kaiser, and Yen-Wei Liu. Special thanks to Sam Dolnick, Nina Lassam, Ravi Mattu, Beth Weinstein, Kate Carrington, and Melissa Tripoli. Thanks for listening. Talk to you next time.