The Exchanges

Q pretty in the ballpark. I guess the question is, you know, you guys released for a while ago, you're working on five or, you know, more upgrades to four. I mean, I think to Jamal's point about Devon, how do you stay ahead of open source? I mean, it's just, that's just like a very hard thing to do in general, right? I mean, how do you think about that?

A What we're trying to do is not make the sort of smartest Set of weights that we can, but we're trying to make is like this useful intelligence layer for people to use. And a model is part of that. I think we will stay pretty far ahead of, I hope we'll stay pretty far ahead of the rest of the world on that. Um, but There's a lot of other work around the whole system. That's not just that, you know, the, the model weights and we'll have to build up enduring value the old fashioned way, like any other business does. We'll have to figure out a great product and reasons to stick with it and, uh, you know, deliver it at a great price.

AI assessment note: “we'll have to build up enduring value the old fashioned way, like any other business”

Q these are not services that are meant to be pipes that are just providing a set of APIs to a smart set of agents that ubiquitously work on behalf of eight billion people. What do you think has to change in how we think about how apps need to work, of how this entire infrastructure of experiences need to work in a world where you're agentically interfacing to the world?

A I'm actually very interested in designing a world that is equally usable by humans and by AIs. So I, I, I, I like the interpretability of that. I like the smoothness of the handoffs. I like the ability that we can provide feedback or whatever. So, you know, DoorDash could just expose some API to my future AI assistant, and they could go put the order in whatever. Or I could say like, I could be holding my phone and I could say, okay, AI assistant, like you put in this order on DoorDash, please. And I could like watch the app open and see the thing clicking around and I could say, hey, no, not this or like, um, there, there's something about designing a world that is usable, Equally well by humans and AIs that I think is an interesting concept. Same reason I'm like more excited about humanoid robots than sort of robots of like very other shapes. The world is very much designed for humans, and I think we should absolutely keep it that way. And a shared interface is nice.

AI assessment note: “I'm actually very interested in designing a world that is equally usable by humans and by AIs.”

Q Yeah, why'd you start, why'd you start it? Maybe just explain UBI and why you started it.

A Um, so we started thinking about this in 2016, uh, kind of, about the same time, started taking AI really seriously, and the theory was that the magnitude of the change that may come to society and jobs and the economy and sort of in some deeper sense than that, like what the social contract looks like, um, Meant that we should have many studies to study many ideas about new ways to arrange that. I also think that You know, I'm not like a super fan of how the government has handled most policies designed to help poor people. And I kind of believe that if you could just give people money, they would make good decisions and the market would do its thing. And, you know, I'm very much in favor of lifting up the floor and reducing, eliminating poverty. Um, but I'm interested in better ways to do that than what we have tried for the existing social safety net and, and kind of the way things have been handled. And I think giving people money is not gonna go solve all problems. It's certainly not gonna make people happy, but it might, it might solve some problems, and it might give people a better horizon with which to help themselves, and I'm interested in that. I, I think that Now that we see some of the ways, so, 2016 was a very long time ago, ah, you know, now that we see some of the ways that AI is developing, I wonder if there's better things to do than the traditional, um, conce…

AI assessment note: “so we started thinking about this in 2016, uh, kind of, about the same time”

Q So, so I, I think it's like, what are the apps that could be unlocked if AI was sort of ubiquitous on people's phones? Do you have a sense of that or what would you want to see built?

A Uh, I, I think what I want is just this always on, like, super low friction thing where I can Either by voice or by text or ideally like some other, it just kind of knows what I want, have this like constant thing helping me throughout my day that's got like as much context as possible. It's like the world's greatest assistant. Um, and it's just this like thing working to make me better and better. Uh, there's, there's like a, and when you hear people like talk about the AI future, they're imagine, they imagine there's sort of two Different approaches, and they don't sound that different, but I think they're, like, very different for how we'll design the system in practice. There's the, I want an extension of myself. Um, I want, like, a ghost, or an alter ego, or this thing that really, like, is me, is acting on my behalf, is, um, responding to emails, not even telling me about it, is, is sort of, like, It, it becomes more me and is me. And then there's this other thing which is like, I want a great senior employee. It may get to know me very well. I may delegate it. You know, you can like have access to my email and I'll tell you the constraints, but, but I think of it as this like separate entity. And I personally like the separate entity approach better and think that's where we're going to head. Um, and so in that sense, The thing is not you, but it's, it's like a always av…

AI assessment note: “what I want is just this always on, like, super low friction thing”

Q Got it. What's your preferred model of organizing OpenAI? Is it Sort of like the move fast, break things, highly distributed small teams, or is it more of this organized effort where you need to plan because you want to prevent some of these edge cases?

A Um, oh, I have to go in a minute. Uh, it's not because it's not to prevent the edge case that we need to be more organized, but it is that these systems are so complicated and concentrating bets are so important. Like one You know, at the time, before it was, like, obvious to do this, you have, like, DeepMind or whatever has all these different teams doing all these different things, and they're spreading their bets out, and you had OpenAI say, we're gonna, like, basically put the whole company and work together to make GPT-IV. And that was, like, unimaginable for how to run an AI research lab, but it is, I think what works at a minimum, it's what works for us. So not because we're trying to prevent edge cases, but because we want to concentrate resources and do these like big, hard, complicated things. We do have a lot of coordination on what we work on.

AI assessment note: “We do have a lot of coordination on what we work on.”

Q is really interesting. A lot of people would say that if we create a GI, that would be like an unintended consequence of something gone. Horribly wrong. And they're very afraid of that outcome, but open AI makes that the actual mission. Does that create like more fear about what you're doing? I mean, I understand it can create motivation too, but how do you reconcile that? I guess why?

A I think a lot of, I think a lot of the, well, I mean, first I'll say that I'll answer the first question in the second one. I think it does create a great deal of fear. Uh, I think a lot of the world is understandably Very afraid of AGI or very afraid of even current AI and very excited about it and even more afraid and even more excited about where it's going. Um, and we We wrestle with that, but like, I think it is unavoidable that this is going to happen. I also think it's going to be tremendously beneficial, but we do have to navigate how to get there in a reasonable way. And like a lot of stuff is going to change and changes, you know, pretty, pretty uncomfortable for people. So there's a lot of pieces that we got to get right.

AI assessment note: “I think it does create a great deal of fear.”

Q startup, and on the open web, and then quickly The accuracy or the value of these models will probably shift to these proprietary sources of training data that you could get that others can't or others can get that you can't. Is that how you see this thing evolving where the open web gets everybody to a certain threshold and then it's just an arms race for data beyond that?

A Doesn't, so I, I definitely don't think it'll be an arms race for data because when the models get smart enough at some point, it shouldn't be about more data, at least not for training. It may matter data to make it useful. Um, look, the, the one thing that I have learned most throughout all of this is that, uh, it's hard to make confidence statements a couple of years in the future about where this is all going to go. And so I don't want to try now. I, I will say that I, I expect lots of very capable models in the world and You know, like, it feels to me like we just like stumbled on a new fact of nature or science or whatever you want to call it, which is like, We can create, you can, like, I mean, I, I don't believe this literally, but it's like a spiritual point. You know, intelligence is just this emergent property of matter, and that's like a, that's like a rule of physics or something. Um, so people are gonna figure that out, but there will be all these different ways to design the systems. People will make different choices, figure out new ideas, and I'm sure, like, you know, Like any other industry, I would expect there to be multiple approaches and different people like different ones. You know, some people like iPhone, some people like an Android phone. I think there'll be some effect like that.

AI assessment note: “I, I definitely don't think it'll be an arms race for data because”

Q Do you think we talked about releasing it differently this time? Are you thinking maybe releasing it to the paid users first or, you know, a slower rollout to get the red teams tight? Since now there's so much at stake, you have so many customers actually paying and you've got everybody watching everything you do. You know, is it, is it, uh, you have to be more.

A Still only available to the paid users, but one of the things that we really want to do Is figure out how to make more advanced technology available to free users too. I think that's a super important part of our mission. Uh, and, and this idea that we build AI tools and make them super widely available. Free or, you know, not that expensive, whatever it is so that people can use them to go kind of invent the future rather than the magic AGI in the sky, inventing the future and showing it down upon us. Uh, that seems like a much better path. It seems like more inspiring path. I also think it's where things are actually heading. So it makes me sad that we have not figured out how to make GPT four level technology available to free users. It's something we really want to do.

AI assessment note: “Still only available to the paid users, but one of the things that we really want to do”

Q music model being trained on the whole corpus of music that humans have created without royalties being paid to the artists that, um, that music is being fed in? And then you're not allowed to ask, you know, artist specific prompts. You could just say, hey, pay me a, play me a really cool pop song that's fairly modern about heartbreak. Uh, you know, with a female voice, you know.

A We, we have currently made the decision not to do music, and partly because exactly these questions of where you draw the lines, and you know what, like, even I was meeting with several musicians I really admire recently. I was just trying to, like, talk about some of these edge cases, but even the world in which if we Went and let's say we paid 10,000 musicians to create a bunch of music just to make a great training set where the music model could learn everything about strong song structure, um, and what makes a good catchy beat and everything else. Um, and only trained on that. Let's say we could still make a great music model, which maybe, maybe we could. Um, you know, I was kind of like posing that as a thought experiment to musicians and they're like, well, I can't object to that on any principle basis at that point. Um, and yet there's still something I don't like about it. Now that's not a reason not to do it, um, necessarily, but it is. Did you see that ad that Apple put out? Maybe it was yesterday or something of like squishing all of human creativity down into one really thin iPad.

AI assessment note: “We, we have currently made the decision not to do music”

Q you had some fear about that. I think is that accurate? Cause the cynical side is like, well, this is a capitalistic move. And then the, I think. You know, I'm, I'm curious what the decision was here in terms of going from open. We, the world needs to see this. It's really important to close. Only we can see it. Well, how did you come to that, that conclusion?

A What were the discussions like? Part of the reason that we released ChatGPT was we want the world to see this. And we've been trying to tell people that AI is really important. And if you go back to like, uh, October of, not that many people thought AI was going to be that important or that it was really happening. Um, and A huge part of what we try to do is put the technology in the hands of people. Now, again, there's different ways to do that, and I think there really is an important role to just say, like, here's the weights, have at it, but the fact that we have so many people using a free version of ChatGPT that we don't, you know, we don't run ads on, we don't try to, like, make money on, we just put out there because we want people to have these tools, um, I think has done a lot to Provide a lot of value and, you know, teach people how to fish, but also to get the world, um, really thoughtful about what's happening here. Now, we still don't have all the answers and, uh, we're fumbling our way through this like everybody else, and I assume we'll change strategy many more times as we learn new things. You know, when we started OpenAI, we had really no idea about how things were going to go, um, that we'd make a language model that we'd ever make a product. We started off just, I remember very clearly that first day where we're like, well, Now we're all here. That was, you…

AI assessment note: “Part of the reason that we released ChatGPT was we want the world to see this”

Q ago that looked incredible. And then some kid just published it under an open MIT license, like open Devon. And it's incredibly good and almost as good as that other thing that was closed source. So maybe we can just start with that, which is tell me about the business decision to keep these models closed source. And where do you see things going in the next couple of years?

A So on the first part of your question, um, speed and cost, those are hugely important to us. And I don't want to, like, give a timeline on when we can bring them down a lot because research is hard, but I am confident we'll be able to. Um, we want to, like, cut the latency super dramatically. We want to cut the cost really, really dramatically. Um, and I believe that will happen. We're still so early in the development of the science and understanding how this works. Plus, we have all the engineering tailwinds. So, I, I don't know, like, when we get to intelligence too cheap to meter and so fast that it feels instantaneous to us and everything else, but I do believe we can get there for, you know, a pretty high level of intelligence, and, um, I, it's important to us, it's clearly important to users, and it'll unlock a lot of stuff. On the sort of open source, closed source thing, I think there's great roles for both, I think, um, You know, we've open sourced some stuff, we'll open source more stuff in the future, but really like our mission is to build towards AGI and to figure out how to broadly distribute its benefits. We have a strategy for that. Seems to be resonating with a lot of people. It obviously isn't for everyone and there's like a big ecosystem and there will also be open source models and people who build that way. Um, one area that I'm particularly interested per…

AI assessment note: “On the sort of open source, closed source thing, I think there's great roles”

Q Okay, sure. I mean, I think the whole industry is waiting with bated breath for the release of GPT-V. I guess it's been reported that it's launching sometime this summer, but that's a Pretty big window. Can you narrow that down? I guess where, where are you in the release of GPT-V?

A Uh, we, we take our time on Releases of major new models, and I don't think we, ah, I think it will be great, ah, when we do it, and I think we'll be thoughtful about how we do it, ah, like we may release it in a different way than we've released previous models. Um, also, I don't even know if we'll call it GPT-V. Um, what I, what I will say is, you know, a lot of people have noticed how much better GPT-V has gotten, um, since we've released it, and particularly over the last few months. I think I think that's like a better hint of what the world looks like, where it's not the like one, two, three, four, five, six, seven, but you, you just, you use an AI system, and the whole system just gets better and better fairly continuously. Um, I think that's like both a better technological direction. I think that's like easier for society to adapt to. Um, but But I assume that's where we'll head.

AI assessment note: “we take our time on Releases of major new models”