The Exchanges

Q not sort of become, uh, what you, you envisioned, but it did, it did bring you into this world that would become a really big part of your life. Y Combinator. Yes. You were in the sort of inaugural class of Y Combinator. Um, Paul Graham, famously one of the main co-founders. Um, I guess you met him and, and he liked the idea and, but what's the story? Yeah.

A Um, I, so I was like, you know, interested in this idea, playing around with it. Uh, a guy that lived in my freshman dorm named Blake Ross hosted something, I believe on Facebook, but memory could be wrong there now. That there was this new thing, uh, and it was called the Summer Founders Program at the time, and, you know, people should check it out. The deadline was, like, the next day or something. I had followed Paul Graham online, was like a big fanboy. Uh, so I immediately applied with my co-founders, and they invited us out to interview. That was the first, like, meeting the YC, the four YC founders, Paul, Jessica, Trevor, and Robert, that was, like, the first time I felt like, alright, I have, like, finally found the kind of people I want to be around. I didn't really know about the startup thing. You know, you, I heard about it as like a kid during the dot-com boom, but I wasn't really paying attention. Um, I had met a few VCs on campus, but I could like tell I didn't like them. The world felt impenetrable. Um, but I met the, I was like, wow, these are my people and everything they say resonates. And even the other founders I met there interviewing, that was this immediate click for me of like, this is awesome. This is what I want to do. These are the people I want to be around. This is so cool.

AI assessment note: “called the Summer Founders Program at the time... so I immediately applied”

Q that can detect disease in humans, um, based on, you know, hundreds of thousands of data points. Um, it can mean a machine that can beat the best gamers or chess masters. I mean, all of these things, but, but of course, that's how we see it now. When you talk about artificial general intelligence, what do you imagine in your mind, what do you think it actually could do?

A The first thing that I think it'll be and where, where people will use something like, wow, that is, that is real intelligence of some sort. What I imagine in my head is a gigantic skyscraper full of computers, and we all text it like we would text a A coworker, a friend, or whatever, and increasingly powerfully, for whatever you want, it can answer it. So you can say, like, I need life advice, or I need, like, great medical advice, um, and it'll help you with that. Or, or if you're just trying to, like, search for something to find information, it can, it can help you with that. It can look at your email and say, you know, here are all the important emails that came in overnight. I've drafted a response to all of them. If these look good, hit send, and you'll be happy with all of them. You'll want to hit send. Um, already today, you can say, I'd like this art, please. Um, can you generate an image that looks like this for me? Um, can you write this code for me? Uh, eventually, now this will require more compute, but if we society at allocate to compute to that, you can say, like, could you go off and find a cure for cancer? And it can spend a lot of its compute cycles doing amazing scientific progress like that. But, but I imagine that we get to a world where Uh, each of us has sort of like a, what feels like an AGI companion that we are talking to all day that is helping us b…

AI assessment note: “What I imagine in my head is a gigantic skyscraper full of computers”

Q What was it about, about computers as a teenager that you remember feeling so passionate about? I mean, was it, was it a community of other people that you connected with locally or even beyond where you lived in St. Louis?

A That was certainly part of it. Um, I also was into like ham radio as a kid, and the one thing that I noticed across both of them was the fact that you could like, A, there was interesting engineering and science problems and technology to figure out. There's like the fun of using a computer at all, especially back then when like they didn't work that often, you had to like upgrade parts of them yourselves. That's just like a cool puzzle and technology is interesting. But definitely one thing that I felt across both of them is this ability to connect with anyone around the world instantaneously, which was incredibly appealing to me. The ability to sort of, like, learn anything, talk to anyone, was, was just, like, and still is, like, pretty magical. I think we've, we've all just now accepted that as standard, but if you rewind the clock, like, 30 years, things were very different.

AI assessment note: “That was certainly part of it... ability to connect with anyone around the world”

Q Why did you found it as a non-profit? I mean, initially, I know that you have a for-profit subsidiary now, but why did you originally make it a non-profit?

A I think this technology really does deserve to belong to the world as a whole. It's going to have such a profound impact on all of us that I think we deserve, like we globally, all the people, all, all of humanity deserve a say over how it's used, what happens, what the rules are. Um, we deserve to share in the benefits. We deserve to all have access to it. And that was why I I'm like very pro capitalism. Obviously I think capitalism is great. Um, but I think AGI is sort of an exception to that. In a well-run society, I, I think even if we rewound the US, you know, 50 plus years, this would absolutely happen by the government.

AI assessment note: “I think this technology really does deserve to belong to the world as a whole.”

Q if you're building an autonomous vehicle, you know what you're building towards, but you're really not building towards anything in particular. You're building towards something that we humans can't even imagine, right? And that's the thing, like, it's, I guess it's a matter of just figuring out how to develop technology that can think like a human beyond how a human thinks, can process information in highly complex ways. Yeah.

A Um, you, you, you bring up a great point. The other company I'm involved with is this nuclear fusion company called Helion. And there are a lot of similarities between the two companies in terms of like a very hard Scientific and engineering problem. But the biggest difference is like what to do, what success looks like, what's going to happen on the other side of this, what problem to go solve next. You can look at it backwards or forwards. It's very clear. It's all hard, but it's very clear. There's no uncertainty about what to go after, what's going to happen, what the company needs to do. Um, and another nice thing is like, it's basically all upside. There's, there's none of the harder issues to contend with. At OpenAI, we, we just don't have any of that certainty, and one of the things that has been, like, a little bit surprising to me is how difficult it is to get advice about how to run a project or a company in the face of such uncertainty. It hasn't happened a lot of times, and I am, like, wistfully envious of the, the Helion world, because it's, you know, it's still super important and super great, but it's so much clearer, and we just have to, like, turn over one card as we go.

AI assessment note: “At OpenAI, we, we just don't have any of that certainty”

Q Right. Because at the time, YC was focused mainly on, like, software, right? Yes. And you wanted, you wanted them to focus more on things like nuclear energy, self-driving cars, like, like actual physical things that you use.

A Not necessarily physical things, but I would just say, like, hard technology. Like, there's a thing about Silicon Valley, which is it's, it had not been that well geared towards the time and capital intensive science projects. Um, so I wanted to do that, and I kind of was just, like, thinking about, you know, I had, like, basically taken a year off. It had been really fun. I had, like, learned a lot about everything. I had met a ton of smart people. I had read, like, many, many dozens of textbooks. And there was, like, all of this stuff that I wanted to do. And, uh, I was like, you know what? In some sense, yes, being a VC is, like, an admission of failure. But in some other sense, what an incredible platform. You know, I could, like, do whatever I wanted here. Um, and, and I love YC so much.

AI assessment note: “Not necessarily physical things, but I would just say, like, hard technology.”

Q jobs will be taken over by, you know, artificial intelligence Uh, you know, devices or, or software, whatever it might be. But, um, this labor situation won't be around forever. There will be higher unemployment rates. And, and so what happens? I mean, do you, do you, I mean, if this technology is, is going to become so good, which it will be, what happens to the, to the humans?

A You know, there's a huge amount of pole clutching about what, what's going to happen when AI replaces all of the jobs and what we're going to do. And I have A few thoughts about that. Um, one, I actually think we're in a humongous labor shortage, like much more than it seems. There are many people who love their jobs and derive great meaning from it, you know, but I think those people don't recognize it's a privilege that most people don't have, and most people would love to work less or not or do something completely different, but we have sort of this, like, pay-to-play world right now where you have to pay to live, And I, I get why some people think that's important, sort of, in theory, but I personally really don't think that's the society we should strive for. Um, I think, like, traditional work, in the way we think of it, should be optional, and you should be able to do less of it. I really love work. It's my hobby. It's my passion. I, I think it's great. But I know a lot of people don't, and a lot of people would choose to spend their time in a very different way if they didn't have to. I also think a lot of people who like work would clearly like to work less. We're seeing this in this post-pandemic world where, you know, a lot of people are like, oh, I'd go into the office two or three days. I'd really rather only work two or three days, too. I have a lot of other thin…

AI assessment note: “traditional work, in the way we think of it, should be optional”

Q think about these things, and so you know that all of these wonderful things, you can answer questions for us, it could solve cancer maybe, it could, it could give us a diagnosis, it could identify the best doctor, that will happen, and then there's gonna be other things too. So let's talk about some of the other things, and, and what keeps you up at night about those things?

A You know, I've, I've had three recurring bad dreams over the course of my life. When I was in high school and college, I had this one about, like, missing a test. It sort of evolved when I started my startup to, like, thinking in the dream that I was, like, still in school and I was missing stuff because I was running the startup. Um, after the startup, I had this, like, recurring bad dream for a while just about, like, failing, which was just incredibly shameful and Very, very tough. Uh, you know, startups are great when they're working, but when you're failing, it's, like, quite brutal. And the current one is about, like, what if we're wrong about all of this? Um, I think there will be, like, good and bad with any new technology, and I can accept that as long as the good is order of magnitudes more than the bad. Um, and that is truly what I believe will be the case. I think the world is going to be incredibly different, but there's no stopping that now or ever. And I think we will be able to responsibly deploy current technologies. Uh, and I think we've, you know, one of the advantages of our nonprofit structure is we do a lot of things differently than a normal company would. But for me, the real question is, can we solve the existential issues? And I am more optimistic about that than I've ever been before. Um, I think a lot of the, the work we've done around alignment and,…

AI assessment note: “And the current one is about, like, what if we're wrong about all of this?”