The Winding Path of Progress | Sam Altman | Talks at Google

Jorge Cueto 0:00 [upbeat music]

Jorge Cueto 0:08 Hi, everyone. Welcome to our Talks at Google event. Uh, I'm gonna go through some logistics first. This is a fireside chat style talk, so we'll take about 40 minutes to talk through some questions, and then in the last 20 minutes, we'll open up to the audience to submit questions, um, both in person here through a mic in the back of the room, and also through the Dory at go/ask-sam. And I'd like to thank everyone who has helped out to make this talk possible, including our facilities team and everyone on the PM speaker series organizing team. So it's great to have Sam Altman here with us today for Talks at Google. Sam is the president of Y Combinator, which is widely regarded as one of the top startup incubators in Silicon Valley. He went to Stanford and studied computer science, and was the founder and CEO of a, a mobile location-based startup called Loopt, which was funded by YC as part of its first class of startups in 2005, and acquired by financial services company, uh, Green Dot in 2012. In 2014, he was named president of Y Combinator, and since then he's worked on a wide range of initiatives from YC Research, which is a nonprofit branch of Y Combinator that focuses on doing pure research around moonshot ideas like universal basic income. And he's also worked on OpenAI, which is a nonprofit AI research company looking into finding ways to create safe, friendly artificial intelligence that can actually help all of humanity. And it's great to have you here.

Sam Altman 1:48 Thanks for having me.

Jorge Cueto 1:49 So just as I mentioned, you're involved in a wide range of things from YC to OpenAI, and most recently you released the United Slate. So I wanted to ask you how you think about prioritizing the projects that you work on.

Sam Altman 2:04 Uh, you know, I think, uh, optimal time allocation is, is probably like an AI complete problem. Um-

Jorge Cueto 2:09 [laughs]

Sam Altman 2:10 ... I think if you can get to spending like 1% of your time perfectly, that's, that's really good. Uh, and so I think this idea of figuring out what to focus on and what not to focus on is, uh, both really hard, uh, and still significantly under-invested in. Um, the frameworks that I have used, the sort of two big frameworks that I've used, uh, to figure out how to allocate time, um, one is impact maximization/regret minimization. Um, so I try to look at those two curves together, and I try to think about where I can have the biggest net impact on the world, um, net positive impact on the world. Uh [laughs] easy to have a big impact. Um, and, uh, and then also just regret minimization. You know, you get, you get to live once. It's really important that you do what you wanna do, and that you spend time with the people you'd like to work with, and work on the things that you find personally fulfilling. And so if I think I'm gonna really regret doing something or regret not doing something, uh, even if I think it's not the best use of my time for pure sort of net impact on the world, um, I'm still willing to be like, to take that really seriously, and I think that makes me do better at the things I do that, that do help the world. Uh, so, you know, the broad things that I've learned that I like to do, um, one is teach people. Um, another is create economic growth. Uh, I really do believe that one of the things that is most fundamentally going wrong right now in the country is that we don't have enough economic growth, and the little that we do is not evenly distributed at all. Uh, and so I think in a democracy, um, you, you really want everyone's lives to get better every year. We're basically insensitive to the absolute quality of our lives, and extremely sensitive to relative differences year over year to our neighbors. And it's, it's really important that everyone's life is getting better constantly. Um, and I think economic growth is important to that. I think that, um, I think that AI is, um, gonna be the most important technological trend of our lifetime, so I sp- spend a lot of my time on that. Um, and, you know, I, I, I try to think about things on those two strategies. Uh, the other framework besides the sort of impact maximization, regret minimization that I found really useful is spend a little bit of effort trying a lot of things, and then relentlessly prune down and focus quickly on the ones that you like and the ones that seem to be working. So in some sense, this is the Y Combinator model, um, of fund a lot of startups with a little bit of money, um, and then, you know, most don't work out and some work really well, and you spend more time and more money on the ones that do. Um, and this is the, uh... Hey, guys. Um, this is the thing that I've tried to apply to, um, my life more generally, uh, is this idea that I can try a lot of things with a little bit of effort. It's very hard to predict exactly what's gonna work and what hasn't. Um, but then the, the hard part about that, and the thing that most people don't do is you really wanna relentlessly focus down on, on the ones that, that do work. Um, and the last thing I'll say about prioritization is the other hack I have learned is if you can get one really great partner with you on every project, um, that will cover up a lot of the slack, because if you try to do multiple things at once, crises come at the same times, and that's really hard if you have to do it all yourself.

Jorge Cueto 5:33 And focusing more specifically on productivity, like what are some life hacks that maybe you apply to make your everyday more productive?

Sam Altman 5:42 Um, I mean, I think there's like a lot of crap written about productivity secrets on the internet, and people sort of like get into this thing where they spend more time like trying to be productive about their productivity system than actually getting things done. Um, I, I will say, I think that... Well, I'll say two, I think, pieces of advice that aren't obvious. Um, one is I think far more important than any particular system is just figuring out the right things to work on. And so all of the time that people spend with, like, this new productivity app or that or whatever, would be better spent, like, really trying to think diligently about, I have the same number of hours as anybody else. What am I gonna spend them on? And getting that right is more important than exactly, like, being perfectly productive with those hours. Um, a big part of that is not doing things that waste time. Uh, I think if you can just, like, focus on the things that are important and not do the things that waste time, um, you can be fairly sloppy with productivity otherwise, and you'll still get far more done than most people. Um, that's... it's really hard to do, though. The other thing that I think people don't think about enough is figuring out your own personal, like, rhythms of productivity. Um, and there's huge variance I've noticed between people that figure this out and don't. Um, so, like, for me personally, it turns out that I am most productive if I go to sleep late, wake up late, and then keep the first, like, three or four hours of the day and don't schedule any meetings, like, work from home, like, get through my list of stuff then, and then, like, pack all my meetings when I'm, I'm kind of less productive at just grinding stuff out or thinking creatively in the afternoon. Um, and, like, it took me some number of years to figure that out, 'cause it didn't, like, fit well with the work schedule I was naturally in. But then I was like, "All right, if this is, like, the thing to, that makes me most productive, then I'm gonna, like, make my whole schedule work to support that." And that was, like, a really important change for me. Uh, so I think figuring out your own personal, like, optimal times to work on what kind of different things, uh, people don't really talk about that much and, at least for me, it had a huge impact.

Jorge Cueto 7:51 Shifting gears a little bit, but still trying to get your perspective on different things, uh, related to day-to-day. Uh, one of the key issues that has come up recently is bias in the workplace, and both conscious and unconscious bias. And I wanted to get your perspective on what are some of the strategies that you implement personally, whether in the way you interact with people that you encounter or in how you approach the decision-making process in using strategies to minimize your own bias.

Sam Altman 8:22 Look, I think this is an important conversation happening in Silicon Valley right now, and there's a lot of opinions on a lot of sides. But I would just like to say the following. I think, um, no matter what you believe about biology, no one that I respect does not believe that women and racial minorities and a number of other groups face an absolutely unfair playing field, um, their entire lives. I think people start getting told directly, indirectly from a very young age, um, "This is the kind of thing you should do or can do or whatever," and that has an effect, a compounding effect on anyone. And I think trying to counter that is really important. And I think, um, again, I think there are things people can, reasonable, well-meaning people can disagree on, and I think unfortunately we spend all of our time talking about the disagreements, and we don't focus enough on the agreements. And I think almost all smart, reasonable people will agree, um, that the society we grow up in has a hugely unfair pa- playing field. Um, and I think because of that, it is not enough to talk about unconscious bias. I think that is a real problem, to be clear. I think, um, we are all a product of the society we grow up in, and we all have biases that aren't our fault, but still a responsibility to counteract. Um, but one thing I don't like about the discussion in Silicon Valley about unconscious bias and how that's the problem that we need to fix is I think it is not nearly sufficient, and it ignores... It's a good thing to fix, but it ignores the fact that, you know, decades or centuries of society have built up a very uneven playing field. Um, and that is why we do need programs, um, to try to, uh, proactively counter that. So, um, I think it's, you know, I think it's really important we not lose sight of that and that unconscious bias training alone, although currently very fashionable, will not fix that. That said, um, I do believe unconscious bias is a problem. We try to counteract it, A, by talking about it and doing training, which I think does help. Um, but B, I think one of the things that we have done that unfortunately other investors have not done as much is just have a very diverse, um, partnership. You know, we have, um, six female GPs on our team. Um, I, and I'm f- that's, you know, probably a large part of the, um, women in top investment roles in Silicon Valley. Uh, and that's, that's really bad. We have, um, you know, the CEO of our core program is Black. Um, and I think, I, I hate to play the kind of like who's the most discriminated stock game, but I think Black entrepreneurs in Silicon Valley have an exceptionally hard time. Um, and I think by having a more diverse team, it helps us, um, have broader networks and also think about our own unconscious bias all the time.

Jorge Cueto 11:17 Mm-hmm. And the issue of politics or maybe even bringing these issues up in the workplace is some- sometimes seen as a taboo, and yet you've been pretty vocal about your views on current political events and also other issues that are coming up in Silicon Valley. So I wanted to ask you how you walk that fine line between expressing your opinions, but also minimizing any repercussions that that could have on the day-to-day business of YC.

Sam Altman 11:44 Well, now it's not even controversial. Um, you know, like, now all, all the tech CEOs are talking about politics. Um-

Speaker 11:51 [laughs]

Sam Altman 11:53 When I started doing it, um, a couple of years ago, uh, kind of at the beginning of the rise of Trump, it was controversial. Um, two, two things were going on. One is I think no one took me seriously. Um, so most people were like, "You know, this is a ridiculous thing. It's gonna go away." Two is that, uh- I think in normal times, it does make sense for business leaders of large organizations to remain apolitical. Um, you know, I think it actually does make, that makes a lot of sense. Uh, it's, it's a huge distraction, um, it's hugely time-consuming, and it has all of these like weird negative effects like, you know, if you piss off the president, he can do bad, bad things to you. Um, however, however, these are not normal times. Uh, and I think the, you know, like when the future of the republic is at risk, um, the duty to the country and our values transcends the, the duty to your particular company and your stock price. Uh, and I think I started that, uh, a little bit earlier than other people, but at this point, I'm in very good company, and it doesn't seem to be that controversial anymore.

Jorge Cueto 13:07 Mm-hmm. What are some things that you wish people knew about you that they don't know about you?

Sam Altman 13:15 Uh, you know, at this point I would just like to have like my own little quiet private life back.

Jorge Cueto 13:21 [laughs]

Sam Altman 13:21 I don't feel like, um, I'll say nothing at all.

Jorge Cueto 13:24 Okay.

Speaker 13:25 [laughs]

Jorge Cueto 13:25 Fair answer. Moving on to talking a little bit about YC and what you look for in companies. Do, are there any qualities in founders that you think are overrated or underrated?

Sam Altman 13:40 Yeah. I think the most underrated quality, um, the most underrated quality of all is being really determined. Um, this is more important than being smart. This is more important than having a network. This is more important than a great idea. Um, the hardest thing about starting a company is the level and the frequency of bad stuff that happens to you, and most people that are good in really other ways eventually just get killed, the company gets killed by, um, stuff going wrong. And, you know, so much about being a successful entrepreneur is just not giving up. Um, when we have funded people who have a great idea, perfect background on paper, and, um, you know, a great product and still failed, uh, it has usually been that they're insufficiently determined. So I think this is the most important non-obvious skill of a founder. Um, of course, you need a good product and a good market and to be smart, but that's really obvious. Um, the degree to which being like a, you know, three or four standard deviation outlier on determination is a required skill of a CEO, um, is not something that was obvious to me when I started. Uh, that's also bad because it's really hard to select. It's really hard to identify that. You know, as we have said more publicly how important that is, uh, people applying to YC have gotten better and better at telling us stories from their past life about how they overcame these impossible odds to get through something. Um, and, and unlike intelligence, which is very difficult to fake in a, you know, one-hour meeting, you can definitely fake determination in a one-hour meeting. Um, so that's one thing that really matters. Um, another thing that really matters that is non-obvious, uh, is independent thought. And, and this I think is an even more unusual skill than determination. Um, I think both of these you can make a conscious effort and build up, but I think independent thought is one of the hardest skills to, um, to, to build up a- because, like, we are all spewing of social pressures from birth. Um, we are all pushed to think like other people. Uh, and if you think in your own lives about the number of people you spend time with that you would say are true independent thinkers that consistently have new ideas you haven't heard from other people and think about the world in different ways, um, it's probably a very short list. And yet, these are the people that start all the interesting companies. The consensus ideas everyone tries, Google will eat your lunch at, um, and they're also not kind of the, the, the really big trends of the future. You want startup ideas. If you picture a Venn diagram, and here you have, um, is a good idea, and here you have sounds like a bad idea, um, you want that tiny little overlap. And, and those are the kind of ideas that are the hardest to identify, and the ones that, um, even if you do manage to notice them, most people will talk you out of. So I'd say those are two non-obvious skills that we look for.

Jorge Cueto 16:55 Mm-hmm. And what's one challenge that YC companies face repeatedly that you've noticed?

Sam Altman 17:02 Um, hiring engineers, uh, when Google can just sort of like throw unlimited cash at anyone they want-

Jorge Cueto 17:09 [laughs]

Sam Altman 17:10 ... has become very problematic for startups. Um, I, I think the good thing about that, I always try to like, I always try to find the good in any bad situation, and the good thing about this is it means that if a startup does not... A- all of the startup, all of the really good startups have really important missions. Um, eventually they figure it out. They may not have it on day one, but they eventually get to this like missionary mindset. And it used to be that even if you didn't have that, like you could just sort of get a bunch of mercenaries to come work for you, and now you can't 'cause you can't out-pay Google. And so, um, one positive side effect of this thing has been that startup, the, the importance now of a startup having a really clear and really important mission on day three has gone up a lot, and I think that's leading to better startups because otherwise you just can't recruit. Um, a- another common problem that startups have, and this doesn't sound like a great insight, uh, but most startups still don't ever build a product that people want. And it doesn't seem to matter how much we talk about this, it doesn't seem to matter how much anyone talks about this, people still keep trying to do anything but this.

Jorge Cueto 18:27 Yeah.

Sam Altman 18:27 And if there's like one thing that a startup has to get right, it's build a product that people really want. Not a little. If they want it a little bit, they kind of... You won't generate enough momentum. Like, you gotta build something that some people really love. And, um, after failing because of insufficiently determined founders, this is the number two reason that startups that seem really good otherwise fail.

Jorge Cueto 18:51 Moving on to talking about different technologies, what would you say is the biggest challenge that we're facing in terms of making progress on artificial intelligence right now?

Sam Altman 19:03 Well, I don't know if any of you saw this, but, um, OpenAI beat the best, um, 1V1... B- the best single player Dota players in the world, um, last Friday. And I, when we started that project at the beginning of this year, I did not think that was gonna happen this year, and I wasn't even sure it was gonna happen next year. Um, and it was very wild to watch that happen, because it was almost purely self-training. Um, and it was this very complex environment, um, and the AI was just playing itself and getting better and better and better. Um, in fact, the final bot that we had that beat all the humans handily, um, o- one day later, it lost 60-40 to a one day more evolved version of itself, just to give you a sense for the rate of improvement. Um, and, and so I think people are kind of asleep at the wheel here.

Jorge Cueto 20:01 Oh.

Sam Altman 20:01 Um, there are problems, but again, in the same spirit of always trying to figure out, like, the good, not the bad, um, we are making unbelievable progress, maybe too much progress. Um, but I think, I think of all the things I worry about with AI, the technical barriers to progress are not the top of the list.

Jorge Cueto 20:23 Mm-hmm. There's another camp that says that maybe we're placing too much emphasis on the threat of AI. Uh, what do you think about that?

Sam Altman 20:32 Yeah, I think we, we don't talk enough about the benefits. Um, I think that AI has the potential to, uh, eliminate, like, nearly all human suffering in the next couple of decades. I think we can have a world of abundance, we can eliminate poverty over time, we can probably cure a whole lot of diseases. Um, there are all these wonderful things that the technology can do. Uh, and I think we're already seeing that in just how much better a lot of consumer products we use every day have gotten. Um, people like disaster porn. People are more interested in talking about the end of the world than they are about life getting 10% better every year and having that compound, which gets a lot better. Um, so, you know, you can, if you're a journalist, you can write an article about how AI is going to end the world and get a lot of clicks, and, you know, a page view bonus or whatever you get. Or you can, like, write an article about how AI is, like, gradually making all these problems 10% better, and probably no one reads it. Certainly, no one shares it. And so I think, for whatever reason, the way we're wired is to talk much more about the downside of this than the upside. But the upside, I think, is gonna be huge. It already is huge.

Jorge Cueto 21:44 Mm-hmm. And what do you think about cryptocurrencies? Do you see-

Sam Altman 21:50 Um [laughs]

Jorge Cueto 21:50 ... sustained growth?

Sam Altman 21:52 Uh, yeah. I mean, I think, you know, it's gonna go up and down. Um, like I bought-

Jorge Cueto 21:58 [laughs]

Sam Altman 21:59 ... my Bitcoin a long time ago. I plan to hold them for a really long time. I try not to watch the price ticks, but it's so addicting that-

Jorge Cueto 22:05 Yeah

Sam Altman 22:05 ... I can't help myself. Um, I think... [laughs] This is not investment advice.

Jorge Cueto 22:14 [laughs]

Sam Altman 22:14 Um, I think, I think the only super compelling use, proven use case we have seen so far is store value, and as a replacement for gold, I think we are seeing real adoption there, and real collective belief that actually makes it have some value. Um, however, if that's how it's gonna play out, then I think, um, Bitcoin should dominate. Biggest network first, biggest brand, most collective belief, whatever. Uh, and so I have been surprised by the continual strength in all of the altcoins. Um, I think there are potential other truly valuable applications of the blockchain. Um, you know, Filecoin is a YC thing, so I know that one pretty well, and I can see that being big. But I think most of what's going on feels like a complete, outside of Bitcoin, feels like a complete speculative bubble, and I feel bad that a lot of people are going to get burned. Um, so I think probably the right thing to do, if you believe in it, um, is to, you know, buy Bitcoin and then not think about it for another five years.

Jorge Cueto 23:21 Mm-hmm. Is there any specific product area or technology that you think people are sleeping on?

Sam Altman 23:28 Um, sure, a lot. Um, AI, as we mentioned, I think people are asleep at the wheel in a really big way on. Um, nuclear fusion, I think is within some single-digit number of years of working. Uh, and because it's been so bad for so long, um, people are just kind of burned out and not really taking a new look at new materials and stronger magnets and better computer models that's gonna enable this to work. Um, synthetic biology, again, it's like, people talked about it a lot a couple of years ago. It didn't quite work as fast as people were hoping, so we have like this hype cycle, and then it really falls off, and now we're in the part where I think the interesting work is happening and, and people aren't paying enough attention. Um, I- I guess I'll s- the, the, yeah, that's a topic I can go on for a long time, but I'll limit it to three.

Jorge Cueto 24:23 Sure. Is there any particular problem that you think technology can't solve?

Sam Altman 24:29 How to make us nice to each other, uh, technology has clearly been quite bad at solving. Um, I won't say it can't, um, but I haven't seen it yet, and I certainly don't think technology can by itself. Um, and I think the... I think we are kind of in this situation, um, at least in the developed world, where the world keeps getting better, um, and we keep getting unhappier. And, you know, there's, like, a decent amount of data on this, and I think, um, I think technology is not entirely to blame, but it's certainly not blameless. And, um, I think... 30 minutes left, I couldn't even start the conversation, but I think the number of, like, things that technology has done to make us, um, more isolated, um, feeling relatively worse. I, a f- a friend of mine, um, a little bit older, said she never used to be unhappy because she had no idea what she was missing. And now that she has to watch, like, I, I'm very bad at pop culture, I'm gonna pick a name at random.

Jorge Cueto 25:37 [laughs]

Sam Altman 25:38 Kim Kardashian, like, fly around in private jets on Instagram all day, she's jealous, and it, like, makes her very unhappy, and it didn't used to happen.

Jorge Cueto 25:45 [laughs]

Sam Altman 25:46 Um, and I've been thinking about that a lot in the last couple of weeks, and I, I don't have a solution to that, but I understand why that's a problem. So I think, um, figuring out how to make us, like, happier and nicer to each other in particular, I think one thing that technology does that's really bad is, you know, we have some probably long-standing evolutionary pressure that even if I really don't like you, I'm very unlikely to come stand next to you and say, "I fucking hate you, you're a jerk." Like, you know, whatever else people say on Twitter.

Jorge Cueto 26:17 [laughs]

Speaker 26:17 [laughs]

Jorge Cueto 26:18 A lot of things.

Sam Altman 26:19 Because there's this, like, you know, we're like these pack animals, and we have to, like, live with each other and, you know, help each other survive. Um, but somehow, like, on Twitter, that goes away. And, um, and I think, like, I'll save my whole rant about Twitter, but I think, like, that is a platform in particular that rewards saying the most aggressive, snarky things. That's how you get likes and retweets and sort of value out of the platform. But a lot of technology does this. It, we don't have whatever kind of, like, the human be nice to each other in person instinct is, and we have these platforms that reward being bad, reward being a jerk. Uh, and, and so I think... I'm not optimistic that technology's going to solve that problem. I think that's gonna have to be people solving that problem.

Jorge Cueto 27:09 Mm-hmm. In an interview with Vanity Fair in 2015, you mentioned that you were optimistic about the future, and that was right before all of the election stuff happened and the situation that we find ourselves in. So would you say that you're still optimistic about the future?

Sam Altman 27:26 Yeah, of course, of course. Um, like, progress is not, it's, it's not a perfectly exponential curve. It's not, it's not even a straight line. Um, and you know, we are clearly in a challenging period now. Um, but I think if you look back at the last few hundred and then few thousand years, um, if you zoom out enough that the squiggles on the curve kind of disappear, um, the world is getting so much better, and I think that's gonna keep happening. Um, and I am, you know, even with everything going on right now, um, I'm delighted to be alive right now and not 100 years ago, and certainly not, you know, 200 or 2,000 years ago. Um, and I think we do have technology to thank for a lot, and we have better governance. Just, like, the number of people living in a democracy, uh, you know, 200 years ago was very low, the percentage. And as horrible as things are, the fact that we get to stand up, uh, and speak our minds without fear of, you know, being thrown in jail for political opposition, and the fact that we get to vote again in three and a half more years, um, I think that's amazing, and I think it's easy to take that for granted. But, uh, yeah, I think the future is gonna be a lot better. I remain very optimistic.

Jorge Cueto 28:47 What do you see as being a big challenge that we're facing as a society right now?

Sam Altman 28:52 Uh, how we deal with a world where the natural forces are for wealth to concentrate into the hands of a smaller and smaller number of people. As I mentioned earlier, um, I think people are more sensitive to relative quality of life than absolute quality of life, and I think technology is naturally a force, it's a giant lever that tends to create way more wealth, um, but really concentrate it. So I think one of the most tone-deaf things people say in Silicon Valley is, um, you know, like, poor people should be happy. They get this Android phone for not much money. They can access anything in the world. Um, and they wouldn't have that without us, so why are they complaining? And okay, like, there is some truth to that. I do think it is cool about the world that, like, the richest person and someone living in absolute poverty, um, carry around the same phone. That, that is something that would not... There was no analogy to that other than maybe, like, if you got a really bad disease. There was no, like, equality like that, you know, 400, 500 years ago. Um, however, that point, which is always what Silicon Valley falls back to, I think is, like, maybe not the most possible, most tone-deaf possible response, um, but up there. Like, people wanna feel like they have agency. They wanna feel like they have a voice in the future. They wanna feel like they can participate. Um, and they wanna feel like they're not just kind of, like, given this baseline while- You know, like they toil in the provinces, and Silicon Valley just gets all the money. Um, and I think people who don't see that are just not thinking clearly. Um, and so I think one of the greatest threats, something else I don't think technology can fix on its own, is how we get to a more just world. I, I really, really fundamentally believe that, um, economic justice is the most important thing you can do for social justice, and if you have all of the resources going to a small set of people, um, even if everybody else has, having their absolute quality of life raised very quickly, that's not enough.

Jorge Cueto 30:56 You recently announced a united slate which puts forth some policy proposals as well as an invitation for candidates. What would you say success looks like, like for that?

Sam Altman 31:07 You know, if we can run five, six candidates in the 2018 cycle, um, on that election, on, on that platform, and have two of them win their elections, um, I think that'd be an incredible start. And, you know, maybe it doesn't work at all because may- maybe the kind of, maybe those issues, although I believe they're important, are not yet ready, um, for, to convince the public at large. Um, but I think it'll be a good start, and I think over time it would be really good if people on the progressive side, um, built up a kind of longterm organization focusing on winning elections and shifting public perception at all levels. Um, the right has done a fabulous job at this. Congratulations to any of you, um, who are on that side. But I would like, uh, the, the left to do a b- a better job at that. I don't, I, I don't think we're putting forward our best game there.

Jorge Cueto 32:02 In your invitation you also mentioned that you'd be willing to work with Democrats, Independents, and Republicans. Is there any issue in particular that you see anyone regardless of political ideology coming together and agreeing on more so than other issues?

Sam Altman 32:16 Um, I mean, I think there are, there are a lot of issues. A- again, I think people agree on way more than they, they disagree on. So I think, uh, you know, as we went around the state and talked to Californians, um, Republicans, Democrats, San Francisco, LA, Fresno, Shasta County, wherever you go, um, the price of housing is like the biggest issue for most sort of regular people that are not Google employees. Um, maybe even for Google employees. Um, it's really, really bad. I mean, this is like, this is ... If you could fix, I think, one thing that would have hugely positive secondary effects, um, it's like bring the cost of housing down by a factor of 10. It'd be transformative for society. It's a- Eliminate government regulation. What? Eliminate government regulations. Eliminate government regulations? I- Buy stories. Yeah. Like, I think many regulations have good elements. I'm not, I'm not kind of the libertarian anarchist, all government regulations are bad. But I do think the, the sort of no more building, no building taller has been disastrous for the state. Uh, and I think, like it is the worst possible allocation of resources to have people tie up every free penny they can find into the place they live. Um, like think about anything else we could spend that on. Think about what it would be like if people could like live close to where they work rather than commute an hour and a half to get here each day. Um, and that is something that Democrats, Republicans, Independents, almost everybody agrees on. Uh, or at least there are people in all of those camps that, that do. Um, so I think like, again, easy to talk about the divisions. It's really good that when you go talk to regular people, um, they all agree on what the most important issue is, and they all wanna fix it in the state.

Jorge Cueto 34:05 What do you think the tech industry's role is in government and politics?

Sam Altman 34:10 Um, look, I think we are all citizens of this country, and we all have, like, you know, there is a, there is a right to participate in the, uh, electoral process, and I think there is also an o- there is a duty to do that. And I think, uh, when ... I think, I think tech is going through such a boom right now, it is easy to say, "Hey, I'm just gonna focus on, like doing my thing. Um, you know, building products, making money, and whatever, and I'm gonna figure out someone else's, let someone else take care of the rest." Um, one of the kind of like disappointing things that I have learned as I've gotten to spend more and more time with increasingly influential people and kind of how the world works, is that, um, there is no plan, and there is no group figuring out everything, and it's, it's kind of like up to all of us. Uh, everyone hopes. Like the reason conspiracy theories are so appealing is that like everybody wants there to be a conspiracy. Like, you wanna think that like someone has a plan. Um, that, you know, there's like, there's all this stuff happening, but someone's got a centralized plan, and it's all gonna work out. And like I think one of the sad realizations of growing up is that there are no grown-ups. Um, no one has this master plan, and if, if we don't participate, the thing can just go off the rails.

Jorge Cueto 35:29 What do you think are some of the most effective ways for employees at a large tech company like Google to get involved in government and politics?

Sam Altman 35:37 Um, everyone wants an answer other than this one, because it would be more convenient, but I think one of the answers is to just run for office. Like, we have this process. We have this like system and set of rules that we've all agreed to, um, and everyone wants some way to hack around the edges of that rather than just participate directly. Um, and that's okay. There are good things to do, but I think just taking the problem head-on is, uh, not enough people try that.

Jorge Cueto 36:08 Mm-hmm. And there's rumors going around that Mark Zuckerberg is gonna run for office like every other week-

Sam Altman 36:15 [laughs]

Jorge Cueto 36:16 ... and then there's also been rumors about you running for governor, and we've seen- Uh, tech people run for office in the past. Meg Whitman ran for governor of California in 2010, but she lost by a significant margin to Jerry Brown. So my question is, do you think that it's actually feasible for someone in tech to run for office and win, or is there too much of this perception of the tech industry as being elitist such that people won't connect with voters?

Sam Altman 36:43 I, I think there is. I think at least people should try. I think the tech industry is the most hated in the Bay Area, and if you go out throughout the rest of the state or the country, there's a lot of people who think it's really cool. They aspire to it. They wanna, they wanna participate in it. Um, and I think it is easy to get too negative a view of how technology people are perceived here. I don't know if a tech person can win the presidency. I have a feeling we're gonna find out. Um, but what I am confident about is that tech... Not... People from the technology industry could start winning local school board, and city council, and, you know, congressional seats, and that'd be a great thing to start with.

Jorge Cueto 37:20 Is there any person currently living that you would want to run for president? Who's your ideal candidate?

Sam Altman 37:32 Ideal, I can't answer that on the fly. That's... I, I have... I would like... It's a good question to think about. When you said that, I was like, "Oh, I should figure out who that person is and go try to convince them to run."

Jorge Cueto 37:40 [laughs]

Sam Altman 37:42 Um, but I don't have an answer, like, ready to go in my head.

Jorge Cueto 37:44 Okay. Well, with that, I'm going to open up the floor for audience questions. So if you have a question, feel free to go to the mic in the back of the room, or also submit your question to the Dory.

Speaker 37:57 Go for it.

Jorge Cueto 37:57 And, mm-hmm.

Speaker 37:59 Hi. Um, you mentioned the anecdote about your friend being unhappy and, and Kim Kardashian, and it reminded me of a quote, and it said, uh, "People aren't unhappy because they want to be happy. They're unhappy because they want to be happier than other people." And you mentioned that technology may or may not be able to solve this, so I'd like to hear some of your ideas or success strategies on how to solve this on a micro and macro level, whether it be delete Twitter to societal level.

Sam Altman 38:30 That's a good thing to do.

Jorge Cueto 38:31 [laughs]

Sam Altman 38:33 Um, that's a good thing to do. I, uh... Yeah, I deleted all the social apps from my phone. Um, and I, you know, I can still, like, check them on my computer when I want, but, uh, I don't have that, like, constant urge to, like, push button for dopamine hit here. Um, and that's, that's been really good. Um, you know, I think after talking to a bunch of these people about the things that make people happy, there's, like, the obvious stuff which is, like, spend time with loved ones, um, that everyone knows even if they don't do. But there's a bunch of things that are, like, very not... Or, or less obvious to me that seem to have huge measured effects, um, like taking time to think about the things that went well as opposed to things you're upset about, or, like, going for a walk outside every day no matter what. Um, and I think as we get sort of to abundance and unlimited resources, this is gonna be a more important topic. Um, what I would encourage you to do is to, uh, just, like, start reading. There's a lot of literature on the subject no one seems to care about enough to spend time, but I would say start reading it. Um, and hopefully I'll finish my blog post on it soon. Thank you.

Speaker 39:51 Hey, Sam. Thanks for coming. Uh, wanted to ask a question about the larger landscape of VC in general. Uh, heard a rumor that YC's looking into a growth fund, think it broke on TechCrunch like a month ago, like, potentially raising a billion dollar fund. Uh, wanted to get your thoughts on growth verse venture and how you see growth evolving, uh, for the Bay Area in general. Uh, you see private companies staying private-

Sam Altman 40:13 Yeah

Speaker 40:13 ... longer and longer.

Sam Altman 40:15 So, um, w- we already have a growth fund. It's called, uh, YC Continuity. Um, we raised the first one in 2015. Um, it's about halfway invested. Uh, and it, it basically follows on in capital, um, in YC companies. One of the things that I wanted to do after I joined YC was start to fund hard tech companies, um, nuclear fusion, synthetic biology, quantum computing, self-driving cars, long list. Um, and one of the, the... One of the good things that I realized is that nobody else was funding those companies, and so we could have our pick. Um, one of the bad things that I realized is nobody else was funding those companies. Um, and so we could fund them all we wanted, and there was no follow-on capital. And I think... So that was the actual, the genesis for our growth fund, is that there was this class of companies that I think are really important and really valuable, and they weren't getting funded. Since then, we've expanded it, and we, we fund lots of other companies. Um, I will certainly say that, um, there is no shortage of growth capital for software companies in Silicon Valley. Like, you know, it may have gotten harder to re- raise a seed or an A round. It probably has somewhat. But if you have things working, uh, like, when you wanna go raise a B or C round and you have, like, this beautiful exponential growth, like, you will not be able to keep up with the number of term sheets you're getting. Um, that is the stage that I think just has a huge amount of capital allocated to it. As, as companies are staying private longer, as public market investors have a harder and harder time finding alpha, um, they are more and more willing to do stuff like this. So, uh, that part of the market, um, is not suffering, at least not yet.

Speaker 41:58 Thanks.

Sam Altman 41:59 Sure.

Jorge Cueto 41:59 Let me do a question from the Dory. Uh, we have a question about universal basic income and how we've... Google has been supporting GiveDirectly's research on UBI in Africa, and you're a proponent here in the US, so what are your thoughts on how, when, and where the universal basic income approach might be most effective?

Sam Altman 42:23 Yeah. Um- I don't think universal basic income is a solution to all problems. Um, I think, in fact, of the bigger problem that we were talking about earlier, of what makes people happy and fulfilled and feel needed and valued and have meaning in their lives, um, it is not sufficient to solve that problem. And I don't think it'll replace, like, the entire other social safety net, like, the people who are like, "Eliminate everything else," you know, "No healthcare, no schools, no minimum wage, um, put basic income," I don't believe in that either. But I do think, um, that if we do all of our jobs, if we, Silicon Valley, do all of our jobs, we will create, um, more wealth than the world has ever seen before, and less jobs. And in that world, um, I think there is a moral obligation to eliminate poverty. I think poverty is a obviously very bad thing, but it is worse even than people realize. Um, if you, if you look at the studies on the long-term psychological damage that living in constant poverty does to people, and how it means that you never get a chance to really invest in your own future 'cause you're always just trying to survive, I think there's just a huge amount of wasted potential. And so, you know, there are a lot of arguments about UBI. This was another thing, like, when we started this a couple years ago, I did not predict at all that this was going to, like, ignite into this big national debate. It was like, we're just trying to hire a researcher, you know? Like, maybe this is a good idea. Can we just do a project? And it's just been like [explosion noise] Um, so I, I don't think I quite understand why it has become such a national issue so quickly, but I do think that, uh, longer term, um, if we have the ability to eliminate poverty, we will get a lot more out of the world as a whole.

Jorge Cueto 44:11 As you started the research, or the experiment in East Bay, have you noticed anything that maybe you didn't expect, or has it changed your attitude towards it in any way?

Sam Altman 44:20 You know, we're like seven or eight months into a five-year study.

Jorge Cueto 44:23 Mm-hmm.

Sam Altman 44:23 So not yet.

Jorge Cueto 44:24 Okay.

Speaker 44:26 Uh, hi, Sam. I'm thinking, uh, more in general of founders, but this could be general too. But what advice do you frequently give but you find most people never actually follow through on?

Sam Altman 44:38 Uh, I mean, no one ever listens to any advice of any sort. Um-

Speaker 44:42 [laughs]

Sam Altman 44:43 So all of it. Uh, I think, I think the thing that people... Well, the piece of advice that people later say they wish they had listened to, um, is that it is very easy to get sucked into a path in life. And, you know, you can do things like say, "Oh, I wanna start a startup someday," or, "Oh, I wanna be a, you know, AI researcher someday, but first I'm gonna go do this other job to like, you know, make money and gain experience and whatever." And it's so easy to, like, get sucked into a path where you spend your entire life doing something that is not really what you wanna do. Um, that's probably the piece of advice that I have given people that they have most often come to me like five years later and said, "I really wish I had listened to that."

Speaker 45:32 Thank you.

Sam Altman 45:32 Sure.

Speaker 45:35 I'm curious to hear your thoughts on relative rates of progress in AI between China and everywhere else. So you mentioned nobody has a plan, but China has a plan for state level centralized investment in, in AI, whereas Elon Musk is calling for regulation that would slow down-

Sam Altman 45:50 Yeah

Speaker 45:50 ... presumably everywhere else.

Sam Altman 45:52 Uh, China does plan, that's true. Um-

Speaker 45:55 [laughs]

Sam Altman 45:56 They don't, it doesn't always work. They don't always have global coordination. Um, you know, like right now we're in this kind of world where it seems like you have, like, the internet in China and the internet in the rest of the world. Um, maybe they have a plan for their piece of it. Um, I, I don't honestly know how far along China is with AI, and I don't think anybody else honestly really does either. Um, I think it would be bad to get into an arms race with China over AI. I think that's something we should try to avoid. Um, and I think there is a real opportunity, although it might, we may not have the leaders in place to do it right now, to make this a, a joint kind of like worldwide project, um, rather than another space race or nuclear arms race.

Speaker 46:52 Thanks.

Speaker 46:55 Uh, hi, Sam. Uh, I was wondering if you could expand a little bit on funding hard tech, because I think there's a big problem with, uh, private capital allocation, and, like, $70 billion was put into VC last year, which is a drop in the bucket in terms of worldwide investment. So do you see, like, YC maybe raising an $100 billion fund in the future?

Sam Altman 47:13 Yeah, maybe. Um-

Speaker 47:14 Like, like how, how, yeah, how can we solve this?

Sam Altman 47:16 I, I think capital is allocated really badly. Um, and unfortunately, there are huge efforts, like if you have this really valuable thing, which is you get access to invest in great startups and most of the world doesn't, um, then there will obviously be a super return there, and you're gonna work really hard to protect that. Um, and so I think, um, I think expanding access to invest in startups, which is happening, but slowly, is a really good thing to do. And I think, um, you know, like crowdfunding, equity crowdfunding, we're still in the early days of, but is an important trend. Uh, and I think that will change the capital allocation. I think you're seeing, I think you're seeing startups in all these different ways go to non-traditional funders. Um, and that's been really good. We first saw that with consumer hardware. We're now seeing that in a lot of other places. Um, you know, I think the other thing that's happening finally is there's just, and there's bad to this too, but there's just a huge amount more capital coming in to fund early stage private companies, or mid stage private companies. And I expect- That in a world where interest rates stay even close to as low as they are, um, we are in the very early innings of that trend. So I think capital will get... You know, my, my general model is that the world's this very complex financial system. Capital sloshes around looking for the best return. Um, there are periods where there's a really out-performing return, but then more capital gets allocated there, and I think we're seeing that now.

Speaker 48:49 But, uh, do, do you think that a nuclear fusion company could raise $100 million in an ICO or something like that?

Sam Altman 48:55 Yeah. You- or maybe I wouldn't suggest that they do an ICO.

Speaker 48:59 Sure.

Sam Altman 48:59 Um, but I do think they can raise $100 million. Uh, and I don't think they could have three years ago, and I think that's a really positive trend for the world, that this stuff is now, um, at least somewhat possible.

Speaker 49:11 Awesome. Thanks.

Sam Altman 49:12 Sure.

Jorge Cueto 49:12 Another Dory question. What are the most common reasons that qualified candidates get rejected from YC?

Sam Altman 49:19 Um, well, there is one particular thing that is, uh, that happens a lot with super talented people from large tech companies. So since I'm here, I'll mention that one.

Jorge Cueto 49:34 [laughs]

Sam Altman 49:34 Um, which is you are a really great person, really talented, really smart, um, really driven, and you don't have an, a good idea, or you don't have any idea, so you make up a bad one. Um, and there's this, like, myth about startups that the idea doesn't matter at all. Um, and I think that's just, there's existential proof that that's not true. Um, and I think there is, there is this particular failure mode of people from the big tech companies, which is like, "Hey, I'm really talented." I'm like, "Yes, you are." Um, "So I'm gonna, like, start a company, and I'll figure out what to do later." And I think that isn't usually what produces the great companies. And so we have learned, um, that it's much, much better if you're otherwise a qualified candidate, um, to have a, have a good idea, and I think that's the, that'd be the common failure case for people coming out of Google.

Jorge Cueto 50:29 Mm-hmm.

Speaker 50:31 Hey, Sam.

Sam Altman 50:32 Hey.

Speaker 50:32 So, um, what are your thoughts about, you know, where you're gonna go or what you're gonna do with your life after YC? Or does that come across? Like, are you gonna take a political office seriously?

Sam Altman 50:41 You know, I don't even, like, know what I'm gonna do in three weeks. Um, I, uh, I plan to run YC for a really long time. Um, I, I do plan to get continually more involved with politics, but I'd like to support others, um, not run for office myself, at least not anytime soon. Um, you know, like, I am a big believer in do the things that you, like, think about in the shower in the morning when you can think about anything you want, and your, kind of, mind is just off. And the things I keep coming back to are, like, how can I make YC like 100 times bigger? How can I make sure that the arrival of AI goes well? Um, and in the short term, how can I help our political system from going off the rails? So, or any more off the rails than it already has.

Speaker 51:27 [laughs]

Sam Altman 51:27 Um, so you know, I kind of plan to keep working on those things for hopefully, you know, a really long time.

Speaker 51:35 Good to hear. Thanks.

Speaker 51:37 Hi. Thank you for, uh, being here. Um, I have a question about the general, um, uh, way Silicon Valley works because it seems like it's a system which, in which you have, like, the most talented minds dealing with maybe not the most important issues. Like, you have the sharpest mind working on social apps or mobile apps, but we still don't have a cure for cancer, or we still don't have a, another way of, like, having energy or, like, really important stuff that the, the world should handle. Um, so my question is do you agree with that? What is your perspective on that? And how do you think maybe this can be addressed?

Sam Altman 52:14 I don't agree with that. I think, um, I, I used to view my job as a capital allocator, and now I view my job as a talent allocator. So most of my day is, like, spent meeting with really smart people and trying to convince them to work on important problems. And you know, I think in 2007 what everyone said is, "The best minds of our generation used to build spaceships and now they're, like, moving numbers around on Wall Street." And in 2017, what people say is, "The best minds of our generation used to build spaceships and now they get me to click on ads." And you know, there's always some truth to that. There's always someone to pick on there. Um, but, like, I think OpenAI has some of the smartest minds in our world, of our generation working on having AI go well. I think Helium, which is that fusion company I mentioned, some of the smartest minds of our generation working on nuclear fusion. Um, YC, I think last time I counted, it's funded eight companies working on a cure for cancer, incredibly talented people, or cures for cancer, cancers. Um, I think you can always say there's all these people, like, working on bad stuff or stuff that doesn't matter. But you know, I think Google is still, like, a really great thing for the world. Uh-

Speaker 53:25 This is why we are here.

Sam Altman 53:26 It's, yeah. And I think it does matter, and I think it's always easy to pick on people and say, "You're not working on a cure for cancer, you know. You're wasting your time." But you're making this thing better that people use every day, and their lives would be a lot worse if it went away, and you're making it better every day. And I think it's, it's, it's really easy to pick on people, um, and say, you know, "That person's not spending her time right." And I think it always says more about the person that says that than the person they're pointing to. And if you're doing something useful for the world, if you're doing something you enjoy, if you're sort of, like... Even if you're having a small impact, but on a product that a lot of people really use and love, um, I think that's really valuable, and I think the people who say this generally have a lot of insecurity about how they're spending their time.

Speaker 54:10 Hm. So you don't think, like, Silicon Valley is, like, avoiding the tough, uh, problems?

Sam Altman 54:17 I invite you to, like, come sit in my office for a day and, like, listen to the people coming through and what they're working on. Uh, I really don't think that, no.

Speaker 54:24 Okay, cool. Thank you.

Speaker 54:27 Uh, hey, Sam. Uh, seeing that you're a largely influential public figure, uh, could you speak to the privileges that might have led to that, um, and helped you with your success, and, um, how you may be a good or bad ally? And then also, seeing that a lot of our consumer products are largely biased, um, spec- specifically to the people who create them, which is largely Silicon Valley, um, engineers like myself and others in this room, uh, how do you keep your ideas and thoughts diverse in, um, addressing intersectional needs?

Sam Altman 55:02 Yeah. Great question. Um, I basically had, like, nearly all of the possible privileges. I had wonderful, loving parents. Um, I grew up in a safe house. I'm a white guy. Um, we had enough money that I was able to pursue the things that I was interested in and go to a great college. Um, and it's never lost on me how if I had been born a mile in a different place to a different family, um, different skin color, different gender, I wouldn't be where I am now. I view that as an, um, obligation to try to make the world more just going forward. Um, I think anyone who is really successful and doesn't ... Like, I think anyone should try to do the best they can with whatever hand they're dealt, but if you're dealt, you know, like, four aces, then, and you win, then I think you have an extra obligation to try to sort of make the world a little better. Uh, so I try to be really thankful of what everyone's done that has allowed me to do this. Um, I also try to figure out how to pay that forward. Um, and I think, I think anyone who is really successful, or almost anyone who is really successful, has, um, like, privilege, luck, skill, and hard work, and I think people who try to say it's just one or just the other all tend to be wrong. Um, so I'm thankful for that, and I try to pay that forward. Um, I ... In terms of biases in the product, I think, uh, this is one of the reasons that diverse teams are most important. The moral question aside, uh, that consumer products, the teams, the companies that I think have done the best job addressing this head-on have had a very diverse set of voices around the table, and I think that is always the strategy I recommend, 'cause that's the only one I've seen consistently work.

Speaker 56:53 Cool. Thank you.

Sam Altman 56:54 Sure.

Jorge Cueto 56:55 We're on time now, but if you can do these last three questions quickly, we can get through them.

Sam Altman 56:59 Sure.

Speaker 57:00 Thank you for coming, Sam. Uh, my question is on machine learning and, and artificial intelligence companies. Uh, it seems to me that the scarce or valuable thing is data in those areas, and less so the algorithm, because they're mostly open source. So I was wondering how you think about that when getting pitches from comp- uh, companies.

Sam Altman 57:17 I, I used to believe that. I now believe that it's gonna be compute, not data.

Speaker 57:21 Hmm.

Sam Altman 57:22 I think data is important, but there will be a lot of it available. Uh, and just my own experience with OpenAI, um, to really be at the forefront here, you just need massive amounts of compute. And so I used to ask companies, like, how they're gonna get a lot of data. Now I ask them how they're gonna get a lot of compute.

Speaker 57:40 Okay. Thank you.

Sam Altman 57:41 Sure.

Speaker 57:42 Hey, Sam. Thanks for coming. Um, I know you were talking about, like, you know, people are becoming unhappier and the world food problem, all this kind of things that are kind of, like, quantifiable, like, physical needs of people. Have you or had companies that come to you kind of like try to solve the spiritual needs of people? Because we can identify the physical needs, but what about, like, spiritual needs or research or companies?

Sam Altman 58:00 We, we have had a few. Um, none of them have really worked yet. Uh, but you know, one that, like, stuck out in memory was a company came to us and said, like, churches, like, you know, organized religion had this really important effect that was totally separate from the, the religion itself, which was this c- this tight-knit community, and how do you build that in a world where most people, or a declining number of people, believe in religion and go to church? Um, and so I think there are people thinking about things like that, which are sort of these non-obvious attacks on the problem that are interesting. Uh, but none that I could yet point to as, here's this thing that's really worked well.

Speaker 58:42 Hey, Sam. Uh, so you partnered with the ACLU earlier this year, which you got mixed reactions to. I'm wondering what you've learned from that whole experience, uh, what successes you've had, and what partnerships you're looking forward to with other nonprofits in the future.

Sam Altman 58:58 I, I, I've been ... I was really excited about how that went. That was, uh ... We had never done anything like that before. Um, we often try new things. Usually they don't work. Sometimes they do. That's one we would do again. Um, we would do something more like that. I think there are these really important organizations in the world that can use our help to build better technology teams, um, and that was an experiment that went well and that we'd love to try again. Um, one of our software engineers, Cajun, went there for, I think she went for, like, eight weeks, almost the whole program. Sat in their offices, helped them. We got a, you know, call from them later about how well it went, and been able to help them, um, put a, you know, expand and supplement their team. Um, and that, you know, I think that's something we'd like to try again.

Jorge Cueto 59:43 And one last question, because it's gotten so many-

Sam Altman 59:44 Sure

Jorge Cueto 59:45 ... votes. Elon Musk recently called a, for preemptive AI regulation at the National Governors Association. As a chair of OpenAI and friend of Musk's, what is your opinion on this issue, and what specific actions can we take to minimize future risk?

Sam Altman 59:58 The specific thing I would support today is just, um, insight. Like, I think the government should understand where the edge of capabilities are and how it's evolving, um, because I think no one, certainly not the government, knows what the regulation for AI should look like today. But I'd be in favor of starting that education process.

Jorge Cueto 1:00:19 Awesome. So yeah, that was our last question.

Sam Altman 1:00:21 Thank you all.

Jorge Cueto 1:00:21 Thank you, everyone, for coming. [audience applauds]

Sam Altman 1:00:26 Thanks for having me.

Jorge Cueto 1:00:27 Yeah. Appreciate it.