[0:00:03] All right. Is this thing on? All right, it is. Good. All right, we're going to wrap up the day. I'm going to do a fireside chat with Sam Altman. He is the president of Y Combinator and considered one of the best early stage, seed stage investors in the world. So really excited to hear from him. Sam, welcome. Thanks for being here. Thanks for having me. All right, so you've been at Y Combinator for about a year now, right? Well, I've sort of been there for like 10 years.
[0:00:37] Actually, 10 years ago today, YC funded my company. Wow. Okay, but you've been leading Y Combinator for just over a year. A year and a half, yeah. So how has that been different than what you expected? Like, what have you learned? You know, I think one of the things that has surprised me on the positive side is how many incredible companies exist in the world. You don't see this outside of YC because you only see the 200 we fund a year.
[0:01:06] But at this point, 15,000 companies a year apply to YC from around the world. And probably half of them are fundable, high-quality companies. So we haven't yet figured out how we fund 7,500 companies a year, maybe someday. But I think the depth that is left in the sort of entrepreneurial ecosystem that we don't get to has been one surprise. Another surprise is how well the same YC model works for lots of different companies. So we've now funded
[0:01:35] Four or five nuclear energy companies, rocket companies, quantum computing companies, synthetic biology companies. And the same lessons for YC, which intuitively most people think wouldn't work, work really well. So this focus on speed of iteration, doing things quickly and cheaply, that has worked for a much broader range of people, of things than people would have thought. And I guess the negative surprise is sort of the amount of external media requests I get and just trying to figure out how to deal with that.
[0:02:04] I bet. So how do you see YC scaling over time? Are you planning to fund a lot more companies? You just talked a little bit about how you've been funding some very different types of companies. I've got some other questions about that. But do you think that YC changes in that you just start funding more and more companies over time? Or is it going to be more about where you focus? Well, I think it'll be both. I mean, we want to fund more good companies, not just more companies. But I think if you just look at the sort of trends in the world,
[0:02:34] We need to figure out some way to be able to fund a thousand companies a year or more than a thousand companies per year. But really what we want to be is sort of this like hub of innovation in the world. And we're kind of flexible on how that looks. But whatever the smart people are doing that's going to create massive new value in the world, we want to support. And I think if we can put out, you know, if we can somehow help put out into the world more new innovation than Google does, that's great.
[0:03:03] And if not, then are we making the best use of our time? I don't know. How we get there, I think that's sort of flexible and continuing to evolve. But we fund companies now from all over the world and all these different sectors, a lot more of them. And then there are other things that we'll do in the future that we don't do now. And there are things that we do that we'll stop doing. But I think that's kind of the vision we all rally around. So you talked a little bit about funding different types of companies, right?
[0:03:33] Biotech, nuclear energy, I think even some nonprofit stuff, unless I'm mistaken. Now, some of those, I mean, particularly on something like biotech or nuclear energy, that has a pretty long lead time before these things turn around. How do you think about that versus what we're more used to, which is investment in these smaller companies that can turn things around very, very quickly? There's a huge premium that you get for being willing to hold an illiquid security for a long time.
[0:04:02] So we're delighted to do that. I think most people put way too much of a premium and they try to like, you know, buy a stock right before earnings and sell it three days later. And, you know, the market, there's like a lower, you don't get as much value out of that. If you're willing to hold things for 10 years or longer, then, you know, I think you get much better returns. And we are not traders. We like to hold things for a very long period of time. We invested in Airbnb six years ago.
[0:04:31] I really hope we're still holding those shares in another six years and actually quite a lot beyond that. And I think that I've actually, I personally have never sold a share in a company unless a founder has asked me to. For like, because they need liquidity in a round or something. But I really try hard to be a very long-term holder. And I think that is one of the things that's different about YC from other investors.
[0:04:53] You know, we don't have limited partners. We're our own money. So we can hold shares on this sort of 20-year outlook. And venture funds with a 10-year life can't do that. And so one of the reasons that people always say, you know, why do we have all these software companies? Where are the nuclear energy companies as a popular example? You need investors with a very long-term time horizon. And I think we have that. And although people think of us as funding these sort of companies that become successful quickly,
[0:05:21] I think the secret to our success is that we invest with this infinite time horizon. One of the secrets. That's pretty awesome. You have this opportunity to see all these different kinds of companies and all these different kinds of ideas. But where do you think we're wasting time as an industry? Or another way to say it is, what frustrates you when you see some kinds of companies and some kinds of ideas that are maybe played out?
[0:05:48] Versus where do you think people, and a lot of people maybe in the audience today, should be focusing more of their attention? Okay, I was actually ranting about this this morning, and it's like a pet peeve of mine, is when largely journalists, but lots of other people that never create companies either, love to complain about how YC or other investors fund small companies, founders fund small companies.
[0:06:20] If you think about the most important companies of the last, say, 15 years that got started that we use every day. When Google got started, the press mocked them for like, why does the world need a 13th search engine? There's not even a portal, it's just a search engine. Facebook, come on, a social network for college students, who cares? Airbnb, people used to make fun of us for funding Airbnb all the time.
[0:06:42] Because they're like, oh, who's going to stay on an air mattress in some stranger's apartment? And honestly, that didn't sound that appealing to me either. But these things evolve. And exponential growth is this very powerful force. And people always make the mistake of calling an idea small or stupid because they don't understand how it's going to evolve. So my advice on this always is to ignore the people that don't do anything. Start whatever company you want.
[0:07:11] And just continue to pull at that thread and see where it goes. A lot of the times it turns out to in fact be small. Sometimes it turns out to be Google, Facebook, or Airbnb. I never try to criticize anybody for what they're doing because I've just learned after doing this long enough that if you bet on smart people, they eventually figure out their way into the big idea.
[0:07:38] Where you would like to see people investing more time? Just to dig in on that one a little bit more, because I think it's interesting to get your perspective. So we mentioned the nuclear energy companies earlier. I think that if you could pick one thing and snap your fingers to improve the lives of the poorest 3.5 billion people in the world, really cheap energy would be that. Cheap energy that is environmentally friendly and safe and all of that. So I think that energy is an area I've been particularly interested in.
[0:08:06] I'm much less interested in the sort of things that compete with subsidies to be barely as cheap as oil or natural gas. But if we can really figure out a way to make energy for a penny a kilowatt, that would be transformational. And I think that's something that people don't work hard enough on. I think that we all sort of gave up on nuclear in particular, but just the energy problem in general. And I think that is like,
[0:08:35] This incredibly high leverage thing that factors into basically everything else. Most wars get fought over energy. There's huge economic effects. That would be really good. I think AI will probably most likely lead to the end of the world, but in the meantime, there will be great companies created with serious machine learning.
[0:08:59] I actually just agreed to fund a company, not even really a company, sort of a semi-company, semi-nonprofit, doing AI safety research. But I think the machine learning stuff is just working so well at this point and has so much longer to go that I think that's really important. And then I think, I could actually rant about this, go on about this for a long time, so I'll just do one more. I think that synthetic biology is really interesting.
[0:09:29] One company that I met with earlier today is working on a generic antiviral. And the reason they decided to dedicate their lives to this is they got really scared by synthetic biology. And if you can create a really bad virus in a lab and sort of unleash it into the world, you'd like to be able to defend against that. And so this company is working on this early but very promising anti-antiviral specifically for
[0:09:56] The synthetic disease risk. I think there's a lot of other amazing things that are happening with biology. The things that we look at, by the way, for new areas of startups are cost and cycle time. And when the cost comes down and the cycle time comes down, you can have a startup exist where you couldn't before. So for example, genomics, really good example. The cost and cycle time has come down faster than the Moore's Law cost and the cycle time even quicker than that.
[0:10:24] You can now have a company that does a diagnostic of infectious disease by just sequencing everything. That's how cheap it's gotten to sequence DNA and RNA. So you can take a sample of blood or whatever, sequence every piece of nucleic acid, anything in there, and then match those. It becomes a computer science problem, and you can say, ah, this gene is from this bacteria, this one's from this virus, here's what's in the sample. So there are all these new things you can do in biotech because of cost and cycle time that weren't possible even three or four years ago.
[0:10:54] All right, I'm just going to ask one more about nuclear energy. You just talked a bunch about biology, but I want to bring you back to nuclear energy because I think it's so interesting. Why do you think that it feels like the world has given up on nuclear energy? Do you think it's like a PR problem or is it a technology problem? I think it's like a little bit of both, but we'd be able to fix the technology problem if it weren't for the PR problem. There's some debate about...
[0:11:20] Are people afraid of nuclear energy because of the bomb or because of Three Mile Island? And I think, you know, people have convincing arguments on either side, but, like, the fundamental problem is, like, people are much more sensitive to sort of, like, theatrical extreme risk than they are to sort of, like, boring, slow plotting risk. So, like, nuclear energy has an unbelievable safety record.
[0:11:51] You know, it's like 1,000 or 10,000 times safer than coal. But most people, would they rather live next to a coal plant or a nuclear energy plant? They pick the coal plant all day long. People just really, you know, and when they die in 30 years of lung cancer, it doesn't feel as sort of dramatic as dying in a nuclear meltdown. And no, it's true. Like, this is like the human risk miscalculation that always happens, right? People always underweight the boring, slow stuff.
[0:12:19] And overweight the quick dramatic stuff. And so I think that has been a really big problem. And that has prevented us from doing a lot of new technology development on nuclear. I think the world is beginning to change there. And I think that, especially if fusion works, that will be a big step forward. But even if not, I am very convinced that the 22nd century will be powered by
[0:12:46] Nuclear energy of some sort, either solar and wind or sort of nuclear reactions on Earth. All right, let's get back around to entrepreneurs. So I personally have been just profoundly wrong about a couple of entrepreneurs that I met in my life. Like, met somebody really early on, walked out of a conversation with them, and I'm like thinking to myself, this person will never be successful in life at all. And then they went on to do...
[0:13:15] Unbelievable things afterwards. So I'm curious to know, my own prediction record is not very good, so obviously I should not get into venture capital most likely, but what are the signals that you look for? Like how do you tell if it's somebody that you want to bet on? You know, no one gets this right all the time, and that's sort of why it's fun. But the things that we look at, you know, clarity of vision is really important.
[0:13:41] And I don't know why this correlation is as strong as it is. I think it's because if you can't express yourself clearly, you can't recruit, you can't get press, you can't raise money. But as we've looked through our history of now the thousands and thousands of companies we've interviewed and tens of thousands we've looked at, this correlation is very strong. So founders that can express in a few sentences very clearly what they're doing and why, that's really important. That's the main we look for. And even if the idea doesn't sound great, if it's at least clear, that's helpful.
[0:14:11] The level of determination is, we've talked about this publicly, but that's the thing we look for the most. Some evidence that you're not gonna give up when things get hard, either because of things you've already done with this company or other times in your life, but most startups fail in the first couple of years because the founders give up at some point. And so really trying to figure out founders that are not going to do that, that's really valuable.
[0:14:39] And part of that is founders that are sort of passionate about what they're doing. A common failure case is you have sort of this MBA that has this very polished signing idea with a lot of market research and picked it because he or she thinks they're going to make a lot of money. That is not nearly as good as someone like a Brian Chesky who just like cares about the idea and is somewhat unpolished in the details and the financials. I actually still remember sitting
[0:15:08] With Brian and going through a financial model with him for the first time. And being like, you are financially illiterate. I don't even know how to explain this. He's learned a lot. It's okay. Are you still here, Brian? All right. It's fine. Yeah, the point is that doesn't matter. So then we look at intelligence. And actually, intelligence is usually relatively easy to tell in a 10-minute conversation. Determination is much harder. And then beyond that, everything else becomes less important.
[0:15:39] Got it. So I think that... Oh, sorry, can I add one more thing? Yeah, go for it. If the co-founders are fighting in the room, which happens a lot more often than you might think people would do, that's a very bad sign. All right, so if you're going to fight, do it out of the room. All right, so I think a lot of engineers these days sort of face this decision, well, actually not just engineers, like lots of people, about whether they want to go start a company, they want to join a really early stage company, they want to join a mid,
[0:16:12] How do you help people think through those decisions? They're all incredibly personal and idiosyncratic to the situation. But I try to share with people the two general frameworks that I have used in my life to make decisions. And they're somewhat at odds with each other. But one of them is sort of the...
[0:16:41] The impact maximization equation and the other one is the regret minimization equation. The problem is that sometimes they get sideways with each other. But I think in general, you really should try to live your life without regrets. And if you are really set on starting a particular company, even if you know that you could go be employee number 30 at Airbnb and Airbnb is already going like this.
[0:17:10] And you'd have more of an impact, but you're going to regret for the rest of your life not starting this particular company because you're so in love with it. You might decide to do that. However, I think this other thing, like making an impact on whatever you do is sort of the satisfying thing in life, in professional life. And so I think most people find that their job satisfaction goes up with the impact they have on their company or the world or whatever. And if you start a company that,
[0:17:41] Is a mild success or join a company that's a huge success, you probably have more of an impact in the latter case. So these are sort of the two things I talk to people about, but the hard part is weighing them and getting the probabilities right. Okay, so your personal philosophy, I'm told, is great products win. And the importance of listening to what your users want, then improving the product, going back and listening again.
[0:18:13] Who do you think is doing this really well? And maybe even who doesn't, if you can say. And if you don't do this well in the beginning, how do you recover from it? Or can you recover from it? Well, I think there are cases of a company that does not make great products but has a killer sales and marketing team that survives for a while or has some incredible monopolistic effect where they can survive for a while. Microsoft.
[0:18:44] For years was given an example of that second case, although I think they're doing much better work recently. But I would say there was a long stretch of time where Microsoft's products were not good, and the company managed to survive. But I think if you look at the companies that endure over decades, they tend to win on product. It is true that it is critical to get sales and marketing right, because without that, you just don't get the flywheel going.
[0:19:13] The most valuable tech company in the world is Apple, and I think they make just the best products across the board. I don't know if Google is still number two, but I think Google search dominates everything else, and also most Gmail and Google Maps dominate their respective products. I think Airbnb won on product. I think that Facebook, although they came late to the game, had the best product by far, and they've managed to expand that.
[0:19:43] I don't ever try to let people off the hook for sales and marketing. Because if you build the best product in the world and don't get anyone to use it, that still kind of sucks. But I think it is very hard to win in the long, long term without the best product. I think that you can get acquired two years in because you had a really good press or marketing strategy. But that is not how you create sort of enduring value. Eventually sort of these things converge.
[0:20:15] Sorry, one more thing. What that means also is why engineers and designers start the companies that win. And I think that is like, you know, as bad as you are at everything else, if you can just make a really great product, you can still sort of win as a startup, which is cool. So sort of on the topic of listening to users, right? What do you think is the...
[0:20:42] What's the right mechanism to that? You could argue that one way of listening to users is by conducting a whole bunch of A-B tests and seeing which one gets more people through a user flow. Or you could argue that it's going and sitting down and having lunch with a customer and finding out what really matters to them. Obviously, there's probably a bit of both, but what do you like the best? Yeah, it's definitely some of both, and there are weaknesses to both. Sometimes users are not good at explaining what they really want, and sometimes the metrics just lie.
[0:21:11] I think kind of the north star on all of this is that it's more important to make a product that a small number of users love than something that a lot of users like, you know, like 6 out of 10. It's much easier to go from that first thing to a lot of users loving the product than it is to go there from the other way. So I think as you're getting user feedback, you should focus on a small number of people really loving the product. And the way to do this...
[0:21:40] Usually involves lots of one-on-one conversations. Now, you still want to watch the metrics, of course, but figure out who your best users are, get to know them, figure out where there are more of them, and figure out how to make them like it more. I think that's a better strategy than, you know, the common way people optimize this is they go, in their early days, they go look at the users that have turned off and say, why didn't you like this? What can I do to get you to use it? And I think it's like a 10x better strategy to go to the people that are using it and say, how can I get you to use it more?
[0:22:08] Because you really are trying to get these users to love a product. Airbnb certainly did that. Great. Okay, so you've talked a little bit recently about the tech bubble, right? Or I think the way you've couched it is that you don't believe that there's as much of a tech bubble as we've seen before. I just wanted, I don't want to put words in your mouth, sorry. No, it's fine. I'll tell you where there is a bubble. Risk-free rates, that is a bubble. You know, eight years of like,
[0:22:39] Negative real rates, that from the US government, that is a bubble. And somehow, we all have agreed to ignore this kind of crazy, unprecedented thing and talk about whether there is a tiny bubble in this one segment of the economy. And it just boggles my mind that we've all agreed to this collective mass deception. So I don't think the interesting question is whether or not we're in tech valuations or 20% ahead of where they should be. Maybe, I don't care.
[0:23:06] I think the interesting question is, what about this mega bubble we're definitely in that no one is talking about and will not go on forever? Explain a little bit more what that means to you for the room here. Sure. It's well known that low interest rates and high equity prices are correlated, and there's sort of an equation about this trade-off. But you can't divide by zero in this equation. And so when you have the government selling treasuries at...
[0:23:36] Roughly zero. Everything else gets whacked. And it inflates every other asset class. If you can borrow money, basically for free, you can buy a bigger house because your mortgage payment will be lower. Rates go back up to 5%. The amount of buying power people have in their monthly payment for a home comes down a lot. Housing prices go down a lot. If people can earn a return on capital again in the bond market, then maybe they sell some of their equities and get some in the bond market. And there, I think, this is like a really long conversation, so I won't kind of go through the whole thing, but there are...
[0:24:05] Lots of other sort of more subtle ways that I think this all correlates. And we have really never seen this before. Like eight years of near zero, seven years of near zero rates. Like no one's ever seen that before. No one really knows what that means. There was a corporation in Europe a few months ago that issued a negative nominal rate bond. A corporation, not even a government. And so I think it was Nestle. Anybody remember if this was?
[0:24:35] I think it was Nestle. But basically, people are willing, rather than store money in their mattress, they are willing to give the cash to Nestle and get back a little bit less than 10 years for the privilege of Nestle holding it and not losing it. That seems a little bit crazy. Something weird is going on here. And I think it is an interesting mental thought process and one that I sort of spend way too much time on thinking about how that unwinds. All right. So another interesting topic that a lot of people have been
[0:25:07] The way that people get themselves in trouble is they invest in growth before the product is really working.
[0:25:38] This is another long conversation. We can talk a little bit about it. But you have these VCs that are incented to have you kick up your burn a lot and spend a lot of money and raise more. And they will push you to do this before the product is really ready. What works and what I'm finding, I don't hate all hybrid companies. I hate most of them. But it's okay if you first get the product right and then spend a lot of money marketing.
[0:26:04] But if you don't do that and the product is sort of only okay, then you're wasting all these sales and marketing dollars because the users don't stick around anyway. And you have these VCs that, you know, they have this fund and they invest some of it in you, but they only get their management fees, their 2% a year management fee that pays their salaries if they invest all the capital. So they want to put more capital to work. Also, the way they raise their next fund is by having you have a marked up intermediate valuation.
[0:26:32] They also want you to raise another round again. And so you have, I think VCs generally push companies to get the burn up way too high because they're trying to deploy more capital and get the valuations marked up higher. And again, sometimes this works out, but it often doesn't. And I think that it's almost never the wrong choice to get the product great first and then ramp up the burn. And if you do it the other way around, if you ramp up the burn and then pray you get the product right,
[0:27:02] That does work out every once in a while, I admit. Like, sometimes lightning strikes. But I don't think it's a good strategy. Okay. Referencing a pretty recent blog post that you did where you were giving advice to a friend about things you learned. Pretty popular post, by the way. One interesting thing you mentioned is that you succeed by picking the right thing to do. And then you follow that up with, this is crucial and usually ignored.
[0:27:32] Can you talk about that a little bit? Maybe even especially in the context of engineering? You know, so that post, this was like my life advice on turning 30. It came from this like email that I dashed off to a 19-year-old friend at like 3 in the morning without a lot of thought. And then basically the next day I like fixed the grammar and published it. And had I known it was going to get like a million views and be published in a bunch of newspapers and stuff, I would have actually spent more time on this particular topic. Because I feel like I gave a very bad answer to how you figure out what to work on.
[0:28:03] But I think this is, you know, I can give a short answer here and I will, but I'm planning to actually write a longer piece about this because I think this is a critical thing that I totally glossed over. The common advice on this is sort of do what you're passionate about. And I don't think that's necessarily bad advice, but I come back to those two frameworks that I talked about earlier. You know, the impact maximization and the regret minimization.
[0:28:32] And I think using these about any particular company decision is sometimes hard. But using these about your sort of life career path is actually easier. And if you trust yourself, if you trust your own instincts about where the world is going to go, you can actually get this decision right by thinking through these two frameworks. The problem is that most people
[0:29:02] Are wrong about where the world is going to go. And I'd say the people in this room are much more likely than average to be right. But you're also going to be susceptible to parents or experts pontificating on TV telling you the wrong thing. So I think trust yourself about where you think in the domain you understand the world is likely to go. And then sort of shape your career to get there. And you have to kind of ignore what other people say. And this is the hardest part. It's really hard.
[0:29:32] To ignore what other people say. We're all so much more susceptible to that than we think. In my own case, at this point, YC is a fairly big deal and well respected. But when I agreed to take it over, it was still sort of like people are like, this thing is a joke, it's never going to work. And I just sat there and I had a bunch of people telling me not to do it. And I sat there and I tried to think really hard about how is the world actually going to evolve and how is this particular company actually going to evolve?
[0:30:02] And I limited myself to thinking about opportunities I understood well, because it's hard to decide without those, and I had kind of four. And I used those two frameworks and ended up here. All right, let's shift gears for one second and talk a little bit about engineering culture. So if you just had to pick a couple of things that you were gonna focus on.
[0:30:28] To build a great engineering culture, and you probably can't through only looking at a couple things, but let's scope it. What do you consider most important? And also, I think it would be interesting to hear, do you think that there are any companies out there that are doing anything very novel on the culture front in terms of creating a great engineering culture that we could learn from? Yeah, I think there's a lot of good engineering management hygiene, but I think the thing that most companies get wrong,
[0:30:58] That's really important is when you have super talented engineers working together, it's easy to have just like a lot of bad interpersonal stuff happen. And the engineering cultures I respect the most work really hard to make sure this doesn't happen. And so you don't like let conflict fester. You don't keep around people that are, you know, great engineers but really toxic to the culture. You sort of put this idea that people need to like the environment they work in, the people they work with first.
[0:31:26] Most companies pay lip service to that, but then when it comes to the trade-off of missing a deadline or keeping the team happy and motivated, they won't do that. A common question that founders come to us with at YC is like, my engineers are only working nine hours a day. I want them to work 15 hours a day. I keep beating them with this whip, but it's not working. What is the trick to getting engineers to work harder?
[0:31:53] And we always tell them if they like the work and believe in the mission of the company, people will work really hard. And everyone wants some shortcut to that. They're like, well, can't I just tell them they have to sit in the office more hours? Only the bad founders ever say this. But the conversations are memorable when it happens because it sounds so crazy. And really, a good engineering culture is one where everyone believes what they're doing is going to really matter.
[0:32:23] And everybody really likes their work environment and the people they work with. And I think everything else that people say about the mechanics of management and sizing projects, that's all true. But I think those are the things that are important that people miss. Do you see anything interesting happening in companies out there today where you think people are doing things differently? I mean, Zappos has this whole holacracy thing, and I think Medium has published some stuff. Yeah, I'm so skeptical of all that stuff. I think that's all just like...
[0:32:52] I don't actually say what I think that is, but I'm skeptical. Just say it. No, no, no. I think the most interesting meta trend that is going to happen to engineering culture, it's much easier to have a really good culture in a small group of people. And things that used to take engineering staffs of thousands are now doable by staffs of hundreds and soon will be doable by staffs of dozens because the tools and languages have just gotten so powerful. And so I think...
[0:33:21] One way to make the culture better is to keep the team as small as you can for a long time. And this is something that a lot of our companies are doing. You know, like YC itself, I intentionally keep the team very small because I know that means better culture. And I think the most interesting engineering culture management style thing that's happening is this bias towards figuring out a way to keep the team much smaller than anyone thinks possible. Love that.
[0:33:52] All right, let's see. Let's shift gears again a little bit. But what threats do you think our industry is facing in terms of stifling pace of innovation or just things getting in our way in general, sort of more macro than just the tech industry? But what do you see out there that could be getting in our way? And do you think that that stuff is on a positive trajectory or maybe even trending in the wrong direction? I think there's been a growing fear of technology since 1945.
[0:34:23] And maybe it took a little bit longer for the full effects of the bomb. Or maybe other countries had to develop it before people got that. But I think there has been this anti-tech movement that has gotten a lot stronger since that happened. And sometimes it's subtle and you don't notice it. And sometimes it's like the anti-vaccinators and it seems really crazy. But for a long time in human history, there was this thing that more technology was always better.
[0:34:53] And then only when it kind of got to some point in the curve was it sometimes, well, sometimes more technology is just worse. And so I think there's been this undercurrent for a long time. I think you have that trend that's been going and doesn't seem to be stopping. The risks that technology present the world get scarier every year. So does the upside gets better, but the downside gets worse. And so that is this like,
[0:35:23] That instability, whatever you want to call it, I think is a big force against tech. And again, it's not wrong. And then the second major force is the wealth inequality one. Wealth inequality has been increasing ever since there was any technology at all, right? When maybe people had a 2x range in ability, when we were hunters and gatherers.
[0:35:49] Not just ability, but sort of like ability and just the circumstance you were born into. And that just keeps going up and up. And I think that is like I think the biggest societal problem we face right now because that is likely to continue going and the solutions that I think are likely to work about that, like, you know, I think a basic income would be really great. And I'm actually okay with, you know, 1% of the people.
[0:36:18] Producing the wealth for the world. And I think you need to find ways for everybody to feel fulfilled. But that doesn't bother me sort of by itself. But it's politically untenable to even try to get something like that done. And so you just have it causing this tension and backlash on tech. So I think those are the two trends sort of in the way of tech that keep getting worse. All right. So this might not be a great question to end on. Let's do it. At a conference where we've talked so much about machine learning.
[0:36:49] But you wrote a couple posts recently where you put forward this idea that, you know, machine intelligence could ultimately bring about the end of mankind. So, I mean, that's a scary thought. And it's one of those things that you hear it and, you know, on the face of it, you're like, ah, that's crazy. You know, machines are never going to be that smart. But then you think about it a little bit.
[0:37:12] And it's like, well, maybe actually it's not so crazy. How should we think about that? And should we be even thinking today about being more responsible in how we implement this technology? Yeah, you know, I think the crazy people are the people that think this is not going to happen. The sort of egocentrism it takes to say there is something so special going on in my brain that it will never be replicated in silicon is just like, I can't respect someone that believes that. Actually, I have one friend that I really respect that believes that.
[0:37:41] But I think where deeply religious people get particularly conflicted about this. But in general, I think, yeah, that's a good argument, right? It's going to take us a long time to figure out how the brain works, but we're going to. And maybe the whole thing is quantum coherent, but probably not. Probably this is something that we can represent pretty quickly.
[0:38:09] And we need more powerful computers, but not way more powerful. The brain has like 100 billion neurons. I think about 100 trillion synapses. If you just think about what it takes in terms of memory and processor to represent that, maybe it takes us till 2030 to get there. It's not 100 years away. And we'll figure out the algorithm. So I think, yeah, we should be doing a whole lot more to prevent a bad outcome here.
[0:38:38] If I were Barack Obama, I would be spending, I would sort of commit $100 billion to R&D for machine intelligence safety. Because right now, the answer is we have no idea. There are a few deeply underfunded groups working on sort of how to have this happen in a positive way. But we have no idea how to make this work. And it's hard to know where we are on this curve. Anyway, on that cheerful note, thank you for having me.
[0:39:07] Sam, it was awesome having you. Thanks so much. Okay, okay. I'm just going to wrap it up. That was amazing. Thank you all so much for coming here today. I just want to give a quick shout out and a huge thank you to the speakers who prepared so much awesome content today. So thank you all for everything you did. It's so hard to clap with a microphone.
[0:39:38] I also just want to give two quick individual shout-outs. There's so many people who put work into making today possible, but there are a couple of people who are really at the heart of it. Marissa Coughlin and Natasha Lowe have just put incredible work, so thank you so much.