[0:00:00] Well, thanks everybody for joining us for this, the first edition of our Clayton Community Foundation Speaker Series for 2021. Tonight, at home with Sam Altman, the entrepreneur and venture capitalist, one of the most quoted people in Silicon Valley. Now, of course, the CEO of OpenAI, but most importantly, perhaps a native of St. Louis, in particular, Clayton, Missouri. So we'll get into that, but first a word about our friends.
[0:00:28] At the Clayton Community Foundation, we started in 2008. We're a partner to the city of Clayton, focused on four things largely, art, history, parks, and sustainability. The foundation has already raised about $10 million since it started, including $2 million for the Centene Commons at Shaw Park, which will transform the old ice rink into a year-round recreational facility for people of all ages. And with the support of the community,
[0:00:58] The foundation is now fundraising to save the pocket park on Maryland Avenue, just east of the library. And that's going to be very exciting. Stay tuned for more details about that. The Clayton Neighborhoods video debuted in December, just last month. It features St. Louis County historian Esley Hamilton, who's sharing insights on eight neighborhoods.
[0:01:21] The first video highlighted the Hillcrest neighborhood, and subsequent videos are going to take a look at Carswell, Beamiston, and others. The foundation is also working on a historic site, rather a historic marker program, and the first one, which will be unveiled soon, features the old attic school, which was Clayton's second school catering to...
[0:01:42] Young African-American students. And if you haven't seen the Clayton AutoCast tour, please do so now. This is an audio tour taking you to public art and historic sites in Clayton. The Ralph Clayton Society honors donors who have given $1,000 annually. And you can learn more about that and other great programs with the Clayton Community Foundation at claytoncommunityfoundation.org. Okay, so much for the preface. Let's get to the heart of the program. Sam Alton, what can I say?
[0:02:11] He has recently partnered with Elon Musk to start OpenAI, researching artificial intelligence. Just got a billion dollars from Microsoft to do just that. Was president of the Y Combinator. This is an accelerator that nurtured and funded little companies that became big companies like Airbnb, Dropbox, Stripe, and Reddit. He was the founder of Loot. Started that at the age of 19.
[0:02:39] And then sold it at the age of 26 for $43 million. Forbes Magazine in 2015 called Sam one of the 30 under 30 in venture capital. Businessweek described our guest as one of the best young entrepreneurs in tech. And as I said earlier, one of the most quoted people in Silicon Valley, Sam Altman. Thank you so much for sharing your time with the CCF this evening. How are you? I'm great. Thank you for having me.
[0:03:05] Well, it all started in Clayton. What do you remember cutting your teeth in Clayton, Missouri? I was just thinking about that. It was sort of like this idyllic place to grow up. We had like a gate out of our backyard and there was like a hole cut in the fence of Captain. I remember like very proudly the first day that my parents let me like walk to school by myself through that little like up an alleyway and then through this sort of hole in the fence and just sort of like goofing around in the backyard. And yeah, it was very lovely.
[0:03:35] Did you haunt Caldees? Yeah, a lot. Still do, actually, last time I was in town. And at Captain's School, I understand at the age of eight, you had your first computer. So in the second grade, you were becoming computer literate. Is that accurate? Yeah, yeah. It was actually, it was in some sense like an easier time then to get started with programming. You just sort of, we had these old Apple computers that would, 2GSs that would just sort of like boot up immediately into this like very simple programming environment, so.
[0:04:05] It's sad for me that kids don't have that in quite the same way today, but I think there's some companies that are trying to make it easy again. We got one at home when I was maybe in third grade, something like that. Did you take it apart and put it back together? Yeah. I actually just bought the same computer that we had at home, which was a Mac LC2. I found one on eBay.
[0:04:32] And I was amazed how much I remembered about the actual hardware of it. So that was fun. I hadn't seen one in like 25 years or something. And from the Ralph M. Captain School, you took your studies to Y Down Middle. Yes. Most people don't like middle school. How about you, Sam Altman? I mean, it wasn't perfect, but it was like pretty good on the whole. I wouldn't want to go back to middle school, but yeah, I have, I think, fond memories at this point.
[0:05:01] Now, of course, you're the CEO. High school was great. High school was John Burroughs, of course. Yeah, I think high school is easier than middle school. Well, talk a little bit about your current mission at OpenAI. I am correct in saying that you're partnering with Elon Musk in this capacity. We started OpenAI because we thought that this thing that was happening of us.
[0:05:28] Building digital intelligence, us like humanity collectively building digital intelligence, is going to be one of the most important milestones in human history. And it can either go really well or really badly. And we did not think there was enough effort to make sure that it happens safely and in a way that sort of humans broadly benefit. It's very hard to think about what the world is going to be like when we have superhuman capacity inside of a computer.
[0:05:57] Computers that can learn anything, that can think billions of times faster, smarter than the smartest human on every topic simultaneously, and that eventually sort of become self-aware and have their own desires and will and none of the limitations that humans have. And that is such a hard thing to think about that I think most people just don't, or at least not for very long. And we wanted to start an organization that was dedicated to kind of getting across that event horizon.
[0:06:25] Safely and in a way that humans as a whole were as happy about as possible. I think this is going to be a bigger technological revolution than the three great ones so far, the agricultural revolution, the industrial revolution, the computer revolution, all put together. Because fundamentally, our phenomenal ability to think, to learn, to reason, to generate new ideas, that's what makes us human. That's a huge part of what makes us human. And when a computer starts doing that, which is happening more and more every year,
[0:06:55] Everything is going to change. So we started this org to see how we could contribute. Now, science fiction writers have written about the harm that artificial intelligence can cause, Terminator 2, that sort of thing. I mean, do you fear that a computer that can do everything that a human mind can do and more could possibly pose a threat to civilization? For sure. The most important research we do is research around safety.
[0:07:25] So how do you build these systems? When we started, it was maybe how do you build these systems at all? And now that we're more confident that it's quite possible, our most important research is how do you do this in a way that aligns with the collective set of human preferences? How do you make sure these systems act in the interest of humanity that manage to be an extension of individual human will while also still like following the rules and not using their immense power to sort of like...
[0:07:53] I'm more optimistic now that that's going to be possible than I was when we started. We've had a good last couple of years of safety research, but I think it's still quite challenging. And yes, I think the sci-fi writers have been right to be fretting about this for the last 50 years. It's interesting to go back and read the sci-fi sort of each decade as we learn more and more about what...
[0:08:23] How the systems were going to work and kind of what the feel of the intelligence was going to be. I think a lot of sci-fi still portrays that it's going to be very human. And my guess is that it will be quite intelligent, quite general purpose intelligence, but in a pretty alien way, where it will do some things that are just wildly better than what humans can do and still struggle with some things that we do. And so it's been interesting to watch as we've learned more and more about the technical roadmap.
[0:08:51] How the sci-fi has updated every decade, and now everything's moving so fast, how it updates every year. Well, I'll tell you what, when it comes to language, it's updating really quickly, almost to the point where some computer-generated voices and language sound identical to people down the street, Sam. Yeah, voice has been there for a few years, but really last year we released something called GPT-3, which could actually come up with the language, the sentences.
[0:09:21] That in many cases are human-like. And I think in a few more years, in many cases, we'll be superhuman. You won't be able to tell if something was written by an AI or a human. I'm excited for this. I think we will be able to interact with computers in ways that have been only in the realm of sci-fi when we can talk to them in natural language and they can really truly understand what we want them to do and manipulate language sort of as flexibly as we can.
[0:09:50] I think that is going to lead to a whole different way that we interact with computers. Really, truly understand us. The jobs question is tough. And I'll tell you why. So jobs change by technology all the time. If you look at this back the last several hundred years, on average, about 50% of the jobs turn over every 75 years.
[0:10:20] Because new technology comes along. And we can adapt, at least at that pace. If it was 50% turnover in five or 10 years, that would maybe be hard. But when we think about policy to address the unavoidable changeover of jobs, it's hard for two reasons. Number one is that what the new jobs of the future will be is quite difficult to predict. If we were sitting on the other side of the computer revolution, it would have been very hard to predict the job of computer programmer.
[0:10:50] But also because technology itself often doesn't go the way you think. If you had asked me 10 years ago to predict the order in which AI could do human jobs, I would have said first blue collar labor, then white collar labor, and then finally like the last thing humans would hold on to is creativity. So music, art, things like that. And to my great surprise, I would have been exactly wrong. And the order in which AI is sort of rolling out is,
[0:11:20] Creative applications first. So already it's quite good for making visual art and music. White collar jobs second. So people doing AI lawyers, for example. And then it turns out that robots are just really hard. And so doing physical things in the real world, I expect that to be last. That's actually fascinating because you're right. People think just the opposite.
[0:11:50] Of course, the business applications are pretty obvious, though, because now companies will be able to use artificial intelligence to lower the payroll, among other things. Yeah. Again, as AI sort of joins the workforce, I think we will see that unfold in lots of ways, both wholesale replacing jobs, but more commonly, and I think this is going to be better for all of us, and AI will do...
[0:12:19] The 90% or the 80% of your job that you like the least and give you much more leverage to do the 10 or 20% that is hard for an AI to do that you really enjoy. Do you think there's any governmental applications? I mean, recently we've seen where the government's ability to handle a crisis is limited, shall we say. Do you think that artificial intelligence can maybe tackle problems like modern day pandemics in a way that might be better than human beings have?
[0:12:48] Yeah, I think for very complex coordination problems or where you have to do like a lot of sophisticated modeling, AI today can probably make much better decisions than humans can. It's just too much to think about all at once. So what was the interest in Microsoft in partnering with you with a billion dollar investment? By the way, that's a billion as in B, baby, oh baby. That's a lot of money, Sam. The sort of, I was going to say the dirty secret, but I don't even think it's like.
[0:13:20] The surprising thing is that a lot of AI progress over the last few years has come from not coming up with smarter and smarter algorithms, but just running them on bigger and bigger computers. And so Microsoft is incredibly capable of building super large supercomputers that span multiple data centers. And we partner with them to do that.
[0:13:48] We're interested in doing the research to try to build this general purpose intelligence. They're interested in taking it to create products. And so that's like a nice partnership for us both. What do you see in the near horizon, how artificial intelligence through the work of OpenAI will manifest itself that people on the street might recognize?
[0:14:15] The one that I think some people have seen now and is a good preview of what is to come are some of the self-driving features on cars. It's far from perfect, but it can do amazing things. And once in a while you see this glimmer of like real reasoning in the newest self-driving systems. But I think the next, another big one soon will be that you will chat with systems like either via text or via voice.
[0:14:44] And you'll be talking to the computer and it'll be so good you don't realize it. So maybe a customer service call? For sure. Very soon. How about world leaders? Instead of speaking from a lectern or a podium, just having a computer generate a speech for them with some sort of electronic signature? I think that electronic signature is going to be important. I think another thing that has started to happen but is going to happen much more in the next couple of years is sort of the extreme version of DeepFix.
[0:15:14] Where there will be videos of world leaders saying anything you want, incendiary stuff, you know, claiming to have bombed some country on Twitter and they'll be totally fake. But it will be, the AI will have gotten so good at generating that, it'll be almost impossible to tell without some sort of cryptographic signature. So I think that's gonna become really important. Sam, before AI, OpenAI, you were with Y Combinator, a Silicon Valley based company that,
[0:15:43] Nurture, coached, funded small companies that became big ones, such as Instacart, DoorDash, Stripe, Airbnb, Reddit, and Dropbox. In fact, you were there with the founders of Airbnb on the very early days before they were a household. We're in America, weren't you? Yeah, it's a really wonderful program. I didn't start that one, but I did run it for a while. I took over from...
[0:16:12] The people that founded it. It's sort of like a school for startups or sort of like a replacement for business school, except it's only three months long and we give you the money instead of you paying us tuition. But we have tens of thousands of companies from around the world apply. We would choose about 400 a year to fund. And these are just a handful of people and an idea and maybe a prototype.
[0:16:40] And it is amazing how quickly they can make progress. And then we would sort of coach them, help them, and then later connect them to investors that could write larger checks than we could. But that was the model. It was just sort of find the most talented but unknown entrepreneurs in the world and give them some money and help them as much as we could. Now, if you took a look at the market caps of all the Y Combinator alums and added them up, do you know what that number would be?
[0:17:11] Good question. It'd be a lot. We only own a tiny, tiny slice, maybe like 400 billion, something like that. 400 billion, maybe. What was it like? What were the founders of Airbnb like, for example, before they were rich and famous as it were? One of the things that's surprising about Y Combinator, the Y experience of that is that how quickly the founders of these companies go from being sort of like very...
[0:17:43] You can see a physical transformation in the way they walk into a room from the beginning to the end of the three-month program.
[0:18:12] I'd say people start off on the whole like absolutely unremarkable and then very quickly through the challenges of having to run this company become relentlessly determined and that really comes across and it's very fun to watch. I think it's probably like being a teacher in some sense. Are you the one who said for a young startup it's better to have 100 people who love your product or service than 100,000 people who like it?
[0:18:42] Yeah, this is like a deeply counterintuitive thing. And it's so surprising to me how true it is. The explanation that I would give is it's much harder to build a product that people truly love than it is to get more people to love it once you... And so most startups want to like scale out and they take this thing that's okay and say, oh, we'll just get more people to do it. And that gives us sort of like more total thrust.
[0:19:12] But the other way is to keep the same number of users and just make the product much better and then expand from there. And almost without exception, if you look at our best companies, this is the way they go. They first focus on the small number of users, make the product incredible, and they know that once they do that, basically no matter what else they do, it's going to grow. I know you've been asked this question too many times, but humor me, will you? What do you look for in a startup?
[0:19:44] So there's a lot of things. I'll try to talk about a few of the non-obvious ones. One is just, can this founder become this relentlessly determined person that it takes to sort of face all of the trials and tribulations of a company over a decade and continually learn and adapt and rise to meet the occasion? So that's like a big one. Another big one is, is this person a fast mover and a fast improver? That for me personally is one of the biggest.
[0:20:16] And that's very hard to tell in a single meeting. But if you get to see them two or three times, and you can note how quickly they're making progress and doing the things that you discuss in between meetings, you can get a great sense of that. And so the rate of improvement and the rate of action is a huge thing that I look at. A third is how founders end up having to be the chief evangelist for the company. You have to recruit people. You have to talk.
[0:20:43] To investors to raise money, talk to the press, sell to customers. And can this founder kind of be this infectious evangelist? Does this founder make me want to join their company? And if not, that's sort of like fairly hard to teach. It's hard to like really authentically fake the level of passion for a particular thing that takes to get there. And then finally, like, do I think that the business makes sense? Do I think the market is going to grow quickly? Most investors want a market that's very big right now. I think that's a mistake.
[0:21:13] I'm a big believer in you want a market that is small today but gonna grow quite rapidly. Quality of the product, I guess that's it. That's the short version. You once said, and I'm paraphrasing here, that you noticed that successful startups are headed by people who didn't grow up in the richest households or the poorest ones. Can you kind of follow up on that? Yeah, I don't.
[0:21:43] It was more like an observation than a diagnosis of the reason. But like looking around at most of the successful founders we funded, they kind of grew up upper middle class and not with very little deviation in either direction. How did your parents feel when you dropped out of Stanford at the age of 19, Sam Altman?
[0:22:09] Well, a cool thing about Stanford is that they let you go back no questions asked for a long period of time. I forget exactly how long, but a long time. And so I sort of was like, oh, I'm going to do this. I have no idea what's going to happen. I don't remember exactly how they responded. It wasn't bad. They were just like, OK. Like, are you sure? But at the very worst, it was like, are you sure? And maybe it was like, that sounds great. I don't remember. It wasn't. It was just like, OK.
[0:22:37] That's kind of a blessing for a young entrepreneur, isn't it? Huge blessing, huge blessing. And okay, so you started a company called Loot and you were 19 years old. Loot kind of identified where people were, right? Yeah. And you had seven years with that before you sold it at the age of 26 for $43 million. Yeah, it didn't go quite how we wanted, but I think a good thing about Silicon Valley is
[0:23:07] It's okay to just say, well, we learned something and onto the next thing and we'll try to make this next one better. Boy, it just, there seems to be a pattern here between you, Bill Gates, Steve Jobs, Michael Dell, a lot of guys who dropped out of college. I think college is, there are things about college that are very good. But I do think, and you know, I think this has now become like a, especially in the last year, a more mainstream belief is it is,
[0:23:39] Given the costs and amount of time, it is easy to imagine that there are better ways to spend those four years. I certainly felt like I learned. I loved the time I spent at college, but I certainly felt like I learned more in the two years right after college than the two years during. Now, Stanford allows you to go back. Are you thinking about doing that someday? I can't go back anymore.
[0:24:09] There's a limit as to how many years you apply. There is a limit, yeah. Something tells me though that if you apply, you might get admitted. I don't know. Maybe. There might be some Stanford connections in the Clayton community who can pull a string or two for you, Sam Altman. A couple of questions in no particular order of importance, if you don't mind. At one point, you were the CEO of Reddit for 10 days. What was the story there?
[0:24:38] We had a CEO quit during a board meeting, which is like a very rare thing to happen. Never seen it before or since. And it took us a little while to figure out the succession plan. We were caught totally off guard. So everybody, you know, the management team sort of, yeah, technically reported to me. I had joined, I was a recent investor and joined the board and kind of was the only person local. So I got very involved for a few days.
[0:25:06] We had an interim CEO, and then after that, the original founder returned to run the company. Well, I hope you agreed to a nice pension when you left the firm. I did not get one, but that's okay. It's very enjoyable. People talk about the diversity or the lack thereof in Silicon Valley. I haven't been out there. What is the demographic makeup of Silicon Valley? Unfortunately, it mostly looks like me.
[0:25:37] There has been a lot of effort to diversify Silicon Valley and there's been moderate success, but not enough. I think it's like a hard problem and it's like hard to talk about without sort of it, you know, devolving into like a lot of fights and name calling. I think it is finally on a better trajectory and we're making material progress, but it takes a long time.
[0:26:05] To get these changes to actually sort of propagate all the way through the system because it sort of changes what you, it really needs to change what you sort of study in high school and college. And then it takes however long it takes you in your career to be able to become an investor. And then I think change really happens. So I think it's moving in the right direction, but it's been slow. Can somebody fund some scholarships for minorities? There's a ton of that.
[0:26:35] That has not moved the needle enough. I think what is having a better effect are dedicated efforts around mentorship and startup funding. But even that, it's a very long cycle. How about some of the folks from St. Louis who have pretty decent presences in San Francisco and Silicon Valley? One that comes to mind is John Doerr, D-O-E-R-R, big venture capitalist.
[0:27:06] Another would be Jack Dorsey of Square and Twitter, Jim McKelvey also. Do you ever run into the diaspora from St. Louis? I do, yeah. I see them all occasionally, and we always talk about St. Louis. Do you think that St. Louis could possibly be a tech startup hub or a Silicon Valley of the Midwest, anything in that order?
[0:27:32] Sort of what I think is going to happen is that startups and tech are just going to disperse all over the country now. I think like, A, it's just an interesting field. B, the Bay Area has been like horrifically badly managed and the cost of living has just gotten super high. So I don't think that there will ever quite be like another Silicon Valley with the same level of density, but I think there will be lots more startups and technology everywhere.
[0:28:00] Yeah, it has kind of moved around, hasn't it? Because before you were around, there was the Route 128 corridor in Boston. And then that kind of moved to San Francisco or Silicon Valley. And now, I mean, from what I read in the paper, you know about this better than I do. A lot of firms are moving to Austin, Texas, Sam. I still don't know what to make of that. Again, I think like there's just this drive out of California right now for a bunch of reasons. How long and how far that goes.
[0:28:30] Where that settles and if it's one place or just sort of a scatter everywhere, I think we really don't know. Now, would you consider starting an office in Austin or St. Louis? We've decided that for now, like we're actually a relatively small organization by number of people that work there. So we've decided it doesn't make sense for us to have multiple offices. But yeah, it's something I could imagine us doing someday if we were like five times bigger than we are now.
[0:29:00] And sometimes people are looking at San Francisco these days and they see that, at least in the pandemic, it's not the city that it was. The price of real estate is actually going down a little bit, we're told. I don't know if that's quite possible. But is it possible that San Francisco's warts will actually discourage people from moving there? For sure. I mean, it even discourages me from living here. It's really like quite a shame, I think, the state that the city has gotten into. How so?
[0:29:29] I think that the lack of success that the city has had dealing with homelessness and drug abuse and crime and sort of just general unpleasantness and broken infrastructure, like that's all basically true as reported. And it's really just sort of depressing and sort of inhumane. And it just doesn't feel good to be a part of it. I think we're all hopeful it'll get better, but it hasn't.
[0:30:00] It's been trending worse. We're getting some questions via text. We encourage everybody to do that. Plus, we also have some questions out there that we're going to get some answers to. We had some Zoom poll questions for those who've joined us with you, Sam. Someone just suggested that you finish your degree at WashU. Okay. I barely have time to even read books anymore. I actually would love to just...
[0:30:29] I have a couple of years to study, but it doesn't seem super likely. And here's some of our poll questions. What would you like artificial intelligence to do? Give medical advice, drive your car, teach you new things, solve pressing world problems. I'm going to guess, Sam, that you would probably say solve world problems. Yeah, I'll pick that one. Okay.
[0:30:55] And our other question, and by the way, folks, we're asking you to vote on these now, are you willing to share your personal data to make your experiences with technology more tailored to you? Yes, probably, or no, I'm not sure what the last possibility there is. So we'll get the results of that and we ask people to respond to those accordingly. Sam, I don't wanna get too personal here, but you once credited your mom with being one of the
[0:31:25] Big reasons for your success. Very true. She was not, she's not a tech pioneer. She's a dermatologist by trade, right? Dermatologist. But what was it that your mom did that was so helpful to you? Oh man, so many things. I mean, I think like fundamentally there's like all of these things that are like luck of the draw of how you're born, but having good parents who are like, are very sort of.
[0:31:50] Supportive, who sort of make sure you get a great education, who sort of like help you discover what you're good at and what you're going to be and sort of whatever that is, help you do more of it. That's probably like the most unfair advantage you can get in life. And I had a, you know, outlier like 99.99 percentile mom in terms of that. And...
[0:32:16] Yeah, I don't think people ever acknowledge like of all of the injustices in the world, like what a great thing that is to have. But my mom is super great about that. Where did you develop your love of learning? I'd say it's more like a curiosity and I don't know where that came from. You always had it. I mean, if you're taking apart computers at the age of eight, I guess so. I don't, I actually didn't. Yeah, I don't remember. I didn't love school when I was a very little kid. I eventually did.
[0:32:58] Sort of like academically, I would say somewhat of a late bloomer. But I was always like quite curious. Okay. And we have some answers to the poll, I guess. Most people would like artificial intelligence to solve world problems. So they're agreeing with you on that one. And are you willing to share your personal data, name, location, et cetera, to make your experiences with technology more tailored to you? 69%.
[0:33:29] Say probably, but would like some control over my privacy. Are you concerned about privacy? Yeah, I think most people, and me too, I think net out to that answer that was the top of the poll answer, which is that as long as I can have some control, I'm willing to trade some privacy in exchange for better services. But AI is going to push that so far. I think if we look forward five or seven years from now, we'll all basically have like an alter ego AI that listens to everything we ever say.
[0:33:58] It reads everything we ever type. It knows everything about us. And in exchange, all of the information we could ever want is right there at the moment we need it. But the depth of the trade-off of privacy for utility there is going to be bigger than anything we faced before. And I'm not entirely sure how people are going to think about that. Do you believe that we can build a machine that can do
[0:34:28] Everything that a human brain can do? I do. How many years are we away from that, Sam? That part, I don't know. I think that sort of nothing that is not prevented, anything that is not prevented by the laws of physics, we will eventually figure out how to do. And in a human brain, everything you are, everything you think, every memory you have, every part of your personality, everything you know, is stored in about 100 billion neurons in your brain.
[0:35:01] That have about 100 trillion synapses in between them and are kind of connected and trained. They learn in this very complicated way. And then, you know, there's also like we could sort of say there's like some hormones having an effect on neurons that we need to model as well. But for the most part, it's just energy flowing through this network of 100 trillion connections. That's it. That's you. And remarkable complexity comes out of that.
[0:35:29] We will eventually replicate that same thing artificially in a computer. We still have a long way to go, like 100 trillion. It's massively big, even relative to our biggest computers of today. But we'll get there. And we already have one algorithm that can learn. Like, this is a miracle. The big miracle, I think, of building AI is already behind us. Once you have an algorithm that can learn everything else, including eventually, I think, self-awareness emerges because the sort of smarter...
[0:35:54] If a system is going to be really smart, it has to understand every other agent in the world, including eventually itself. And so I think we have this one thing from which everything else emerges. And we know that at a fundamental level, this idea of a neural network, whether it's biological or inside of the computer, can seem to produce and learn and remember almost any behavior. So we have a lot more research to do and a ton of engineering. But yeah, we'll get there. How long? I don't know.
[0:36:24] That's really hard. What about love? So this is a big question. I think all, eventually, I think things, the full range and depth of human emotions will eventually come in these systems, but that may not be in our lifetimes. Okay, questions from those who are following us tonight. We thank everybody for joining us. Here's one. What will be the impact on the human brain if we don't have to work it the same way that we do now? Will our brains be impacted?
[0:36:58] So yeah, in the short term, I think probably people growing up today are already having their own, what we call wet neural networks trained differently because they're faced with different things and sort of what you study in school is different, how you spend your time is different, what you work on your job is different. Memorizing facts is not as valuable as it used to be. So I think that change is already happening. A longer term question about the impact on our brains is what happens when we connect them to computers?
[0:37:28] A company called Neuralink, also started by Elon in San Francisco, is working on connecting human brains to computers. And although it's quite early, they're finding that even with their very rudimentary device, they can have a huge impact on changing the human brain. Today, mostly by sort of producing neurohormones on demand, but they can also cause certain neurons to fire and not fire.
[0:37:56] I think the line between human and machine is going to get quite blurry around the brain. Do you think that young people with constant use of cell phones are already rewiring their neurological networks? 100% guaranteed. And I would guess, not in a good way, although maybe this is like sort of always, you know, it was better when I was a kid kind of thing, and maybe it's fine. But for sure, it's very different. Okay, let's get another question here.
[0:38:26] Could we be more creative if we don't have to do other stuff? I believe that that is almost certain to be the case. Every time we have been able to free up more of our collective brainpower and capacity to think about newer and more interesting problems by removing the sort of base things we have to do, we have stepped up to the occasion and done wonderful things.
[0:38:57] Wondering if Sam read The Crucible by James Rollins. For a non-tech person, open my eyes on things under discussion. It is in my book stack, but I have not read it yet. I'm looking forward to it. Also, any recommendation for reading? I don't think any of the books about sort of general purpose AI and what's going to happen are great. I think a book called Superintelligence.
[0:39:31] By Nick Bostrom is one of the best. I think the best, I think there's like many good short stories. Two that just came to mind. One is called The Last Question and one is called Crystal Nights. I think Crystal Nights about sort of AI sci-fi. Excellent. Well, hopefully somebody's writing these down so we can share them. How about this? Some of your favorite places when you're back in St. Louis, any restaurants?
[0:40:08] Sadly, Coffee Cartel closed, but the mocha frap at Coffee Cartel was basically like my favorite item at a restaurant anywhere for many years. No kidding. Yeah, they closed just about two years ago. Yeah. I might concur with you on that one. What are your thoughts on Japanese robots meant to replace human emotional interaction? I find it uncanny valley and uncomfortable, although at some level I realize.
[0:40:37] That's hopefully not replaced, but that they are going to supplement human emotional interaction. Probably what I think the bigger thing that's going to happen is are people, something closer to the movie Her, where people talk to chatbots and really develop these super deep emotional connections, not to a physical robot, but just something that they're texting in the ether. And that I expect to happen fairly soon.
[0:41:06] And I have very complicated thoughts about that, but I think it is probably unstoppable. Do you believe that artificial intelligence should not try to look like a human being? I don't have a strong opinion about that. It doesn't seem to me to be the most important problem, but I could easily be wrong, and it really is quite significant. Without getting in too much detail, give us a little insight into the one and only Elon Musk.
[0:41:38] What do you want to know? Well, you're probably one of the few people who's actually, I don't think people around St. Louis work with him, know the man. Tell us what it's like to be his partner. He's super smart, really cares about sort of being maximally useful to humanity. Surprisingly funny, which it's just like, actually like laughs a lot and like quite funny, much more so than you would sort of think given the seriousness of.
[0:42:13] How much he works. And just sort of like, yeah, fun loving, pretty jolly guy. Super good instincts on what is possible versus impossible when it comes to sort of engineering the limits of physics. Definitely one of like, just such an outlier for capacity to work. Another question for you.
[0:42:44] Do you believe AI is possible without the next big unlock in computer technology, quantum software, hardware, co-design, et cetera? I do. I do, but I think that is not a majority opinion. I have heard the theories about why there's some weird quantum coherence going on in the brain for consciousness that we can't get to with.
[0:43:14] Classical systems, I think it's unlikely personally. I think we may need bigger chips or faster interconnects or both probably. But I think we're on the linear path, not the kind of we need a new miracle path. Do you get stressed? Sometimes, but like one of the advantages of working, being around startups for a long time is you realize that like,
[0:43:47] Almost nothing is as bad as it seems. And even if you sort of don't know what to do in the first moment, you can figure it out and evolve as you go. I definitely do get stressed thinking about like what's going to happen if we like really screw up AGI. But it's more like that kind of like long-term existential stress rather than the kind of like short-term I'm annoyed at this person or, you know, like something's blowing up at the company.
[0:44:16] That happens sometimes too. The New York Times had a story on you and others out of Silicon Valley that a number of people in your neck of the woods are preparing to live off the grid if necessary in case of some eventuality. And the term used to describe people who do this is prepper, which I don't completely understand. Maybe you could fill me in. It means like being prepared. So that's where prepper comes from.
[0:44:46] I think we used to all get made fun of a lot more before the pandemic struck. And then people were like, oh, maybe we should have like, you know, put slightly more mental energy into this. I think technology is just going to make the world increasingly high vol and a little bit of effort to think about like, okay, what are the kinds of things that could go wrong like a new pandemic? And how can I put a little bit of effort into being ready for that is super reasonable.
[0:45:16] The characterization of the press, like looking for another reason to hate on tech people, which I don't blame them for, I think bears like very limited connection to reality. Yeah. I once interviewed an associate conductor for the Boston Pops, Harry Ellis Dixon, who described the press as unnecessary evils. But you've done pretty well when it comes to press coverage, Sam. Oh, I would say terribly. But again, I think that's okay. Like we're, I think like, you know, tech.
[0:45:46] Got to run around unaccountably for a while and that was bad. And I think like a tough press is like a necessary pain. But I think that kind of, sure it goes too far and it's like, you know, too reflexively negative, but some accountability for tech is certainly much better than none. How much do you work each week on average? And what do you love to do when you're not working?
[0:46:17] Not a crazy amount. I'm not one of these like 100 hours a week, so I can't do that. And I don't understand the people who can, and I choose to just believe they're all taking Adderall or something. But I wish I could do that, but I can't. I probably work like a productive, I don't know, maybe like a pretty productive 60 hours in a week. And what do I like? I don't know, hang out with friends, go hiking, be outside, exercise, read books, cook, eat. Nothing too exciting.
[0:46:50] Do you think most people have no idea how hard successful startup business men and women work? Yeah, I think in the first few years of getting a new business off, actually, no, just all the way through, I think most of these people tend to work super hard. I was once told, if you don't want to work hard, don't start a company, buy a franchise. Interesting. You get to work much less hard then?
[0:47:22] You could buy a franchise that's like a bunch of great clips. You're not cutting the hair, but you kind of own the place. So the franchise is for the entrepreneur who doesn't really want to put in 60 to 100 hours a week. Interesting. Yeah, startups are tough. We always try to advertise this to people that were thinking about applying to YC. But if you are successful, the degree to which it will take over your life, I think most people underestimate.
[0:47:51] Oh, absolutely. And do you have any favorites that you nurtured at the Y Combinator and that really benefited from your expertise in coaching and that you're proud to say, wow, look at them, a great American success story, or maybe a great Israeli or European success story, because I know many of those who went through your boot camp are from all around the world. But any that you are especially proud of to see how well they've done? You mentioned my mom earlier. One thing that...
[0:48:20] She was always very good about, despite our lobbying efforts to the contrary, was never picking a favorite child. I will not pick a favorite child here. But categorically, the companies that are working on kind of what I call the hard tech problems, nuclear energy, sort of infectious disease, stopping infectious disease, going to space, those are the category of companies that I have the most fun with.
[0:48:50] Nuclear energy has typically had some issues, such as where do you dispose the waste? But by and large, it's clean, it's cheap, and it sure would be nice if we could have more of it. It sure would be nice. So you're a proponent of it, you are? Huge, yeah. And I think we're going to get some new progress there in the next few years. One of our companies called Oklos submitted an advanced reactor design to the NRC that sounds like it'll be accepted, first one ever. It's pretty cool.
[0:49:19] And of course, the big question, were you to dispose the waste? Have they addressed this? So one of the things that I think is really cool with the new reactor designs is how much you can reuse waste back as sort of new fuel, B&B processing. And there may still be tiny amounts of waste that need to get stored, but I don't think we're kind of, no one's advocating that we, well, some people are. I am not advocating that we just sort of scale up the old existing designs and generate a lot of waste. I think there's just much better ways to do this now. Sam?
[0:49:50] You have been too generous with your time with your native Clayton, Missouri. This has been a real pleasure. The pleasure was ours, and we thank you so much. On behalf of everyone with the Clayton Community Foundation and the entire city, we thank you for the time you shared with us. If we can be helpful in any way to you, please let us know. Best to all the Altmans, and as they say, to be continued. Thanks for having me.
[0:50:14] Thank you so much. Sam Altman, ladies and gentlemen, thank you so much for joining us. We sure appreciate it. We'll be back next month with another installment. So please stay tuned by following us at claytoncommunityfoundation.org. Have a great evening and go Greyhounds or what's Burroughs, the Rams? The Bombers. The Bombers, go Bombers. Thank you so much.