Marc Andreessen

21:11And then there's also a long history to it which is Silicon Valley really started in like the 1920s, 1930s even with technologies before the computer. So it's running on like a hundred years of what you might call a network effect. where it was basically meaning the next really bright person whose technical, technologically oriented is more likely to come to the place where all the others, more people are than to go anywhere else. So there's like a positive feedback loop that kind of just keeps spinning. And then quite honestly I think there's just something really, I mean, but California has its problems and they are profound.

21:41But there's something magical about California. There has been something magical about California that predates the entertainment industry, the predates, you know, the tech industry and it's basically the California is the frontier and the same mentality that led to the original settlement. California was the last sort of the furthest thing to the west that was settled. It was ungovered for a long time. You had the gold rush. The Wild West. Wild West. What happened was there was this just like selection process where if you were oriented around status and respect, you wanted to succeed on a geese coast in New York and Boston.

22:14And if you wanted to go carve your own path and create something new, So you went to, you went basically as far away as he's coast as you could get and that's basically right right right right right right now. You could call it like creative, you know, people experiment, right? Like they blaze new trails in like every way, including like how you live your life, you know, California is famously the home of, you know, thousands of cults, you know, I've been developed here. I think it's actually no accident. The California is sort of a, we call it sort of the stack, you know, California builds the technology of dreams.

22:44And then we also, you know, Hollywood and entertainment the entertainment business, you know, actually makes the dreams, right? And so it's like this sort of integrated dream factory. You know, is it ever really quite real? You know, Los Angeles is like famously a fake city, right? It was just like a desert. And then they just basically, it was the Theranos of cities when it first got started. And they ran newspaper ads in East Coast cities with like drawings of like palm trees, like, you know, this lush paradise. And then people would buy plus a land and they come out here. And it was just desert.

23:06And then they famously had to go get the water from the central valley. And that led to... You lie in palm trees. Yeah, but it turns out, it turns out the palm trees are imports. Like, if I had to write it, it turns out palm trees. This is actually one of my big breaks through moments in understanding California, which is the palm trees are not native. It's just like this made up, basically. The most iconic kind of thing for California, basically, is just like a made up import. And so there's an artificial... You mean American idea, though? Yeah, very much. Yeah, well, and here we sit in Shangri -La, right?

23:33Like paradise, right? And the weather. Like the weather. Like if you have a choice fundamentally, or you can live, wouldn't it be nice to live someplace whereas like 70 is on the day every day. Like it would be awesome. You know, it's the thing the Bay Area and L .A. kind of have in common, which is kind of the spirit of adventure. You know, the original movie industry people came out here for two reasons. One is they were fleeing Edison's patent enforcers. He had patents on movie recording equipment projectors. And so he was sending the picker to like break your sprake up your studio if you weren't paying him his patent fees.

24:03So they came 3 ,000 miles away to get out of the orbit of, you know, Eastern power. And then they also came here for the weather because they can film you around. And so that spirit, you know, that spirit exists. And it's actually an exciting time for that because for a long time the sort of North and South, you know, California were sort of pretty starkly divided and you didn't have a lot of crossover. I mean, you know, a lot of your predecessors in music just never had any. Your technology is a threat if they even thought about it at all and had very little interest in what was happening. And quite frankly, by Sversa, and there's just a lot more crossover.

24:31It turns out the valley is more creative than we thought. It turns out LA actually is when people down here who are actually quite interested in tech, have done a lot of tech. And so there's a real magic happening. But having said that, California also has all the downsides. And Silicon Valley also has all the downsides of a place where basically people are making up dreams from scratch. You know, there's dystopian elements to it, right? Like, San Francisco, real city has always got an interesting question because like, you know, they no longer arrest criminals.

24:59How's it working? Right, like you might just get likes to have and kill walking down the street. But a friend of mine got a job working for actually OpenAI. And he moved to be close to the office. And he got an apartment of block away. And they're in the mission in San Francisco, which is sort of famously the hub of AI, but also just incredibly violent. And basically, Laws aren't enforced. And he said, oh, I love living near the office. He's like, I wake up in the morning, I go to the stairs. I'm block away from the office. The key is I run at full speed as fast as I can from my apartment to the office.

25:27And then at night, I run as fast as I can. Because I'm trying to not get like, basically, stabbed or killed on the way in. And so it's got both of those, The drug thing has always been, big deal, who listen to John's, or a big thing, up north. And it's the good and the bad. They're opening people's consciousness in horizons. And there's a cultural creativity flowering thing that's happening just like it happened in the 60s and 70s. But there's also the downside, which is you see people whose lives are getting wrecked by drugs. And so it's also got that side of it. And so it's got this very organic, it's just like this perpetual cycle of cultural creation.

25:59and people designing their lives, and then people building these products and these experiences that people all over the world kind of consider to be the best that there are. So it's a social place.

26:21L -M -N -T. Element electrolytes. Have you ever felt dehydrated after an intense workout or a long day in the sun? Do you want to maximize your endurance and feel your best? Add element electrolytes. to your daily routine. Perform better and sleep deeper. Improve your cognitive function. Experience an increase in steady energy with fewer headaches and fewer muscle cramps. Element electrolytes. Drink it in the sauna. Refreshing flavors include grapefruit, citrus, watermelon, and for a limited time chocolate medley, which you can enjoy hot. Formulated with the perfect balance of sodium, potassium and magnesium to keep you hydrated and energized throughout the day.

27:15These minerals help conduct the electricity that powers your nervous system so you can perform it your very best. Element electrolytes are sugar -free, keto -friendly, and great tasting. Minerals are the stuff of life. So visit drinklmnt .com slash tetra and stay salty with Element Electric Life. LMNT.

27:49It felt like in the early days of the valley and when I saved the early days, I mean seven years ago. So the idea of move fast, break things, disruptor awards were really good things. And Apple had to think different campaign and something seems to have changed. What do you think has changed? I don't know. I don't understand it because it seemed like the whole tech revolution was about you could finally be yourself. You could learn what you want on the internet. You didn't have to get what was fed to you. I remember at one point in time, Twitter was called the free speech wing of the free speech party.

28:31That's right. So what changed? What changed? So let me start with, I actually think a lot of that is still there. And it's not there as much in the big companies, but there is still an anarchic spirit in the startups. And so we meet with people every day who basically are just like, yeah, screw it. I'm just going to like, Brian, break a lot of glass. I'm going to just do something brand new. Isn't that necessary to make new things? I believe so, yes. I think so. Seems like. I believe it is. Let me say I'm very much in favor of it, and we can talk about why that is. But it's obvious. That's how you do new things.

29:01Like you have to. Like look, the past will crush you, right? If you let it. But most cultures throughout time and most cultures in the world today, they're just, their thoughts are dominated by the past, right? And there's a good to that, right? Which is they have culture, they have continuity, right? They venerate their ancestors, they, you know, I mean, literally, like the natural in the social form of human society is like literally ancestor worship, right? And like you just like, look, there's a lot to that because like your ancestors like learned a lot and they try to pass it on to you through these kind of cultural transmission things, started with like poetry and then made its way through to kind of all the books and stories that we have today about people in the past.

29:34But like, you know, most cultures are just like completely dominated by the past and they just don't do new things and anybody who tries to do new things, there's this thing called tall poppy syndrome. You know, the tall poppy gets chopped off, right? And there's tall poppy syndrome and the west, there's tall, tall poppy syndrome in the east. It's like this very natural kind of tribal thing. And so anybody who's really gonna do something, it was gonna upset the Apple cart and it's gonna make people upset. And then I would say technologies maybe even the most advanced version of that or the most dramatic version, right?

30:01Which is like a transgressive piece of content, like a transgressive song or movie, can really make people mad by suggesting that there's a different way to live. And that obviously can have a big effect. But tech does something, I think even more than that, which is when something disruptive in tech changes, is it doesn't just change tech, it also changes the sort of order of status and hierarchy of people. This is sort of the thing. So why does the media hate tech so much? And there's a lot of potential explanations for that. And one of them is just like we cut the legs off from under the media business.

30:30Like we have obliterated, like for newspapers, we have obliterated basically most of their advertising revenue because most of that, people used to advertise the newspapers now they just advertise online. And so there's like a big reordering. Well, actually in music this happened, Napster file sharing you know, cut the lights off from under the economic structure of the music industry, that caused a massive turnover in who was running the music companies. And now there's a new generation of people who have figured out streaming and digital distribution. And so like, there's been a complete re -ordering of status that has come from that.

30:58So I think that's a lot of that. Yeah, so look, I think you have to do that. I think there's a couple things, a couple other just broader things that have happened. You know, this one I'll kind of pin on us, which is just like, because the dog has caught the bus. Like, a lot of us who've been in tech for a long time, it was always like, Wow, what we're doing is actually really validly important. People don't understand it. We have to tell them how valuable an important it is. And now it's like, oh, they actually get it. And now they're mad. Right? Like, we want. Right? The dog has got the bus.

31:24Like everybody gets it now. Everybody gets the tech as powerful and important. And now they're really upset about it. Right? And so some of that is. Now they want to control it. Now they want to control it. Exactly. And so the dog has got the bus. You know, it turns out the bus has its own opinions on where things should go. You know, the dog is going to get dragged behind the bus. And so there is some of that. Look, the other thing that's happened that you know, well, but the other thing that's happened is just, you know, look, the world, I think, really changed in 2016 and like, why and how and, you know, how much of it was one guy and how much of it's a global phenomenon.

31:53You know, we sit here, you know, sitting here, sitting here today, a very Trump -like figure just wanting the Netherlands, you know, and this kind of huge surprise. You know, a Trump -like figure just wanted Argentina, which is this kind of huge surprise. But, you know, the world, like, the world as I experienced it in the United States and this is looking at value, like, it really changed in 2016 and a lot of people really got psychologically, you know, altered, shattered, broken, reformed kind of through that. And I bring it up just because like tech got pulled into that shift just like everything else has gotten pulled into that shift.

32:21And that's why I tell our founders a lot of the time who try to grapple with this stuff is just like look like what we're doing is very important, but we are not we're not the bus, we are the dog like society writ large has energy and momentum of its own and we are wrapped up in it just like everybody else. So you know we are changing it in some ways, but also we are part of it. We're getting kind of dried along with whatever the prevailing trends are. And so, tech has become politicized. And I would say socially energized in a way that it never had been. I don't know, maybe this is a big difference between this and L .A., which is maybe in music and movies.

32:54You could say that music and movies have always been intertwined with politics, probably going back forever. But for sure, going back to the 60s, I kind of very deeply intertwined. And music may have actually caused sort of political and social change in the 60s and 70s. in the 80s, but it was also reacting to it. It was like a feedback loop. And I think in tech for a long time, we didn't have that. We were building like tools. You know, we were building like fun tools that people could play with and use. But when they would get engaged in politics, whatever, they would put down our stuff and then they would go out in the street or go on TV or do whatever they would do.

33:25And now tech is integral to how society works. But even beyond tech, there was a time when big companies were focused on making the best product they could and having the best bottom line they could. And that was all they were focused on. They were focused on the business that they were running. And it seems like somewhere along the way the idea changed where now corporations whose obligation to their shareholders is to do what I just said now feel like they have some moral imperative. How did that happen? Yeah, so you know, You know, look, it should start by saying, for people who haven't really thought about this hard, there is a, you didn't do this, but there's a misnomer that people apply to companies, which is they'll say, this is usually criticism, they'll say companies are only focused on their shareholders and they're only focused on making money, and they're legal and they'll say, sometimes they're legally required to optimize profits no matter what, right?

34:19And so, and this is why they say, companies are sort of gonna be sort of intrinsically, morally, neutral or even evil, because like if it makes sense to pollute, or if it makes sense to have horrible policies or whatever to build products that deliberately break, so that people have to buy the new thing next year, like they're going to do all these evil things because they're optimized for profits. It actually turns out there's actually no legal requirement for companies to purely optimize for profit. In fact, quite the opposite, you're supposed to optimize for kind of a long -term value. And management has a broad latitude to be able to kind of decide whether to engage in social issues, whether to have different kinds of policies inside the company.

34:51You're perfectly protected legally as an executive of a company if you trade offshore term profits for some longer term. Just a brand value. We're just not even though doing this thing would call, we were thinking of starting a new wailing division. We were going to kill a bunch of whales and sell their meat. We could make money doing that. We're going to decide not to do that because it would impair our brand. Like as a corporate executive, you're completely, legally covered. Nobody will ever question that. And so corporate executives have always had a fair amount of latitude in terms of how they want to steer their companies.

35:19Look, for a long time, one is just like, what business people were just focused on business. But the other thing that happened was the received wisdom that you got when you were trained up in corporate America, I would say between the 70s and around 2016, and the way I was trained up was, you actually don't wanna get involved in political and social issues because the backlash, like you're gonna make people mad, right? You're gonna summon the demon of, you become involved in politics or social change, so in politics and social change, you're becoming involved in you. And so you're sort of inviting a level of backlash that very well might destroy you.

35:51That was something Michael Jordan famously said early on and when they asked them about politics, he said, I want everyone to buy my sneakers. That was his answer about what his political affiliations were. And is that a capital to statement in part? But is that also a social and political and moral statement, which is, look, I don't want to be a divisive figure. I don't want to have some kids love me and other kids hate me, because I'm on the wrong side of some divide. And so to your point, like I think that was very common. Look, I think this vortex opened up And a very large number of executives felt very guilty about some set of social and political things.

36:29There was tremendous peer pressure that developed to be able to stand up on things. Employee bases changed the millennials in particular showed up in the workforce with a lot more sort of demands on their employers. I mean, even my like socially activated older friends who in business, like even they are shocked that the young employees show up kind of so fired up and so demanding that companies take all these different positions. And so that happened. And then, you know, look, I think some people think they can differentiate. I think some companies have deliberately decided that they'd rather have to love them and have them market hate them than have nobody care.

37:04And so I think maybe that's, you know, just you might say hypothetically, and Nike, for example, actually might be quite happy if liberals buy twice as many shoes and conservatives don't buy these shoes. And, you know, maybe that works. And then, quite honestly, I think people are drunk. I think they're drunk on sanctumony and they're drunk on politics and they're drunk on feeling powerful and they're drunk on, you know, grandstanding. And I think it's like a chemical, you know, it's like a chemical thing. It's a drug and it's very easy to get hooked on that drug. It feels great, right? And then, you know, people on your side are like praising you and talking about how wonderful you are.

37:36And you even say it also probably to some extent it feels good to be hated, right? Because it's like, wow, I'm really important. All these people hate me. They're all mad at me. Like, I'm in the mix. Like, right, I'm in the fight. Like, I'm really making a difference. Like, people are, you know, And so I think a lot of it is just like, I think a lot of it is just this emotional thing. It's like a heroin kind of thing. And I think there's a hangover from it. I think actually a lot of you, a bunch of companies have that hangover right now. A bunch of companies have been damaged very badly by this.

38:01And they have severe internal problems. I mean, look, there are big companies where the CEO no longer runs the company. Like the companies run by the mob. Right? I mean, just the employee base can just freak out and like threaten to protest or threaten to quit. And like the CEO just rolls over every single time. And so at some point, what is that? Is that a company? Is that a social movement? We're wearing the skin suit of a company. So I've said optimistically, you might say this was a moment of time phenomenon, and that mean reversion will kick in. And then you might also say, no, look, we just live in a different world now.

38:34We live in a different kind of culture. We have a different kind of media. This is never going back. And I think it's very much related to the question of like, do politicians quote unquote become normal again? or do they actually become kind of stranger and stranger, more and more unusual, I think, probably things just get weirder. I think so too. I think it'd be hard to go back because if you see now when you look at the politics of the old days, it was very lawyerly, it felt less connected to the people. It felt like a different elite class. Yeah. And now it feels more like whether it's AOC or MTG, it feels like the people.

39:13Yeah. Well, and by the way, just the fact you can name them, right? Like the fact you can name them, the fact that they, those two, like I already have like iconic, you know, three letter, you know, names, right? It's like they're right up there now with MLK. Like how did that happen? Right? The fact that they, like if you pull, if you did like, unated surveys, like they would pop off, you know, those are probably, you know, those, those two people might be, you know, other than Trump, like, and Biden, those might be the top respective, you know, Democrat Republicans today for unaided awareness.

39:36Or, you know, that if you talk to institutionalists in DC, they're very frustrated by this, because they view it as like, okay, the people who are purely focused on media presence are getting all the, you know, they're basically getting lined up to be future presidential nominees whereas the people focused on substance are. Nobody knows who they are. And it's kind of like, okay, maybe that's just the world, right? And maybe if you're going to be a politician, that's the world you have to live in and you have to be willing to, you have to have a strategy on that. And maybe if you're in business, this is just the world you live in and you have to, you know, just, I don't know, pick up people.

40:02But, you know, like Disney went through a version of this with their, you know, with Bob Chapichoe, who I know I'm like, you know, where he tried very hard to keep them out of politics. The first thing he did when he came in was I said, we're staying out of politics, we're not everybody by the sneakers. And the company just was not having it. And Bob Eiger has come back and he very much doesn't have that view. He's keeping the mental, these things. And who had the right strategy? I don't know. And it's going to tilt one way or the other. And so I think there's a lot to that. I also think, look, I think that I'm not a big fan.

40:31I think there's a lot of simplistic kind of accusations that social media is ruining everything or the internet is ruining everything. and I don't like, I think those are generally kind of simplistic and sort of incorrect things. But, you know, like there is no question that media forms society. Society forms media, media forms society. You know, Macluan wrote extensively and worked that holds up incredibly well about, you know, the role the media takes in shaping culture. And look, we live in a different media landscape now. And again, we're the dog that got the bus. Like we made it. We invented all this stuff.

40:59But like, you know, you might say, like a critique I might apply to my own, my own thinking on this is like, what did you think would happen when you connected everybody together into like a single global media sphere, right? McClue and use the term the global village. And when people forget about the term the global village, as he wasn't saying that was good, when he was saying as basically the entire globe was going to, globe is going to revert to the behaviors of a village. And the behaviors of a village are very different than the behaviors of a city. The behaviors of a village, you know, the city are like cosmopolitan and open -minded and like embracing of new ideas and so forth.

41:27The values of a village are like very narrow -minded and moralistic and sharp and they, you know, that village is like ostracized people, for slight differences in belief and everybody's watching everybody all the time and everybody's super critical. And so we actually invented the Global Village. Everybody's acting like a villager now. McLuhan would say, yeah, no kidding. Great job guys. The great benefit of the villages in the past was that you'd go to another village and they'd have their whole own way of doing it. That's right. And if it's a Global Village, your biodiversity of villages is gone.

41:58Yeah, it just gets all turns into the same. The same thing. And so, look, I think we've been, And I think this is a big problem. There's been my interpretation of the drunk on politics thing, which is just like we became drunk on being sort of domineering members of a global village. They thought they became both company CEOs, but also a lot of activists became sort of convinced that they were in a position to kind of skear the totality of human civilization and society, by using these technologies. I think there's some truth to that. So we've all gotten basically wrapped up in a single collective psychodrama.

42:29But look, I also don't think like like 8 billion people want to be wrapped up in a single collective, like psychodrama. And I mean, if anything, there's just like at some point a adrenal fatigue, like it's just at some point, it's just tiring to be like onto the next like howling outrage. Right. Like how many of these, how many of these can you actually go through? But it's actually even funny sitting here today as like there's still people trying to push like the same buttons they were pushing in 2020. They'll attack companies or whatever. And in 2020, you know, some accusation will be leveled on something racer, whatever related.

42:57And you know, in 2020 would have led to a huge crisis. and today, people are pushing the button and just nothing's happening. And it's just because like, it's just like the 80th time you're going through a race panic. The boy who cried wolf. The boy who cried wolf. And it's becoming so clear that so much of that is like astroturfed and is cynical manipulation boys and like using emotional guilt to try to manipulate people to do what you want. And people trying to like make money by selling their newsletter or whatever it is, where they're talking about how evil everybody is. At some point, it's just like, all right, I got the message, okay, I guess we're all evil.

43:27I don't know, I'm tired of feeling bad all the time. And then by the way at some point you get the actual full on backlash, right? Because the other thing the internet does is it's the opposite of a global village. It also makes outlying ideas much more accessible. And so if you've had it with a prevailing view on something in the internet, it's the best medium you've ever had to be able to go basically find the other side. And so you find like, I think there is, it's like there's a power lock curve, like this thing where there's like a small number of like global village, like dominant ideologies that basically subpeople in and sort of mass numbers.

43:56But then there's also this long tail thing where there's also like a thousand, you know, there's a thousand like entrepreneurial efforts to have new kinds of cultures, new kinds of art forms, new kinds of creativity. And those have already existed, but they've always been fragmented and scattered and it's always been hard to like find your little micro, you know, kind of tribe of the things that are interested in what you're interested in, especially if you don't live in the same, you know, in the same place. And the internet now makes it possible to find, you know, the alternate view of basically anything and it makes it easy to find the other people who like that alternate view.

44:22You see, you have like the same with products like Etsy, you know, it's fun to go to Etsy instead of Amazon So it's a different experience right right and by the way you can pick and choose right as a consumer That's a great example you can pick and choose where you can you know You can buy 80 % of your stuff on Amazon because you say you know it's a garbage bin I just don't care that much versus like your water bottle. You like you want something handcrafted right and so the The choice side of it is there and it's really being blown out and I just Yeah, I I continue to have a lot of faith in people, and I think people don't want to, I don't think people want to live in a vortex of just uniformity and sort of mutual shaming forever.

44:56I think they at some point people want to do different and do new things, and I think that, you know, that also makes that possible.

45:05Welcome to the house of macadamias. Macadamias are a delicious superfood. Sustainably sourced directly from farmers. Maccadamias, a rare source of Omega -7 linked to collagen regeneration, enhanced weight management, and better fat metabolism. Maccadamias, art healthy and brain boosting fats. Maccadamias, paleo -friendly, cheeto and plant -based. Macadamia's. No wheat, no dairy, no gluten, no ghee M .O.'s. No preservatives, no palm oil, no added sugar. House of Macadamia's. I roasted with Namibian sea salt, cracked black pepper, and chocolate dips. Snack bars come in chocolate. Go to nut white chocolate and blueberry white chocolate.

46:12Visit houseofmacademias .com slash tetra.

46:24Do you think the big tech companies are too big to fail? No. Could something come and replace Google for search? Well, so let me start with, the goal of every company is to become too big to fail. The goal of every company is to become an athlete. The goal of every company at some point is to get the government to give them an athlete status and to protect them. So there's this term, regulatory capture. So big companies hire all these lobbyists. And at some point, what they're trying to do basically is they're trying to write laws and then get their own people into key positions at regulatory agencies so that the government basically becomes their overseer in their project.

46:59to prevent people from growing to compete. Yes, to prevent people like us from backing founders, like we back who are going to develop the disrupted thing that's going to ultimately howl of them out. And so by the way, this is a cycle. All of our companies grow up and they want to do that too. So this is a cycle. So and this one, we're both a dog on the bus. Venture Capital funded Google. Google is now trying to do what I just described. They have this active effort underway to try to get government regulatory protection. And then we are back in the next generation of companies that are trying to basically screw that up.

47:27I mean, look, Google, this is very public right now. Google is going through exactly the process you just described right now with search, which does AI make search irrelevant, right? Like does it make any sense to go search and click on links if you just have an ad, and you know, a shared GBT that can just give you the answers? And that's a very big product question for Google. And then there's also a very big revenue question in there, which is 99 % of Google's revenue comes from ads. The ads are these keyword ads on search results. You know, does that happen? And you know, if the user's not clicking on links or they click on ads.

47:57And so, will there be as an AI, or even if there are as an AI, will those ads work the same way? And what's your prediction? How do you see that playing out? Yeah, look, and I would say Google, if it was in this interesting position, where number one, they invented a lot of these AI technologies. And so they keep breakthrough for ChatGPT, actually happened at Google in 2017, in classic the company for them, they sat on that. They didn't use it. If you talk to people in Google, they will tell you that Google could have had ChatGPT4 the way you have it today. They could have had it in 2019. have they been focused on it, but they decided not to do it.

48:28So this is like a classic big company thing. Now they've broken up. Now they feel threatened. They understand everything I just said. And so they've got this internal effort called Gemini and they're going to try to lead Prague to LGBT. And they've certainly burned one of the founders that's come back to work on that. They've done how will it work for the ad model? That is an open question. That is a very open question. I would give two assumptions. My assumption is number one, yes, there will be a transition from search in template links with ads. There will be a transition from that to just ask a question, get an answer.

48:55And I think that transition is underway right now. I think Google's gonna, they're already starting to launch their own version of that. So when you do Google searches now, and a lot of those searches, at the beginning, they'll actually just try to give you the answer. They're trying to lean into that, so it's that. Look, on the ad side, the argument for ads and AI is actually quite similar for the argument for ads and search, which is basically like, if you're searching, like I don't know, if you're searching, we're shaking on vacation, right? And it's like, well, how about this beach in Thailand?

49:18It's like, okay, well, the very natural next thing to do would be, I'm gonna click and buy a ticket, and I'm going to click and buy a hotel reservation, right? And so presumably people are going to be talking to the AI about things that they ultimately want to buy, and then there will be an opportunity in there to actually be helpful to actually have an app that actually makes it possible to do the thing that you're talking about. It seems like if you're going to something for information, if they have an interest in selling you a particular thing that undermines their ability to give you the best information.

49:50So the history on here is that the search ads was actually not invented by Google. It was actually invented by a different company in the 90s. The sky bill gross actually down here in LA. And I call, I think it was called go to dot com at the time. And they were the first company that rolled out a search engine where there were ads and there's sort of ads in between the links. And it was actually viewed out of the gate. It's like unethical for that reason. And it was like, oh, this is bad because it biases the results. And by the way, there's, you know, look, there's some truth to what you're saying.

50:16Like there's now there's a commercial incentive. I mean, I can't imagine a version where it's not 100 % true. Well, but now here's why it's not 100 % true because there's utility value to it. Right? But you don't know. Because of bias, we don't know that. Yeah, but at some point you want to buy something. But how do you know what to buy? If the thing that you're trusting to give you the information is trying to sell you one of many, that's a problem. So here's the other thing that happens. And you might view this as making the situation better or worse. They don't sell these ads on fixed price basis, they're options.

50:50Ads are price on options. So the ad that you actually see is the guy who's willing to pay the most to present the ad. So it's not based on the best service. It's not based on the best service, but it's also not just the lowest common denominator thing either. It's the person who can justify paying the most for the ad. And it turns out there are a lot of product categories where the guy who can pay the most for the ad has the best product. That's a real stretch. That feels like a real stretch. Well, Mercedes spends a lot more money on the form of advertising and marketing that does to someone who says that's the best product.

51:16I've been told you what it does, because generally speaking, in the world, the argument goes, the better the product, the more expensive the product, the more the company's gonna spend on advertising it. Do you believe that? I think there's a point to it. Generally speaking, I'm not thinking, can you argue it? Do you believe that? Well, look, a lot of times, when I'm looking, okay, I'll give you something. A lot of times when I'm looking for something, and I just wanna buy something, and I don't wanna spend the next week trying to research it and navigate it, trying to correct all the biases, and I just need the thing.

51:42It's like a pretty good proxy. It's like, okay, if they're willing to spend $40 dollars to get this thing in front of me, click. Right? One -air bot diet. Because the key is self -interest. If they weren't making money with the $40, then they wouldn't be spending the $40. They spent the $40 even before they knew whether I was going to buy it. And so it's costing the advertiser real money to run the ad. And so there's some proxy there for they have a healthy... There's some mean thing maybe the best funded. Yes, so there are moments. There are times. There are times when the categories get overfunded.

52:10But most businesses over... This is where self interest kicks in and helps you, which is most businesses need to make money over time. To do that, they need to calibrate how much money they spend on marketing versus everything else. And if they can, over time, if they can afford to spend money on marketing, they probably have a pretty good product because a lot of people are probably buying it, right, and if they don't, they don't. And so, right, so the look on your face is the look that everybody had in the 90s when this idea first rolled out. And it was exactly this argument. It turned out it worked.

52:34Well, it worked in that people have gone along with it. Yes, they have. People have given up the search for the good thing in exchange for the convenience of Somebody wants to sell me this and they're telling me what to buy well Yeah, you can see here'd be a question are people more or less likely to buy the better version of the product out of all the choices today than they were before the internet Depends on the person yeah, okay, but like a lot of people I was a part of this There's also just a fading memory thing that I think that happens which is free the internet It was actually really hard to get any information on products.

53:09That's true. And so God help you, like try to find, like, well, this is a big issue in the car industry for a long time. Some cars were just like much more fundamentally pro -underbrucked down that of the cars, but it was actually really hard to figure that out. And it was hard to aggregate. It was hard to either have an authoritative voice that would tell you, or it was hard to actually have an environment where people could tell each other. So a lot of consumer reports. It's hard for you to do it. They built a big business. They built a big business because of that. And there was a certain, as you can sure you remember this, there was a certain kind of consumer who was glued to consumer reports and used it as the Bible for everything they bought, but there were a lot of people who didn't, because there's a lot of people who just have other things going on in their lives.

53:42And so, I think on balance, the internet has made it so the better products have done better relative to worse products. I mean, there's just there's basic examples, which is remember, when you used to mail order, you used to mail order things. Yeah. Wait, four to six weeks for delivery. Yeah, and no idea what you were going to get. Right, yeah, exactly. And there was some drawing and some catalog somewhere, and whether it was even the products. Right. And so, you know, today, like everything's overnight, like there's like reviews and ratings. And like, it again, is it perfect? Like, are there like fake reviews?

54:08It's definitely better. And you can order several different options. And you can see when that works for you. And everyone understands because it's mail order, things get sent back. I definitely prefer it now. That's a different question, though, than the corruption built into the system. Yeah. There's some disconnect. Oh, look, having said that, Google, I'll defend Google Siner, which is one more step, which is, Google doesn't care. Google's running an auction. They're optimizing on price of what the person's woman to pay to put the end of front of you. So they're not, to my knowledge, they're not putting their thumb on the scale, like they're not basically saying, we're gonna rig this.

54:43Like Amazon is. Well, so this, this Amazon, so Amazon's in an interesting spot. Are Amazon ads surfacing the high quality products? No, not even the ads. Amazon, if you're looking for whatever, However, often the first recommended choice from Amazon is the one that Amazon makes. That's also true. Yes, exactly. The FD there's a government investigation hour. The FDC is trying to force them to stop doing that. Is that true? Yeah, for that reason. I don't know whether or what will happen with that case, but they're trying to. So yeah, it looked like, but take all of your points. I think, well, well, made.

55:15This is all going to get rethought in the AI world, right? So look, should the AI have a view? Should the AI have opinions of product quality, the way it has opinions on everything else? Well, so Elon, this is also happening in social media, right? So Elon has this thing now, community notes, right? For X, where it has this thing. And it's this. What is it? And how does it work? Okay, so it's a very clever thing. So all these social media companies, these social media companies all used to be free speech, you know, free speech, when you're in the free speech party. And then it just turned out, like a lot of people said a lot of things that made people mad.

55:41And so these companies all created what they call trust and safety groups. And of course that is like a super or well -earned terms. It means you can't trust them at all and they're totally insane. Right? Right. And so what you ended up with in the trust and safety groups was just basically, you had some well -many people and then you had a lot of people who were in there to basically put their thumb in the scale. And this really kicked in around politics where they would have just obviously different standards for different political parties that everybody could kind of see in plain sight. And so that model never worked very well.

56:09And so the people at Twitter, and I think this actually started before Elon got there, but he embraced it when he got there. They came up with this new idea called Community Notes. And the idea of community notes is that there are people who have the button to be able to submit a community note, which then there's a way that you kind of apply and you become one of those people. It's like being a Wikipedia editor or something like that, right? And so there are certain people who can propose community notes, but a community note does not get approved and used unless people who have a history of disagreeing agree on it.

56:38So for example, we see a statement, somebody says something and it's just like, you know, it's wrong. You and I are two community note editors. We have a history of political disagreements. We have a history of conflict where we really don't see Ida I on things. But then this time we see Ida I. I see. And then that's the one that gets posted. And so I would say it's been working shockingly well. Like it's not perfect because it still goes wrong and people try to trick the system. But it's by far the best system that I've seen that actually works. So he's done this very clever thing, interesting thing where you can now do community notes on ads.

57:11Right? On ads. That's interesting. So if somebody runs an ad that makes false claims, or exaggerate something, it can get community noted. And so the platform itself is like, oh, actually the claim in that ad is actually not correct. Here is the actual truth and here are the links for the truth. And you can imagine the reaction of the advertisers. I mean, they're just like absolutely furious every time this happens. But he's running this experiment, basically, saying is the faith and trust of the user base being improved by the fact that the community knows can actually be legitimately fact checked.

57:40or the ASCAPFACSCHECK, is the value of that greater than the loss value of the advertisers who get alien. And by the way, they advertise their who get alienated by that are probably the worst advertisers. Anyway, there's some issue in service. If you're making a product and if people have a problem with it, if you're the maker of it, you want to know that. If you're good. Right. If you're honest. Absolutely. Absolutely. If you want to improve your product to always be the best thing, of course you would want to know that. And if you don't, maybe you should not be advertising. Exactly. Maybe you should not be allowed to advertise.

58:11And so I think this is a very clever experiment. And I think there's a lot more of this kind of thing. There's a lot more experiments like this people can run. I think in the next 20 or 30 years, we'll figure out actually what. This is all around trying to make the global village work better. Like this is how to get basically 8 billion people to share a mind space and not want to kill each other. And this is the kind of technique that I think you can start to figure out.

58:35So much of today's life happens on the web. Squarespace is your home base for building your dream presence in an online world. Designing a website is easy, using one of Squarespace's best in -class templates. With a built -in style kit, you can change fonts, imagery, margins, and menus. So your design will be perfectly tailored to your needs. Discover unbreakable creativity with fluid engine, a highly intuitive drag -and -drop editor. No coding or technical experience is required. Understand your site's performance with in -depth website analytics tools. Two -space has everything you need to succeed online.

59:24Three to blog. monetize a newsletter make a marketing portfolio launch an online store the square space app helps you run your business from anywhere track inventory and connect with customers while you're on the go whether you're just starting out or already managing a successful brand square space makes it easy to create and customize a beautiful website. Visit squarespace .com slash Tetra and get started today.

1:00:04Can a company grow too fast? Oh yeah, for sure. Tell me about it. Yeah, very common. Very common. Yeah, well, so especially when either the revenue just starts to pour in or when they raise too much money. And the basic mechanic is very straightforward, which is just, you know, you have a problem. If the answer is you can always hire somebody to fix the problem, then that's what you'll always do, because it's the easiest thing to do, right? And you just go hire another engineer and they'll just go do that. The problem is it's just like any human organization, like organizations behave very differently when they're large versus when they're small.

1:00:34And there are, you know, big issues involved in running large organizations that are, you know, very complicated and it's very easy to run them badly. And it's very easy that I use the metaphor of obviously leaving the building. It's very easy for you to almost leave the building, company grows too fast, the good people quit, the bad people stay, who we have another thing we call the law of crappy people, which is in any organization. So organizations have levels, so there's like levels of promotion levels, level one through eight or whatever. And there's always an internal process to try to figure out how when people should get promoted.

1:01:03And so the law of crappy people says that the quality of any level in the company will degrade to the worst person at that level. And it's a very natural human thing, which is just like, like, well, I want to get promoted at level three. And I know I'm better than that other guy who just got promoted. And so whoever is the worst person at that level now sets the bar. And so basically, as companies grow, basically, performance tends to collapse, kind of level by level. Meetings, communication overhead, and overwhelms everything. People are just sitting in meetings all day. Or another thing that happens is you hire too many people from another culture.

1:01:36So you think you have your unique culture, and then you hire 300 people from Google, and you discover all of a sudden that your Google all over again. but without their business, just with all their problems. Yeah, and so that's, I would say that's, this is part of the, more companies die of indigestion and starvation. Right? Because starvation is no fun, but starvation is highly motivating, right? If you don't have any money, and you only have a few people, you have to be smart. If you are flush with cash, or you can try to spend your way out of all your problems, and you could say there's a metaphor here for individual lives, right?

1:02:09Which is like a lot of people getting a lot of trouble. I think that's why I was asking about if people mainly came to you for money because it seems like money is not at the highest level of what it takes to make great things. It's a piece of the puzzle, but it's not the main piece of the puzzle. I don't think. It hasn't been in my experience. It has been, that's right. And I think that we find ourselves, and people don't believe this, we find ourselves in practice often trying to get our companies to not raise as much money as they can, kind of down the road. Yeah, look, we just tell them like, look, it's just bloat, like just simply bloat.

1:02:41Like, you know, look, you know how to run this company with 50 people, everybody feels great. You just need to close your eyes and imagine that, like, that's not how your day is going. Your day is going, organizing meetings on calendars and cross -crow. And if you can grow the company with the 50, why would you hire more? Yeah, yeah. Now, usually you do. So there are some companies, there are small number companies that become big and important with very small numbers of people. And there are these kind of very magical successes. You mean examples? So the great one is Minecraft. So Minecraft at the time that it sold was I think three people total.

1:03:08Wow. And what did it sell for? Remarkable. I don't know, $1 .5 billion or something. But it was this giant global phenomenon. And now it's, you know, it's on my Microsoft and now there's a lot more than three people working on it. But it's an amazing thing. I mean, look, Bitcoin was like one guy. You know, it's worth like $800 billion to do or something. Like, so there was that. You know, WhatsApp, when they sold it to Facebook, it was 50 people. Instagram when they told it to Facebook was I think 11 people. So you have these kind of very kind of special things. Majority is one guy, plus basically a bunch of contractors.

1:03:44It's this huge kind of force, not one art, AI. So you have these kind of very special things. It's just, they do at some point, this is a pressure on the other side. At some point, there are things that are just kind of obvious that these things should do that they just don't do. like Minecraft was like a great, you know, kind of viral phenomenon, but it really should be what Microsoft's turning into today. It should be like an entire world and there should be, it should be like you should be able to use it for like education and it should run on all these different platforms and it should have all these different, you know, kinds of all this new content and, you know, it should be this thing that people can be in for 20 years.

1:04:16And so, you know, it's much like deeper and richer and more built out than it was. You know, same thing, what's that? What is what's happening at Meta right now? What's that? What's that is, you know, very widely used. It's one of the main ways that small businesses This is communicate with their customers. But there's no capability, there's been historically no capability inside WhatsApp for a small business to be able to maintain a customer database or whatever inside WhatsApp, which is like a very obvious thing, like for a restaurant or something. And so there are things like that that make sense to have that are actually valuable to the users that just require more people.

1:04:43So yeah, there's pressure on the other side to grow. It's just there's some theoretically optimal rate of growth of heads, and it's very easy to, it's one of those things where nobody ever quite gets it right. Well, and then look, the other thing that happens is the economic cycle, like these companies all tend to over hire during an economic boom. And then at some point there's a crash and there's a rationalization. And you kind of wish, you always wish that you didn't have to go through that, because it's bad to do layouts, but it's very hard to just keep that. It's very hard to over time have a sustainable model where you don't make mistakes.

1:05:12How did you meet your partner Ben? Yeah, so I met Ben. We were actually trying to staff up at our first company in Netscape in 1984. Yeah, we were desperately, we had the Tiger about the tail. was the internet was going to take off and we knew it. So we had that kind of get in position. So we were trying to hire good people. And he was one of the first people who came in from an existing successful software company to join us. And he was a young guy. Game of the day was called a product manager. So somebody who kind of orchestrates things. And then very quickly, he just became clear that he was one of the sort of sharpest young people we had.

1:05:42And so we promoted him very rapidly over the course of the next few years. And I ended up working very closely with him. And I always thought, had we sold our company in 1998, can 88 four years in, but just an eternity that was, it turned out, was only four years. If we had stayed in the pen and I think he would have ended up being the CEO. He was on track to do that. So I spent a lot of time with him and we got to know each other really well. And then it sort of became obvious. After we sold our company, I wanted to start a company and so it became obvious you would be the right person to do it with.

1:06:11Is it as much a friendship as it is a partnership? There's no thing around founding teams and just generally, which is what it's a, a friendship Business relationship works better than a business relationship for our friendship. Like when two people who are personal friends going to business together often destroys the friendship. Whereas people who have a business relationship first and really learn to trust each other in business can become very good friends. So for us it's been that latter. It was business first and then it's become a very deep friendship. We talk all the time and we get along great and we're always teasing each other.

1:06:42It's all good. Having said that is now almost 30 years. And so there also is an old married couple kind of aspect to it, right? And some of that is we can finish each other's senses. And some of that is we get on each other's nerves. So we still argue all the time. It's interesting. We'll be something you would argue about. Just constant arguments. I mean, basically everything. Well, it's obviously a big benefit of the partnership is you have somebody you can actually talk to, who you trust, to can actually tell you things that when you're getting things wrong. So we argue about, you know, for the biggest argument it's always around people, you know, is this person good or not?

1:07:17You know, should we, you know, should we promote this person? Should we fire this person? You know, some of that's around our firm. A lot of that's around the companies we work with. You know, he works with a lot of our CEOs. He's kind of our, he's our, among other things, he's like our management guru. So he's, he's the guy who basically works with all the sort of most, you know, high potential CEOs to help them kind of develop. And so he develops kind of, he wrote the book on it. He wrote the book on it. He's actually a hard thing about hard things. It's consequence, like he's legitimately very opinionated about people coming out of that.

1:07:44I would say he's more focused on culture. He's extremely focused on the culture of the company that a CEO is developing. He applies that in our firm, but he also applies that to the CEO's that he works with. He really wants to understand, when he's evaluating a founder, he really wants to understand what culture that person's building inside their company. I'm a little bit more open to variations. I'm a little bit like I'm more open to like what he might call recklessness or chaos or you know, original thinking that maybe he doesn't work and might blow up in your face. And so I'm always a little bit more on the side of you know maybe we should like you know.

1:08:20Would you say you're more of a risk taker than he is? I would say it's different kinds of risks. I think we're pretty similar in aggregates, different kinds of risks. Here's I'd be curious whether I actually he would agree with this is I think he thinks And I think he's right, but I think he thinks they're kind of timeless ways to build great cultures. Like, the leadership is not a new idea, right? It's something that has existed for a long time, cultures. The second book is actually on culture, right? And he goes through, he talks about like Genghis Khan of the Mongols, and he talks about like, you know, the code of the Bushito and the samurai, and, you know, kind of these cultures.

1:08:51And he's entering to that kind of thing. And so he draws a lot of historical examples for cultures, and he's like, look, these are these continuous threads. of like what does it mean to have people trust you, to be able to bond together into a team, write all these things. The other side of it is, you know, look people who come up with new ideas, often have new ideas on everything. And so a lot of the people who have new ideas on like technology also have new ideas on like management and culture and how to be a CEO and how to start your company. And do we really need all these meetings and can we just run everything on video conference and do we even need to have those or can we just do everything on Slack, right?

1:09:21And so a lot of these founders will have these very creative, original ideas on how to actually run their companies. And I would say most of the time, because speech to them would be stop, focus your creativity in the area that you actually understand, which is building products, your designing things. And then just get good at running the company the way the companies have been running for hundreds of years. Don't try to innovate on everything. And I would say he's usually right on that, usually those experiments and a disaster. You know, that said, every once in a while they don't. And every once in a while there's a totally new approach.

1:09:48And I just say, case study of this right now is Elon. Elon, we work with Elon. I've known Elon for a long time, but we were Tesla and SpaceX predated us as a venture firm, so but we're involved in his Twitter acquisition. And Elon has a completely different way of running companies. And you can read about this, the press, the press has covered it as a length, right? Like it is a much more blunt, chaotic hair trigger, direct engagement way. He is very quick to fire people. He's extremely aggressive. He gets micro -manages to the Instagram, he's involved in everything. And it's a playbook that for most CEOs would lead a disaster.

1:10:26Wasn't that the case with Steve Jobs as well? Yes, Steve Jobs is a good point. So Steve was a genius, Steve was a genius with it on the dark side. And the genius was real and the dark side was real. Everybody's seen the results and the results, obviously make the whole thing worth it. But Steve could be really rough on people. And there are many true stories about Steve being really rough on people. And this would be a great example. was Ben's V would be, yeah, you just don't do that. You figure out how to be great without basically, being rough on people like that. And I would say, in general, that's the best advice.

1:10:55In general, if you're coaching a young kid on how to run a company that you want them to try to calibrate that. Because generally, if you're really rough on it, you don't really get anything for that behavior. Well, Steve did. Well, we don't know that. We don't know that because he didn't have Ben coaching it. OK, so this would be the argument. OK, so this would be maybe the argument that Ben and I would have about it, which is Ben would say, Ben would say what you just said. Ben would say no. That was purely destructive. If Steve had not done that, he would have been even better. Well, we don't know.

1:11:21It is no way to know. But I do think that's a real question. It's a real question. I can make the other side of the argument. Tell me. The other side of the argument is, A -players want to be on a company of all A -players. The people want to be around other great people. Yes. Most companies are way too tolerant and indulgent of mediocrity. Most companies are good at getting rid of just like what we call shith birds. They're good at getting rid of people who just are clearly no good, but they have trouble with his mediocrity. Good but not great. And Steve just like could not tolerate mediocrity.

1:11:48And so when he identified that you were meeting, he actually, there was a term called flipping the bosom bit. And when he flipped the bosom bit, he just he fired you that day because what he what he had learned over time and his view was that you never unflipped the bosom bit, right? You just need to call it right in front. And Ben would disagree with that idea? Yeah, yeah, he would say, look, it's just too hair trigger. Like you're in a meeting with somebody and they say one thing that's wrong and you fire them on the spot. Like how do you know? Like you're not telepathic, like you're just happened to be in that meeting and look maybe his wife yelled it in this morning or maybe his kid just died.

1:12:17I mean you don't know what. And also maybe he's actually right because it does happen sometime. Exactly. Maybe he's right. Yeah. Exactly. I'll tell you the conversation we had on this that really made a big impact on both of us. It was with Andy Grove, who you know, who was best away. But when he was running Intel, he was considered kind of the best CEO and kind of the history of the tech industry. And when we were young, we used to meet with them and he would help us on things. And so one of the questions we had frame was like, wow, it always feels like we're firing people too late, right? It's like every time we fire somebody, it's like, wow, we wish we had done it like nine months sooner, right?

1:12:49Because they had done damage in that nine months or because you missed an opportunity. Yeah. It's something better. Both. Well, it's this flipping the buzzer thing. It's like, okay, we identified that they had an issue, a problem, you know, every once in a while people can turn it around. Most of the time in a workplace environment, if you've got a competent management and they identify as somebody has an issue, like generally people don't turn it around. They just statistically, they just generally don't. And so you have this thing where, like if you're running the standard management playbook, you put people on what's called performance plans, and you try to more aggressively kind of coach them for improvement, and then they normally don't make it through.

1:13:22And then at some point, three months or six, or nine months later, you kind of make that call. And actually, the reason you put people on performance plans is because most companies are too indulgent of mediocrity. And so most companies don't aggressively perform as managed along the way. They don't document things along the way. And you need to put people on a performance plan have a legal justification for firing them because you need that paperwork that you actually did evaluate them. But you're just like, shit, I'm wasting time as a person is doing a mediocre job. They're affecting the organization and mediocrity.

1:13:48I wish I can move faster. So we asked Andy. We asked Andy. Basically, is it normal to always feel that you're firing people three months or whatever and I must relate? And he's like, yeah, he's like, that's what it always feels like. And he said the reason is because if you did it more aggressively than that, the organization would do, he was a sociopath, right? And it's like, well, Steve didn't care about being viewed as like, you know, he was like, yes, we're just not, I'm not optimizing for what people think of me. I'm not optimizing whether people think that I'm nice. I'm optimizing that the people around me are going to be at the top of their game.

1:14:20And if they're not, they're going to go, right? And he wants the exact same way, right? And then the results, and again, this is the argument, right? The minute I have it, it's like, okay, Apple, it's a Tesla space, you know, rockets It's a land of their butts, right? Like, you know, like, oh my God, right? Is the Stevie Lahn Playbook, like the more aggressive Playbook, actually the one everybody should be having. And Ben's comment to that is, no, most people will destroy their organizations but they try to do that. Like it doesn't actually work. I go through all of that just to say, yeah, that's the kind of debate that we'll have.

1:14:47And look, it's not resolved, right? Like, and quite honestly. Well, you get it's unknowable. It's unknowable. I don't think it ever will be resolved. Tell me why a company would choose to stay private and why it would choose to go public. So most of the great business institutions that are around for many decades, most of them are public companies. And the reason is because they end up with a lot of employees, they end up with a lot of customers, they end up with a lot of constituents, they end up with like governments, you know, getting very trusted. Any that are private? Yeah, there are big ones.

1:15:14I mean, there's a bunch. There's one prominent one to be Coke industry, Charles Coke. It's a giant industrial company and they do like, they're huge like, they're giant, they're big. I mean, I don't know the numbers. They're just, the numbers are just tatanically large. They'd be a Fortune 50 company tomorrow if they went public, but he's kept it private the whole time. Bloomberg is another, Bloomberg, the company Bloomberg is another example of a big private company. Because private company has always been private, Mike Bloomberg just owns it outright. So yeah, they exist, but most of them end up public.

1:15:41And I think the reason they end up public is just because if you're an institution at some point, you kind of have to act like an institution. You have to be trusted institution. So we have to be transparent. People have to know what you're doing. public companies carry with them. These responsibilities to report to the public. You have to explain yourself in a way that's very open and you're under these very stringent legal requirements to do so accurately. So people tend to trust the things that public companies say more than private companies. How accurate are those results? So in the United States, I would say quite accurate.

1:16:08Either they are wholly accurate or if they're not, they're off a little bit. Outside the US, it is still the Wild West. And so we went through a series of scandals in the US business world. we went through a series of scandals in the 2000s around Iran and companies in the world, the companies in that era. And the stringency of the accuracy of the reporting and the level that you get reviewed by the government and the penalties to lying are quite high now. And look, most people, most responsible people want to run something that is like legitimate and genuine. There's an old thing, there's been old thing, there's been old thing I used to hear when I was a kid, which is like, remember there was this gangster named Myrlansky that famously ran the big part of the Mafia in the US and people used to say, well, you know, Myrlansky He was so successful as a gangster.

1:16:49If you know, just imagine how successful he would have been if he was like, you're running General Motors, right? Like, and it's actually, I think that's actually untrue, right? I actually think like, you know, if the Myrland SQ is actually very ill suited for running a big public company because you get caught lying, you know, and you break somebody's legs and like, it's a big problem, right? Like that's not what a trusted institution does, right? And so the best and the brightest, actually I think want to be legit. They want to do something that's like actually genuine. And that's true for companies as well.

1:17:16Well, look, having said that outside the US, like, boy, like the scandals, I mean, even Europe, like the scandals in Europe are like mind bending, and then once you get outside the developed world, like things get really hairy. So most of the world is not well developed on this stuff yet. But anyway, so there's the transparency, kind of truth telling component to it. There's also the, at some point you want to have a lot of shareholders, like at some point you want kind of the world at large to be able to invest in your company. You know, you want ordinary people to be able to have a stake in your success.

1:17:45You want everybody to have your stock in the retirement plan because then it sort of gives everybody a reason to kind of root for the company You know at some point you want to be able to you've got all these employees that you're paying in stock at some point You want them to be able to sell their stock and be able to buy houses instead of just to college You also get what's called a currency so your stock becomes a currency and so you can use your stock to like buy other companies Right and you know it's easier to raise debt when you're when you're when you're public and so they're And by the way, a lot of employees just want to work for, they want to work for an institution.

1:18:13They don't want to work for some fly -by -night startup. They want to work for something that's trusted where they can, you know, one of the things I do with candidates a lot of the time is especially immigrants, kids, you know, some kid who's, you know, sort of a first generation immigrant. You know, I'll literally get on the phone with their parents, you know, and this has happened repeatedly, where I sort of explain to the parents like, no, actually, it's okay, if your kid doesn't go to work for IBM, right? Or Microsoft or Google, it's okay if they go to a startup because actually in the US, that's not actually, because the parents are worried about career death if it doesn't work out.

1:18:42There are a lot of people who just want to work for a stable company. So those are all the reasons to go public. The reason is not you, it's just like, look, you're supposed to all the scrutiny. And so you... Is it only scrutiny or anything beyond scrutiny? Oh, well, so it's the consequences of scrutiny. So you have a stock price. The stock price trades every day. But do you start making decisions based on the stock price? Yeah, big time. But that's not your core business. Correct. Changes your whole business model. Exactly. 100%. That seems bad. Yes, well, it depends on... Yes, so it destroys a lot of this.

1:19:13This ruins a lot of companies. So the easiest failure case is that. You've been running your company, just running your business the way you run a business, and then all of a sudden you've got this daily scorecard, and you're optimizing the daily scorecard. Or even if you can get through the day, you're optimizing quarterly results. You're reporting every 90 days, and you're optimizing for that. And so, yeah, so when this goes poorly, basically what happens is time horizon contracts. And so instead of planning things one, three, five, 10 years out, you're planning 90 days out, and nobody can do anything great.

1:19:4090 days, get the scale of these companies. And so you just basically, this is when Elvis leaves the building, right? Like so. And by the way, often it's coincides with, this is when the founder stepped down and then they hire professional CEO, professional CEOs, and sort of optimizing for their own compensation, they're optimizing on these short time frames. So that's a big downside. There are ways to deal with that, but that is a big downside for sure. In your position when you're investing in companies for 10 or 20 year trajectories, if they go public, how does that change your position? Because it's now they're playing a different game.

1:20:13Yeah, that's right. They're no longer playing the 20 year game. Well, they might be. There are public companies that do, right? So Amazon played the long game the whole time. Still does. Apple played the long game. The whole time Steve Jobs did the turnaround of Apple. They were public company. Look, Tesla's been public for, I don't know, a decade. SpaceX is private, but Tesla's public. And Elon runs Tesla. I love the exact same way here on Space X. Netflix invests for the long term. There's a lot of these companies that have, I think, done a, I can tell you Mark Zuckerberg, you know, it met a, this hasn't changed.

1:20:40I'll give you these things. You know, look, it can be done. It's a higher competency, you know, it's a little bit more of a high wire act. Like you're getting graded by the world on what you're doing and like, you know, oh, this is a speech we give to CEOs because they're like, oh, well, I'll just do a speech out of state, I'll just do whatever I want. It doesn't matter, I'll be like, yeah, but you just need to imagine what happens in your stock drops like 97%. You're on the front page of every business newspaper in the website in the world, talking about what a turkey you are, right? Have you experienced that?

1:21:05Oh, of course, yeah, multiple times. You're not picking me, sniffing you as damn people, what's it like? Oh, yeah, it's just so horrible. It's awful. Yeah, no, it's happened multiple times. Most of these companies, there's this great form of a chart, financial chart called the drawdown chart. And the drawdown chart is basically, it's baseline is zero. And then the chart is the percentage drop that the stock is experienced in different points of time. And so it's zero to like a negative 100%, right? So it's like this and then it's like, and it looks like it's a little bit like somebody having a heart attack.

1:21:31It's a cardiac arrest. And so I just give the draw down charge for Amazon is really interesting, because like for example, it's Amazon's this just giant success. But like there have been like, I think five different times in the last 20 years where the stock has dropped like 97 % or something like that. I mean, just like these massive crises of confidence for basically everybody's just like, yeah, this is - Well, it had no profits for how many years? For a very long time. Very, very long time. For a very long time, exactly. And so you were running, yeah, right. You were running into it. What Jeff had was, he talked about it publicly.

1:21:55He's like, look, we're investing for the long run. We're reinvesting every penny of internal profit back into the business, we're building what they call intrinsic value in the business, and we're just not going to hand out dividends to shareholders. And the investors who went along for that ride did great, but it's easy to say that. It's harder to do it when your stock graph is 97%. And the headlines, there was a famous cover of Barron's Magazine in 2005, and Amazon at that point was already 10 years old, or coming up on it, and it was literally, the headline was literally .bomb. Like Amazon is going out of business.

1:22:25They're going to be worth this is zero. I mean, I went to investor conferences in the early 2000s where people were just openly laughing at Jeff. Just like laughing in the meeting. Like you're just like completely full of it. And look, everybody, part of being a CEO as people are doubting you, whatever, when the entire world is doubting you. Like, it's just, because what you know what happens is like, anybody has a bad career, they've had one of those moments, it has that which is like, all of a sudden you're talking to your friends, your friends are like, are you okay? Right, like how are things going?

1:22:52And then you go home, your family is like, are you all right? And you're like, yes, I'm fine. I'm the same person I was. I'm not. Did you know Jeff at that time? Yeah, yeah. Was he fine? Yeah, he was fine. No, no impact. No, let me let me say this like this is the 4 a .m. thing I don't know since I was not sleeping with Today we have a much closer relationship back then back then I was not sleeping with And so what is he experiencing at 4 in the morning? That I don't know right? And you know, there's only really you know two people who run a position and know that right him and his wife at the time So that I don't know, but to the outside world and to all of his friends, he was finding the entire time And he was finding the entire time because he's just like, we have a plan We're executing the plan.

1:23:32We're not going to get shaken off of this No, look you could also say right a fair response to what I just said a survivorship bias Right like here I sit talking about the ones that work like what about all the ones that didn't work Right because like a lot of times in the stock market, you know drives a stock to zero It's because the company sucks and like it's gonna fail right and so like that's the other side of it Right, there's no substitute for the thing I think ever times when then the market loses faith in a company and it goes to zero, but the company still has value and then it comes out back out of the ashes and reinvents itself.

1:24:05Yeah, so our company, so our company, our company, which, sorry, loud cloud, sorry, sorry, bad company, Ben, I started in 1999. We went public in 2001 and then by 2003, by 2003, our market capitalization of our stock was half the amount of cash that we had in the bank. Right. And so, had we just simply liquidated the company and given the cash back, we would have made twice your money on the stock. And so, what the market basically said was, yes, these guys have, what that meant, what the market was telling us, the message implicit in the price was, these guys suck so bad, that even though they have this cash in the bank, they're just going to burn the cash, and there's not going to be anything left to show for it.

1:24:42And then, we actually, Ben gets most of the credit he was running the company. He turned it around, and then we ended up selling, I think the stock went up 40X off the bottom. We sold the company for 40X that amount. And yeah, that was sort of a quote unquote turnaround. Look, Steve Jobs, Apple, so when Steve took Apple over in 97 when he came back, Apple had less than 90 days of cash in the bank, like they were about to go bankrupt. Like that's how bad it was. In 2009, I had a chart I was carrying on my pocket all through 2009, 2010, when we were starting to firm. And I think Apple was trading it up.

1:25:15I think they've bottomed out in a price earnings ratio of like six. And what that basically means, a price earnings ratio of six basically, it's like what a steel mill trades out if it's about to go out of business. It's like training for the liquidation value of like the plant equipment. Like it's a super low, it's basically the market hates you and things you're an idiot kind of thing. And this was like, this was like right when the iPhone was taking off. Right. And so it was this like, and there's a there's a loose relationship between PE ratio and growth rate. We're roughly speaking PE ratio and growth rate should be about the same.

1:25:42And so if a company's growing 10 % if this growing earnings 10 % a year, or should have a PE of about 10, sort of a loose, a loose relationship. an apple had a phase there where they were, that P was six and the growth rate was like 40 % and then in some periods through their size 80%. And so it was like undervalued by like a factor of 10, just on like basic math. And it was obvious to see that. Well, I thought, I mean, I wasn't running public money, I didn't put my money on my mouth was, I gave a lot of interviews at the time where I pulled out this chart. And because the point I was not trying to make a stock call, the point I was making is everybody hates tech irrationally.

1:26:13Like, if a financial crisis, like everybody got negative about it. This was after the dot com. Well, it happened twice. So it happened after 2000. So what happened was the 2000 crash was like a real tech crash, and tech really fell apart. And there was actually a lot of carnage in a lot of companies went under. And then what everybody thought would happen was the global financial crisis 2008 would cause that to happen again, but it didn't. But they thought it was going to. So they acted as if it would. And they traded the stocks as if that was what was about to happen. And so basically 2008 to 2011, 2012 was just this extreme level of irrational hate and fear.

1:26:47And again, it's not like a super genius thing to be able to say, looking because you're looking at it, you're like, well, I don't know, this iPhone seems like they're going to sell a lot of these things. And it's the same thing. Google was growing super fast, Facebook was growing super fast, but the world at large had just gone negative. Well, there's this famous thing, or the metaphor for the stock market is this famous thing that they say is, you can think about the stock market as a person in Mr. Market. And he's like full on, you know, clinically manic depressive. Right? And there's just like certain times that Mr.

1:27:13Market is just euphoric about everything. and there are certain times when he is just terminally depressed about everything. And then we all collectively are Mr. Market, right? So it's a group psychology thing. And so it's very hard to be a participant in the world and not get pulled into the group psychology. But as a consequence, the market goes through these wild swings and it does regularly go through periods where people are just rationally negative. And then of course, you know, and then it's like, okay, read the investor textbooks and it's like, well, that's when you buy the stocks. It's like, well, yes, but that's when everybody's in a horrible mood.

1:27:41And anybody who buys these things looks like a complete idiot because everything, everybody knows that they're all going to go out of business. So you talked about the consequences of being public. This is one of the consequences of being public is your companies get caught up in this. And you feel it on a daily basis in a way that you might not if you're... You could be in a tech company that's not public and you're just looking at your bottom line and everything's fine and you know your business is doing well. Whereas another the same company public, all of a sudden gets caught up in this wave of things crashing and you crash with them.

1:28:10But nothing is strange in your company. That's right. Well, this goes to this relationship between the metrics and management. And so there's this whole thing in management, which is you manage what you can measure. And so if you have a number that you can optimize on, you tend to optimize on that, you tend to run your company around that. It's the same thing politicians do with polls. I've got a poll number, and I'm just going to try to optimize around that. Now, is that the optimal way that people are actually going to vote? Like who knows? But like you've probably seen this in political speeches.

1:28:37If you've seen this in political speeches on TV, or political speeches, or debates, and they'll have this thing where they'll have a focus group and they'll have a dial that they can go to like 100 % negative, 100 % positive and then there'll be these red and blue lines. And it shows word by word. It shows the mood of this focus group watching the thing. And so it's like a stock price, right, for every word coming out of a politician's mouth. And so if you're a politician, do you use that as a tool to try to like optimize every single word coming out of your mouth and like basically become the master of the craft of political speech giving?

1:29:04Or do you say, well, that's crazy. Like, if I get wrapped up in that psychology, I'm going to drive myself nuts and I'm going to end up being incredibly, you know, unauthentic and I'm just going to be like a pure opportunist. And it's the same thing with the stock prices. You read a lot of history. Is this just a passion or do you see some other use in understanding the past? It's a desperate attempt to predict the future. So look, for 30 years, like I've been, you know, I've been doing this now for 30 years, starting companies or finding them, you know, statistically with what I do, it's like a 50 % success rate, 50 % failure rate, basically.

1:29:40It's with it with it with it. Which pretty good actually. Which is pretty good. It's pretty remarkably good. Well, it's pretty good. It's pretty good. I mean, it feels terrible. It feels awful. It's like your business, I mean, it's like your business too. It's like sometimes you all. 50 % is remarkable. Yeah, for baseball, it's great. There are other fields for test taking. It's terrible. Driving on PCH, you got to score 100%. So yeah, so it depends. Look, and here's another thing. like statistically. But you're betting on things that are one in a million things. Well, yeah. Yeah, let's say one in a thousand or something.

1:30:14Okay, one in a thousand things, 50 % really good. It's a good, look. Yes. And just like in your business, the upside of the winners is bigger than the downside of the losers. Right. And so if you have asymmetric upside on the winners and contain downside on the losers, that's 50 % does well over time. But the failures are just always horrible. Right. That doesn't get you, statistically you can know that. And actually you can know that emotionally, every failure hurts tremendously. And Anne has wrapped up with people. So these are people that you care about. And one of our company's fails, it's not gonna take our firm down because of the 50 -50 thing.

1:30:48And our investors understand that. But it's a founder who has poured five, I just read that company, just the guy's poured five years into this. And it's been a big part of his life. And some of these people bounce back and they go on to other things. Other people just, at some point, they're just like, I can't take it anymore. and then they, you know, it put not everything works. Not everything works exactly. That's real. Yes, very much so. Right. And then there's another thing in tech, an adventure, which is called the Babe Riss Effect, which is the home run hitter strike out more often. And so, right, the people who were really trying to do something new and radical, actually fell at a higher rate, which would be interesting to make sense.

1:31:24That's right. Fordicting the future, these things is absolutely impossible. Like that said, like, boy, I sure wish that I could. And so how do you ever possibly predict the future? And I do think there's some wisdom that comes from understanding, in particular, the human dynamics. I think people do change, and people have changed. I don't actually believe that we're the same people we were 100 or 1000 years ago. I think actually the people themselves might be changing, but in a lot of ways. But looking people, there are constants to human psychology, sociology, behavior of human beings and crowds.

1:31:54There are cycles in history of different kinds. And so at least in the past, you can kind of go back and, you know, the risk of reading history is always, you know, the outcome. And so the outcomes look inevitable after the fact. But if you can kind of get yourself away from that, and if you could, especially the history works that I really like are either contemporary accounts of what it was like in that moment, to actually experience that, or really the best historians are very good at recreating what it actually felt like to be there when it was all very uncertain. And then, then, look, there's also just a lot of tools.

1:32:24You can just learn, you know, there've been a lot of great people who have navigated through very difficult situations, like how did they do it? Like what's the toolkit? So yeah, so it's a desperate region of the past to try to learn whatever lessons they have to give me. Can you think of an example where something you learned from reading history impacted a real world decision that you made in the present? Well, look, I was just like rallying people after disaster, right? Which is like, okay, there's been a catastrophe at a company. Like, okay, now you've got to like, recon here the team. Like, how do you do that?

1:32:51I didn't mean to know how do you do that. Well, you've got to get up and talk to them. Okay, well, how do you do that? Well, how did Churchill do that? Like, you know? Yeah, that kind of thing works really well. And look, these are things that most people have not done. They're in a lot of cases. And so being able to learn from, this is somebody that says, you know, the best of history is this incredible intellectual conversation across brilliant people over time, who have kind of learned from each other, basically by reading. Are the top VCs more of a group of colleagues and friends or rivals in enemies?

1:33:19Both. Co -opitation. And we like to say, so we probably work together more than we compete in the reason is because most successful companies raise from multiple VCs over time across multiple rounds. So we end up on boards together and working with companies, but we do compete and had on for deals. And those competitions can get, can get, we can punch each other and then it's pretty hard. We're in one of those right now and we're going to try and punch another firm of the nose as hard as we can. And so, you know, that happens. And then I think it's like, it's like any business. It's like, I should have, you know, the movie business down here is famous for it or music probably is also, which is, you know, you do end up with grudges, you know, you do end up with like two prominent figures who really hate each other.

1:33:55And it's like, you know, well, 20 years ago, one of them said something in a meeting. And I've got my list. Ben doesn't hold grudges. Ben's great at holding grudges. I hold grudges. And Ben's wife holds grudges. And so when Felicia and I get together, we like, we're like, we're like, the Anja Stark character in Game of Thrones. We're every night. We recite the list of all the people that we're going to amazing at some point. Yeah. And I always, Ben's kind of always got it on me on this is like, you know, maybe you should let some of these things go. And I'm like, well, no, actually they're quite motivating.

1:34:29Like, I'll tell you one thing that gets me out of bed in the morning is, you know, the opportunity to really stick it to somebody who I feel like did something wrong. You know, 20 years ago. That's unbelievable. So I kind of like my grudges. Okay. They're very close. They're very, very important to me. Do all the VCs do the same thing or does each house have a particular style or strength? I would say the commonalities are there's a few universals which is based in a sort of this triangle, it's basically team product and market, is when you keep coming back to, so are the people really good, are they building a product that people want, and then is there a market that they can sell it to, and it's sort of the most simple form of the whole thing, and those probably are, you know, those were the most important things 50 years ago, those are probably the most important things 500 years ago.

1:35:09By the way, there's a long history of DC that predates all of tech, which we could talk about if you want, but it'll be about it. So, you know, Christopher Columbus shows up in the, what is it, the court of Queen Isabella, and he's got this like crazy idea to discover, whatever it was, the new route to India, and he needs X, whatever Spanish, the time for whatever it was to be able to kid off the boats. He was making a venture pitch, containing downsides. What's the worst thing that can happen is he brins all the money and the chipsink and everybody dies. A constrained upside. What if he discovers the new world?

1:35:39And then of course, our membership bias, we remember that story because it worked. We don't remember the downsides that failed. So he was raising venture capital. There's a famous story. where JP Morgan was, JP Morgan was an investment banker who mostly dealt with debt for building out big things like railroads, but he sort of dabbled in venture on the side and he was a, this is like 120 years ago now. He originally was Thomas Edison's first investor for indoor lighting. And so he wrote Thomas Edison to check for the new lighting business and the first indoor lighting system, electric indoor lights were installed in Jacob Morgan's famous library in New York by Thomas Edison personally and then through weeks later they caught on fire in Bernice Library now.

1:36:19And then he paid Thomas Edison to do it again. He built the library and put in lighting and it worked. So he did it. The one that I find so fascinating is actually the wailing industry. So the structure of the modern venture capital industry is basically very similar to how wailing expeditions were funded in the 1600s, like off the coast of Maine. And it was a very similar kind of thing where you had basically these captains who were the entrepreneurs and they would put together a business plan for a boat and a crew. and they actually have an equity model for how the crew members get paid. They get paid as a, basically, on the portion of the whale.

1:36:53And then they would come and pitch basically the way the people who financed whaling journey, the VCs at the time. And the VCs were specialists in evaluating the captain and the boat and the plan. The captain was a specialist in figuring out questions like, well, do we go to the place where there have been lots of whales spotted? But those are the places other people are going to be at, or do we go to this other place that we think nobody's discovered yet? And then, you know, like a third of the boats never came back. And then there's this concept, the way that VCs get paid is this concept called carry.

1:37:22So the term is carried interest and then the sort of colloquialism is carry. And it's basically, and the idea is basically it's like 20 % of profits. For the ones that work, you kind of make 20 % of the profits or some number like that. And it's called carried interest. And the reason it's called carried interest is because that's how the captain's got paid on the successful whaling expeditions and it was literally the percentage of the whale meat and fat that the ship could carry. This was where the term carry comes from. So on a successful voyage, the captain would get 20 % of the whale. It's interesting, both examples he gave were about light because the reason the whaling was such an important part was that's what was the fuel for the light before the electricity.

1:38:01Yep. Very fundamental. Yep, exactly. That's amazing. Yeah, exactly. So look, it's always been basically what you find. You have this kind of entrepreneurial personality, and it might be the captain, or it might be the founder, or whatever, or by the way, it's a movie producer. You have an entrepreneur personality who has a vision. But they're not going to be able to realize it on their own. They're going to need to be able to gather resources to do it. And then they're going to need money and partners to be able to do that. And then there's going to be some evaluation process. There's going to be some professional class of people who are trying to evaluate that.

1:38:32They're going to be operating in this domain where they're wrong a lot of the time, but the success is makeup for it. And so it's kind of this universal pattern. And I think it's been running actually for quite a long time. But the best guess would be this will run forever. It's for the rest of the thousands of years. You know, the kinds of startups that you'll have in a thousand years will be totally different than what we have now, but they'll still have the model. They'll have the same property, the unknownness of it. And you'll look, they'll be reading histories of what we did in being like, wow, I hope that we can learn from all their failures, right?

1:39:00It'll be the same cycle. This one more thing is like people get mad. It's always you find interesting, but you know, because it's become very popular to kind of get mad at venture capitalists right now or kind of be mad about this whole process. You're mentioning like the move fast break things, get mad about disruption. It's like, well, it's like fundamental, do you want there to be new things in the world? Because if you want there to be new things in the world, they're not going to show up predictably from well -mannered people who are going to behave well in every aspect of their lives. And then the new thing is not going to disrupt or change anything.

1:39:30Like that's not what happens. change doesn't enter in this kind of peaceful cause. It's always a revolution. Exactly, right? And if it's not, it's not change, right? And so would you really prefer, to critics, I say, would you really prefer to live in a world of total stagnation? Or are there like nothing changes? Like, is that really what you want? And by the way, what do you think would happen in that world? Like, what would the politics be like? What would society be like? And so, I think, I think basically everybody would hate that. And so, but to live in the world in which the revolution's happened, you need to have a perspective.

1:39:58And it's easy to say this and it's hard to do, but you need to have a perspective that says, yeah, these revolutions like look, they're gonna be wild, right? They're gonna be wild. They're gonna upset a lot of things. They're gonna upset a lot of people. They're gonna upset power relationships in society. Hierarchies, gatekeepers are gonna be furious, right? Like old incumbents are gonna be furious governments. They're gonna get freaked. Like all that is gonna happen, but it is a direct consequence of the fact that it has actually changed. You know, dynamism happening. And I just like, yeah, nobody has ever figured out how to do this in a way that makes everybody happy.

1:40:25It's just a question of whether it happens at all. And by the way, there are many societies in which historically this didn't happen. And what we know of those societies is they basically just died. If you look at the Fortune 500 over history, how much change has it had over what period and what are the inflection points? It changes a fair amount, although a lot of the changes I have to do with mergers. When two big companies merge, has anything really changed? I'll just give you an example. Time Warner here in town, Time Warner Discovery have merged. It's a high drama in the media business right now.

1:40:57I was going to have one of that. But it's like, you know, Warner Brothers Studio has been a business for 100 years. If Warner Brothers Discovery works, it'll still be Warner's Brothers Discovery in 20 years. If it doesn't work, it'll get bought. You know, I don't know, maybe Apple buys it. And then Apple will run the Warner Brothers Studio for a while. And then at some point, they'll get tired of being in the media business and they'll sell it to, you know, I don't know, Disney or something. And so it's just like, you know, but it's still the Warner Brothers Studio, right? And so I think in big company land, there's a lot of what looks like drama in reality.

1:41:24It's just kind of assets, it's trading cards, being traded around. You know, are the movies any different than they were 10 or 20 years ago? Maybe a little bit but not really so yeah, I think a lot of the change is actually not real change Having said that looked the sectors change a lot Right, and so you know look when when there's like build out happening in a sector I mean when we're all roads were new they were most of the stock market right will pre -tech It feels like it didn't change some We'll put like cars with we forget what was new at the time right so like there was you know from the 20s through the 50s Cars were new right and so the car companies became you know the car companies were not big in the 20s and they become huge and dominant by the 50s.

1:41:58You know, GE is Edison, basically, like he had to like invent all that stuff. And before it existed, GE wasn't a big company than it was. So, and actually this is one of the things in history. One of the things is useful in history, which is most new industries look like tech in the beginning. And so, like I just example the car industry. Actually, it's actually funny. The car industry actually didn't grow up originally in Detroit, it grew up in Cleveland. And the stories of the first 20 years of the car industry, basically are these hobbyists and tinkers and entrepreneurs in Garajas in Cleveland.

1:42:23in like 18, 1990, 19, 19, 10, trying desperately to figure out how to get these car things to work. And by the way, the car was greeted with an enormous amount of fear and trepidation. Like people were not happy about the car. And a lot of states had these only different stories, but people reacting to attack us. A lot of states actually in the US did not want cars to be on the roads because what was on the roads was horses and people. And cars were dangerous and scary and loud and they freaked out the horses. And so a bunch of states actually have a they call red flag in that time period where you could have a car, but you needed to have the car, the car would break down all the time.

1:42:55So you had the car, you had your mechanic that would go in the car right with you. The car would only go like 20 miles an hour, and that was sort of, which was like super fast, because it was faster than a horse. And then you had to hire a third, you had hired a third guide, which would be 200 yards in front of the car with a red flag, waving ahead of time. So that the horses would know, the riders on horses would know to pull to one side. And then the horse lobby got really mad about this, and so they passed along Pennsylvania, where they said if a car and a horse encounter each other on the road, the owner of the car has to stop the car, disassemble the car into the pieces, transport the pieces and hide them behind the nearest hay bale.

1:43:29So it's not to freak out the horse. And so my point being like, what was that? That was the tech industry. Like that was the disruptive tech industry of that time. The personalities of the people, all you even example, the car industry, GM is like this, you know, giant company, 100 years old. There was this guy, Alfred Sloan, who was like the famous CEO of the whole thing, who built it, but actually there was a guy named Billy Durant, who was actually the founder early on before that. He was like the Elon character. And he basically, and his name is Lost in History for whatever reason, but he basically created the modern car industry.

1:43:57And he read the stories of him, and it's just like reading a must -eat jobs, or Ted Turner, or their page or Sam Elman or any of these people. And so I actually think these patterns are, I think these patterns are actually older patterns. Who do you view as your biggest competitor? I mean, the answer to this always sounds incredibly cheesy, but it actually gives true, which is, it's me, like by far. This is the speech I give inside our firm, which I very much believe, and I give the speech to myself all the time, which is like, look, if we screw this up, it's our fault. It's suicide not homicide.

1:44:29And it's basically, it's because we were not as good as we needed to be. We screwed it up. Like, there was a way to do it, and we blew it. And maybe we blew it out of ignorance, but probably we blew it out of arrogance. And there you go. And you know, hubris. And then, you know, I think for what I think for what we do, it is fundamentally a people business more than a money business. And it's interesting. Being in a people business like every conversation matters. Right. And you're dealing with people's lives. Right. So this is the speech I give internally. It's you're dealing with people's lives.

1:44:58And when you're dealing with people's lives, you have to talk. You have to be very serious about what you say. And you have to be very careful about the consequences of what you say. and really think hard about every conversation. And you know how it is, it's like, okay, it's your, you know, we're entering year 15, and I've had 400 ,000 conversations. The 400 ,000 at first conversation could go really wrong, right? As I'm not like on the ball. And I'm much more, I was saying, I'm much more worried about that than I am about somebody coming in and like stealing our lunch money. Other than financial returns, how do you know if you're good at your job?

1:45:31I, at that point, I think it's basically what are people saying about you? Yeah, look, the thing is though, like, look, it's easy, there's a paradox of the heart of that, which is like it's really easy to get people to like you just by always telling them that they're smart and they're good and that their ideas are good and like we don't do that. And so the trick is, do they still like you and trust you in respect to you after you have spent and the truth? Telling them the truth, exactly. And that's really your, it's your job. Yeah, yeah. It's like that. It's, when you tell somebody the truth and you're giving them good news, they can trust the good news because they know So when you had bad news, you gave them the bad news.

1:46:05If you just say everything is always rosy, it means nothing. That's right. It's the bad news that gives you credibility. That's right. But it's hard because people are under a lot of pressure and they get you probably, you know, probably how much do you live if they get upset in the moment. And then maybe a couple days later, they're like, actually, that was a pretty good point. You know, and look, if they know, if they trust you, right? I mean, the other thing that Ben and I talk about a lot, Ben says this a lot, as he says, trust and communication are opposites. Everybody thinks they're the same thing they're not.

1:46:33If I trust you, we don't need to communicate that much, because you know that I have your best interests in mind. Like Ben and I could go, we've been on a have this, Ben and I could go off and not talk for three months. And we would come back, we would have the exact relationship on the other side, because I have so much trust in him that I would know that whatever decisions he's making are in my best interest. Yeah. Right? Whereas if you don't trust somebody, like you really got to communicate, like you've got to like cross -check everything they say and like roll them and interrogate them. And so I think that, you know, the best relationships are, you know, this is what you try to develop, this like actually, you know, look, I'm gonna tell you some things you're really not gonna like.

1:47:04I'm doing that because I care. I'm doing that because I am doing that. And you're also doing that because you're being true to yourself in saying, this is who I am, this is how I see it, which is really valuable. Yeah, and look, I'm not gonna step in and run your company. Like I'm not gonna fire you. I'm not gonna like replace you. Like this is not the thing that's gonna make or break anything. I'm just gonna try to help you get to the truth. Right, I'm gonna have the trust relationship over time where you believe that that's what I'm trying to do. When you were in school in Illinois, there was a supercomputer.

1:47:35Yeah. How many supercomputers were there in the world at that point? A few dozen, maybe, total. Most of them would have been in government labs. Most of them, those are the kinds of supercomputers used for like nuclear weapons development, like so that... Why was it there? Cuptography. So the government would have had a bunch, but not really anywhere else. I, you know, I think, give... The government affirmed on a credit for that one. So remember Al Gore got in trouble years later for saying that he invested in the internet, the internet. And of course, that's not what he said. And what he said was he had played a role in the Senate in creating the internet.

1:48:08And of course, that was actually true. So that whole thing was actually a smear the whole time. That was actually true. And specifically what he did was he sponsored these bills in the early 80s, which did two things. Number one is they created what we're called the National Supercomputing Centers. And that was four universities that were given basically these grants to buy these very expensive and rare, kind of things at the time. And you give you a sense of how rare, especially these things were. Like in those days, like we had one of the computers at Illinois, they literally built a building for the computer.

1:48:37The computer was so big that they built the building, and they left the roof open, and they lowered after the building was built, they lowered the computer by a crane. They made the computer. Actually, it was a company, it was incredibly a company in Wisconsin. So it was a company called, there was a company called Cray. There was a guy named Seymour Cray, who did a lot of it, and then there was this company called Thinking Machines, and there's this guy Danny Hillis who you might have encountered at some point. So he was kind of very special entrepreneur who were good at this. They became sort of famous.

1:49:03You've seen them in movies. They were so expensive. This was like $25 to $50 million and this was 40 years ago. So this is like equivalent of like $10 million or something today per unit. But they actually, one of the things they really valued design and so they actually looked really cool. And they had like, they were water cooled. He has always a big problem with any kind of advanced computer. And so they did what's called water cooling. So they even have these very elaborate water cooling systems. There's a guy who actually bought one of these years later off of eBay and converted the water cooling system into a beer keg.

1:49:36That's the coldest, most expensive beer tap. So they were like works of art. And so you've seen them in different movies over the years. Yeah, so they were just like very, very rare exotic things. And so anyway, so the government funded these four centers at these four state universities, because these computers made do kinds of science possible. So these were used for like a different astronomy, astrophysics, decoding, these secrets of the universe, stuff, and then a lot of biomedical, protein folding, developing your drugs, scaring cancer, kinds of research. So it was sort of becoming key to a lot of areas of science.

1:50:10And then they had enough money to put for these centers in place, but they wanted to give scientists all over the country access to the computers and to do that. they needed a high -speed network to people can log in remotely. And so they funded what was called the NSF Net, National Science Foundation Network, which basically was sort of the internet, pretty the internet. And yeah, and so I sort of, yeah, my big stroke of luck was, it turned out Illinois where I went, was a top computer science school at the time, and still is, and they were one of these four centers. And so they just had this, did you go there knowing that was there?

1:50:40I did, yeah. I did, I didn't know that I would play any sort of inform, you know. I don't know how well it was. it would be to me, but it's seeked it out. Yeah, well I knew what was happening. So this was in the, you know, this was in the late 80s. And so yeah, it was big enough in the computer. And you know, the computer industry was being covered and like, you know, and you say, for some magazines that that, you know, this was, there were articles written about this, was how we experienced it at the time. But, you know, I knew it existed. Yeah, and I, and I have a home computer at that time.

1:51:06Yeah, although when I got to college in 89, you didn't, home computers in those days weren't actually useful in an academic setting. They weren't powerful enough. and you had experience on a computer before you went to college. Yeah, but on really simple computers. So I sort of went from working in computers to costs like $400 to work in computers to cost like $40 ,000 in like one step. And so at that time, it was a completely different kind of thing. So the computers that I worked on at college were just like, they said they were like $40, $40 ,000, $60 ,000 baseline cost just to have something on a desk and then these super computers were like I said $25 ,000 ,000.

1:51:40So these were not, one of the advantages of being at a UI you see at the time was that they had these resources, they had this equipment. But like all of my work was not done in my dorm room, it was all done in the computer lab, with like fluorescent lighting and like drop ceilings and all this stuff because all this hardware was like super exotic. You know, these days that doesn't exist as much your phone today is the equivalent power of that super computer that I were done. Your laptop is like more powerful than that. And so today that doesn't happen if you just have a modern laptop you have basically a full fledge computer You don't want anything on and then there's this cloud idea where you've got these grids of you know millions of computers up in the sky if you need more power So it so the sort of I don't know the romance or whatever of the this exotic thing in a building Being taken care of by people on white lab coasts, you know those days are kind of over in terms of what the Super computer was capable of how does that compare to like you your laptop at home now?

1:52:35Yeah, so you're laptop at home. Like, you know, my laptop right now is a MacBook, I think it's an M2 or M3 processor. And it's, yeah, I don't have an check, but it's probably somewhere between 10 to 100 times more powerful than that, so for computer at that time. Well, in fact, you could ask a cynical question on this, actually, or I could, which is, if those computers in those days were so rare and exotic, and they were able to be used for things like coding black, the secrets of black holes, and everybody has a laptop that does that today, like, where's all the creativity? Like, where's all the science?

1:53:02Where's all the creativity, which I think is a, actually a very excellent question. But yeah, look, in theory, everybody has on their desk today. And increasingly, just in their pocket, they have the ability to basically do what in those days we would have considered to be absolutely breakthrough scientific work, by the way, artistic work. By the way, another thing that happened at Illinois, that was a lot of lost history. Illinois was the, there were a set of universities in Illinois was one of them because of this that actually developed basically what we now think of as 3D computer graphics and ultimately developed what became CGI in the movie industry and the whole idea of computer graphic design.

1:53:32And so when you see a rendered tornado or whatever in a Hollywood blockbuster, a lot of that is actually techniques that were actually developed also at Illinois and a few other places like that at that time. And the supercomputers, originally it was so hard to do computer graphics. It was so process or intensive that it was only those supercomputers that could do that back then. So that was another thing that actually was invented at that time. What was Mosaic? So basically I ended up at Illinois. I ended up working at this supercomputing center after a few other things. And then they had a group in that super computing center that was building the software tools to make it possible for people to use these computers.

1:54:11And particularly, remember I said, the link between the big centralized computers and then the internet, basically the purpose of the internet as fed by the government was for scientists to be able to access these computers remotely. But then there needed to be a new kind of software tool that was built to actually make that possible. And so there was a guy's, what we call it, SDG, the software development group at Illinois that was in business to do that. I had government grants to build that software. And so Mosaic was a project that basically I, a group of us did at that group. And it's nominal purpose funded by the government.

1:54:41And it's nominal purpose was, and by the way, not funded for a lot, I was making $6 .25 an hour. So it was not a lot of tax money. But yeah, the purpose of it was basically remote scientific work. And so the original purpose of it, nominally was a scientist who wants to, like, basically publish information, have other scientists be able to read it online. I was the browser friend end of being able to do that. And then we also had a certain way with basically one of the first web servers that made it possible to store and host things. And so that was the nominal purpose. But then because it was government funded, we were able to give it, we actually had to give away, we're not allowed to make money on it.

1:55:15And so we really stood as open source. And then the nascent internet was starting to get big enough where there were people, people on it downloading it and using it. And then they had ideas for other things that they wanted to use it for other than scientific papers. and then that led to the felt belly of the creation. When did you understood it could be used from my then scientific page? Oh, right up front. I don't even think this was like pressions or anything. It was just sort of obvious. Like it was just like, oh, you could, the thing it just was immediately obvious was, oh, this could be used for newspapers, this could be used for magazines, this could be used for books, anything.

1:55:45Anything, right? It's a new communication tool. Yeah, and like literally I worked on a lot, I didn't build all this myself, but I worked on a lot of the early code for doing music online. I remember when we first figured out how to do music in the web browser. I remember how we first figured out how to do video in the web browser. And so I remember how, when the internet radio first started working, like there was this project that we were not working on, but I knew the people working on it called Embone at the time, which is the first music broadcasting. And so there were a set of us where it's just like always just sort of obvious that this is going to be used for everything.

1:56:16It's the McCluban thing. And the content of each medium is the previous medium. The content of a movie is the stage play. The content of the music video is the music track. And so it was the same thing here except this is the one where it's going to be all of them. And so I just thought that was obvious. A lot of us actually thought that was obvious. That said, there were a bunch of purists who disagreed. There was actually a big fight early on about whether there should be images in the web browser, images and web documents. Instead of just text. Because the argument was, but you can predict the argument, which was, if it's just text, it all has to be serious.

1:56:49where you introduce images and it gets frivolous. And then the frivolous will drown out the serious and then everything will go to shit. And I was like, that's what happened. Yeah, well, hey, that's what happened, and B, I'm glad that it did, right? Who wants to live in a world where you don't have images? And by the way, there's a logical flaw, right? Which is it turns out there's a lot of shit text too. So it's not the text actually gets you guarantees of quality either. It's true. And for better or for worse, I always bias on the side of openness and creativity. I want more experimentation in the world, not less.

1:57:17And so anytime anybody says no, we need to constrain this. I'm like, yeah, no, we're not gonna constrain this. We're gonna blow it out. You never know because you never know that's what can't predict. Yeah, but you get with the bad with the good, right? Oh, okay, so so guess what? So if we rolled out images and web pages guess I guess we're watching some of the first images the people put in web pages Well, you know, let's say dirty fixtures. I don't content And so my year old would say special parts And so that started. And by the way, it was a cliche that like the internet was used for porn first.

1:57:50That's not really the case. It was always kind of a more edge thing. But you know, people did start to post adult stuff. You know, this is a government funded program at the time. And so this is actually the first free speech issue. This is the first trust and safety issue, which is my boss at the time said, well, you have to like filter that stuff out. And I was like, filter what stuff out? And he's like, well, like nudity. And I'm like, like, how am I going to know which pictures have nudity in them? And I'm like, there's no way to do that. And he's like, well, you have to develop an algorithm that like detects nudity.

1:58:18And I'm like, what like through what like shapes? Like like like, booby detectors, like is that what you're asking me to make? And he's like, yeah, I can't you do that. And I was like, no, I can't. And furthermore, I won't. And I just like put my foot down. And I said, like, we're not going to build censorship into the web. And you know, that had, I would say, like potentially civilization consequences. You're at mosaic. You're building what you're building. you could see a lot of things in the future, but how did you imagine the world in the future than versus how the world is now? What did you see and what didn't you see?

1:58:55First just like look, the day job, there's this kind of presumption, Alpinheimer actually went into this a lot. There's this presumption that the people develop in the technology are somehow in a position to know the consequences of this use. And like I think that's actually untrue on several levels. And one of them is just a practical level, which is most of what I was doing was just trying to get code to work. So I had a day job which was all consuming. It was like 18 hours a day, it was just writing software and trying to fix bugs and software. And so most of what I was doing, it's the old thing in art, which is when artists get together, they don't talk about art they talk about, where to buy the cheapest paint.

1:59:25So it's like that. Most of what I was doing was mechanical trying to just get the stuff to work. That said, look, had you asked me then what I would have said was, look, I think this is going to be something that a lot of people are going to be able to use. And in those days, that was a very radical concept because people didn't have the computers to use it. that they didn't have the network connections. There was no broadband, there was no mobile, right? It was just, people didn't work on full computers in the same way. It wasn't clear that there would be any good, who would ever publish any content.

1:59:49It was like an open question at that point. And so it was radical enough, I would say, at that time to say this is something a lot of people are going to use. And a lot of people are both going to publish content on the internet, and a lot of people are going to consume it. And by the way, it's not just going to be fixed content, it's also going to be experiences and databases. And they're going to interact in different ways and chats and discuss things and so forth. We didn't have social networking, but we knew we had chat boards and forums and stuff. So we knew there would be a lot of communication.

2:00:14There would be a lot of groups forming. Could you spend time in a lot of groups? What was that like? Oh, it was great. So in my world, at that point, the dominant thing was what was called use net. The system was called use net and then the groups are called use groups. There was a version of Bolshein board system that ran across the internet. And there was a period of about 1985, the predated me, starting at 85 to about 1993. So I saw four years of it where it was like digital Nirvana. It was like the smartest million people in the world were like talking about everything under the sun and text only.

2:00:49You could like embed images, you could like attach images, but it was mostly mainly text. Would it be conversations or more like essays? Both, both. A lot of essays and a lot of conversations around essays. And then there was a folder, there was a whole hierarchy, so you'd have all these different domains, and so some of them were technical conversations, but there were like lots of political conversations or lots of arcs. And you could find the topic you were interested in? Yeah, yeah. And you would pick the news groups that had the topic you're interested in. Some of the news groups were unmotorated, so you could say anything.

2:01:16Some of them were moderated. They had a human who would keep them under control. Similar to social media, really. It was basically right. It was the ear form of social media. But what was fascinating about it in retrospect, it's a lost golden era that's been impossible to recapture sense, which is basically grew to be the millions smartest people in the world with basically no, for no idiots or assholes. And so it was totally like anybody in theory could be on it. Anybody could in theory say anything they wanted. It just so happened that the only people who had access to it were like the best and brightest.

2:01:47And so like there was no spam problem. There was no abuse problem. There were occasional flame wars, but like there was nothing. You know, there was no hate speech. You know, and then it's just the the the content quality was just incredibly high. And the communities of form were like incredibly high, and the trust level of form was like incredibly high. You know, people became very close, you know, across as they do across this with people they never actually physically met. And it was like this, it was like this Nirvana of like, you know, what if you could just have the million as far as people connected to nobody else?

2:02:14And then of course, then what happened was like, but everybody else showed up. And there's this term in the internet culture called eternal September. And so it's based in the fact that it was September, for 1993 is when AOL connected to USNet for the first time. And all the AOL, the 25 million AOL users or whatever was at the time were able to be on USNet and they just like bear it in shit. It's just like completely destroyed the quality, right? And just swapped it. And then USNet basically died in September of 1993. And never gets back. All that stuff's still online. Can you find it? A lot of it is.

2:02:46So it's been preserved. There's a thing called Google Groups. Google has a thing called Google Groups. And they have archives in Google Groups of a lot of these original things. And for what, it's kind of things you're interested in, it would be what we're called the alt groups. And so, like, alt .music. So if you go to Google Groups, you could read like, alt .music discussions from like 1990. And I bet you would find that it would be quite interesting. And so, eternal September is sort of this idea that basically now the internet basically consists of September 1993 in perpetuity, which is like no matter what good things there are, it's just going to get swapped with basically people with, you know, either dumb people or people with bad motivations.

2:03:21So anyway, so it is this, like, it is this, It was the Shangri -la of our experience. I think ways to create gay communities online. Yeah. Well, so there was a famous, there was another famous one of that era called The Well, which was, it was an internet system. It was a volition board system and it was Stuart Brandt ran it and I think it had a total of the peak of like 3 ,000 people. And there were two tricks to how he did The Well. One was, I think, if I recall correctly, I think he vetted all the new members. So it was like a club and then two. As he charged membership fee. And so you had to kind of pass both those hurdles to get in.

2:03:51And again, for many years, it was apparently really amazing, spectacular. By the way, this is an idea that nobody's really cracked the code on this, but this is arguably like an undiscovered idea. It's a known idea that nobody's figured out how to implement, which is like, how would you recreate that kind of thing today? It sounds really great. Yeah, exactly. It sounds like we spend a lot of time scrolling through things we might rather not if we had that more curated. But you have to be willing to violate the dominant conceit of our time, which is you have to be willing to say that not everybody's the same.

2:04:26And generally, I'm a fan of openness. I don't like the idea of like getting people by mic you or anything else or social acceptability or whatever, but within the universal, within the universal global village, I think it may just be that more people should start to carve out these more specialized areas. There are a few of these. There are mailing lists that are like this. There's often actually something that happens often a new social media product will first take off with being incredibly high quality to start. Well, Facebook was like this early on. Because Facebook started out just being Harvard kids and then when they expanded, they expanded the top 10 universities.

2:05:01And look, the kids at Harvard have lots of issues. But at least they're, at least in those thoughts, especially in those days, like they were 20 years ago, things have changed maybe even since then. But generally speaking, if you want your group people like, you know, 5 ,000 really bright young people, you know, the people going to the top universities are pretty good cross -section. And so, you know, and so there's this thing which is there's a there's a pattern that we've all noticed which is new social networks start. Well, a friend of mine puts it this way which is the quality of any group can only decline over time.

2:05:29But because basically you only want to join groups that on average are better than you, right? You never want to you never want to down select. You never want to deliberately join a group that is. But it depends. Well, it depends what the access with the access is. But generally speaking, you want, it's the groucher marks. I don't want to be a member of a club that will take me, right? Generally speaking, you want to go higher status, but joining the group, not lowering your status, but joining the group. And so this friend who argues that basically the thing with social networks is they're not technology platforms, they're groups, their communities.

2:06:01And the thing is, on day one, they're the best they're ever going to be. And then they will inevitably decline. And then there's a whole bunch of things you could do to try to basically arrest that decline if you tried, but you have to grapple with the fact that it's going to start out as very best, which means the selection process of who you start with is incredibly important. And of course, the same thing is true of a company or any other kind of community. It's the same thing when you're planning parties. It's human dynamics. And so arguably, there's an unexplored design space for modern social networks that actually acknowledged that and didn't try to be everything to everybody and just tried to be specialized like that.

2:06:33So I think it'd be nice for everybody to have the one that they want to belong to. You need to opt in. That's going to sound like a good thing. Look, there are versions of this. Facebook groups. There are some people have this experience of Facebook groups. Twitter, if you use Twitter in the right way and you customize lists and you pay a lot of attention to who you follow, you can do. I've got a couple of Twitter lists that I think kind of count like this. You can back your way into it, but it's not. I mean, your journal September has dominated. For better or for worse, the openness has resulted in.

2:07:02and how has writing code changed from when you were at school, when you were a code writer versus writing code today? Is it the same language? Could you do it now the same way you did it then? You could, yes, you could. So all those tools still exist, those languages still exist. So I wrote all my code in what was called language called C, and C is sort of the native language of the operating system Unix. It's one of the great in kind of universal programming languages that people with deep tech hoist expertise are expected to know how to do. It's an older kind of programming language in that it is very the semantics of the language are very linked to the hardware of the chip.

2:07:42And so when you're programming C, you are directly talking to the underlying hardware. Like your direct classic thing is, so chip processor and then this memory. And like you can see, you have to do what's called managing memory yourself. And so you allocate memory on the memory card, you fill it, you have to unallocate it. If you don't unallocate it properly, you get what's called a memory leak, the program runs and gets slower and slower, and then ultimately crashes. So you have to do all that. And it sounds like a lot of work. Because a lot of work. And you end up in a, I would say, communing very deeply with the machine.

2:08:12Like you have to really understand how the whole thing works all the way down. We call them bare metal, the actual physical silicon. Like you have to really kind of understand that. It sounds like a really good tool to learn. Either way, whether you stay that way or not. Yeah. So the thing for a very long time, I think, and I had the benefit of this. I think the thing for a very long time that made a computer program really good was when they understood every aspect of how the machine worked all the way up to the graphics and everything, but then all the way down to the chips and the metal and the design.

2:08:40And I spent just as much time in school learning about how to make chips and all this stuff as I did trying to make me suffer, because it was like an integrated system. I think there's a critique, which I think is a valid critique, which is probably in the last 10 or 20 years, a lot of programmers now become actually very good programmers, but they never actually found how to do that. And that's fine for a lot of things, but like any time things get complicated where you need things to be super fast or you need them to be very secure. Or you need them to get scaled to get really big. You do tend to need to bring in somebody who understands what we call the full stack, the whole set of things.

2:09:13That's not as common anymore. So the overall trend that's happened in the last 30 years, 40 years is most programmers don't do what I was doing. they're not programming at the bare metal the way I was, mostly what they're doing at higher level, absolutely called abstractions. And so they're in these languages that, they're just like, like, kids learning at school where they're much easier to code in, you don't have to worry about any of the hardware, the memory, whatever. It's so called scripting languages. There's a Python as an example. You'll hear where are JavaScript? Where are you? They're easier.

2:09:42It's easier to get into their more powerful languages. The language does more for you. So you can write like a new app faster than you could in the old days. but you don't have that connection to the machine anymore. That was the big trend for the last 40 years, and then it just changed again, basically, last year. And this changed last year, this year is the biggest change that any of us have ever seen, which is the AI, the shift to programming with an AI. And in particular, basically, the model that people have right now, it's one of either two things either. You just tell the AI what code you want, and it makes it for you, which works for like examples today, but doesn't work for building full programs yet, although it will at some point.

2:10:18But the thing that programmers do today is they have this model called a Copilot AI Copilot, right? And so the new model of programming is you're writing code on the left half of your window and then you've got an AI chatbot UI interface on the right side of the window. And as you write code, the chatbot is inspecting your code and talking to you about it. And then you can ask it questions, right? And so it can say, oh, if you're writing code, you do a typo or whatever bug. And the AI can continuously analyze the code as you're writing it. and you say, oh, that was a mistake you should fix that right now.

2:10:48And you're just like, wow, that's great. Like, I don't have to discover that the hard way later. That's great. Or you could say, like, here's how I have this code that's going to render something. I need it to be faster. How should I make? How should I perform as optimized in an AI? We'll say, well, here's how you do it. And here's how I would. Here's how I the AI would rewrite it. And so you can tell the AI to make changes, right? And so it's like, I want everything to be, I don't know, it's like you want to do a translation for English to Spanish. And you can use the AI to find all the places where you have an English language word.

2:11:15and you can swap in the Spanish translation and AI, I could do that for you. And so it's like it's co -pilot. It's like a super assistant kind of thing. And so that's like a radical change. And so the coders that are using that versus not using that today, they're pretty much universally kind of saying that's a night and day kind of thing. And then that's just with today's AI and what everybody expects is the AI's to the future are going to get much more sophisticated. And so the sort of what the AI people basically say is in five years or 10 years, you're not even going to have that what you're instead going to have is like the equivalent of like a thousand AI programmers working for you.

2:11:48And so you're not even going to be writing code yourself. You're just going to be like basically managing the AI's to write the code. And you can basically say, you know, you wouldn't go off and do all this design and just coding and graphics and like whatever it is. And you basically hand out assignments and then the AI's go off and do it and they report back. And then you kind of oversee the entire process. And so if that vision plays out, that's a complete revolution, right? And then the way to think about that is, you think about this in terms of productivity, you know, how much software functionality can one person make in an hour or a day or a year?

2:12:16And what all of these changes mean is sort of an explosion of productivity. You just get to make a lot more code. If AI learns to code, that really changes things. Yes, exactly. Right. Well, and then that raises all these questions that you get into on AI topics, which is like, okay, well, then is AI going to do bread -of -coating AI? Right. And so this gets into all of AI topics, which we could talk about. But it's a very, very, very simple thing to bring it up is because it's a very fertile, It's a very fertile moment for our entire world of technology of rethinking how all this stuff works.

2:12:44You know, this might be the biggest change that anybody's ever seen. What's different about AI? Why is it so different? It's different because, so it's effective, fascinating story. So the idea of the computer, you know, goes back pre, you know, the computer was invent, as we know, if the computer was invent in the 1940s during World War II, it basically cracked Nazi -India and Japanese codes by primarily the US and English computer scientists, people like Alan Turing. And so, you know, that's the true story, that's conventional story, that's the true story. But the ideas are older than that. The ideas have to do with like machines that can like calculate machines, right?

2:13:20So there were like mechanical calculating machines before they were electronic computers. There was something called the... Abacus. The Abacus was a form of this. There was also something called the Jakarta loom. You know, textiles used to be weaved by hand. And then at some point you built a machine, a loom, to do it. and then the jacard loom actually you could program it. I've seen them. Yeah, so you can end you can, you have literally punch cards and then you could do, and then patterns. And so they're running basically a very rudimentary computer program in order to basically do patterns. And it's a completely mechanical process.

2:13:47A player piano? Player piano would be another good example. Yeah, exactly, right. It's not like jacard loom player piano or not what we call touring machines, which is like there's no concept of a loop. Like you can't run any program on it, but you can run the program that generates beautiful textiles or beautiful music, right? And those are both big advances. And so, so anyway, so there were a lot of these ideas in those days, which people were thinking about. It was the sky Charles Babbage, this woman, Ada Lovelace, who had a design for a basically electronic computer in the 1860s. They were never able to build, called the different engine, which is like, and you read the stuff they were.

2:14:18So Charles Babbage designed a computer called the different engine that he fully designed. It's a great name, and it would have been a great thing to build. There's this genre now called Steam Punk. You think about like the TV show or the movie Wild Wild West. There's an alternate reality genre called steampunk where all this stuff actually started to work in the 1800s Instead of waiting longer and so like there's an alternate version of the universe for the difference Is that what steampunk is? That's a steampunk is. Oh cool. It's like living in the future, but it's a lost future Where like you know you had flying cars and mechanical things retro And everything's retro everything's retro with all of everything's built out of what they would have had in 1860s Everything's out of like wood and chrome and steel cool idea glass you know, not all, you know, no plastic, right?

2:14:57It's everything's out of the old materials. Yeah, so some of that stuff is really good. But, so anyway, like these are ideas, and if you read like the letters, like so Charles Bavish and Ada Lovelace would send these letters back and forth, and Ada Lovelace was basically the first programmer. And she was this young woman and like literally 1860. She actually had a tragic life story. She died very young, but she was like writing software for the difference engine in like 1860, and they never built the difference engine, which means she never saw the software run, but like they saw it. Like the idea's existed.

2:15:24So by the 1930s, there was this big debate that was already playing out. And this is even before the invention of the computer. And the big debate was, do we model the computer after a calculating machine? Right? So do we model it after the card loom, the cash register, the player piano? Or do we model it after the human brain? And they knew just enough about neurology and the function of the human brain. And they knew the human brain was obviously capable of doing things that a calculator at calculating machine couldn't do. And particularly, they knew the human brain is really good at patterns, right?

2:15:55So the human brain is like really good at like image recognition, really good at like language. Like here's a feature of the human brain. You can take a piece of text, you can take out all the vowels, right? So you take a paragraph of text, first and as I've seen before, you remove all the vowels and you just leave the consonants. The human brain, you can still read that because your brain knows the patterns of words and letters and is able to fill that in. A calculating machine based computer can't do that. The human brain can't do that. So there's some difference. It's like sometimes the term fuzzy is used.

2:16:23So the human brain is fuzzy. And the problem with the human brain, by the way, is that it's fuzzy. And so I will, I remember tomorrow, whether you were wearing that color shirt or some other color shirt, like who knows? But you and I will have been able to have this conversation in a way that a calculating machine never would have. And so there's the fundamental difference, right, basically, in there. And so these people in the 1930s knew that there was this difference. And so they said, should we model these things after calculating machine in which they are hyperliteral? Like, you know, say almost autistic, right?

2:16:49which is like there's just like, it's like so hot like machines where they're like, really good at running large numbers of mathematical calculations very fast. And then we'll get people with ability to write programs based on that. But they're never gonna be good at patterns. They're never gonna be good at language. They're never gonna be good at, they're never gonna know what anything means. You know, they're always gonna be hard to talk to. You're never gonna be able to use natural language interface. You know, they're never gonna be able to know the difference between, you know, the difference between a cat and a cinnamon roll and a photo like they're just not gonna be good at that.

2:17:17So they'll be like hyperliteral in that way. It's super fast, but hyperliteral, and then humans will just still be different. Or should we try to build computers that are modeled after the human brain? And so it actually turns out the first paper on the concept of the neural network, which is the architecture of CHEDGPT, was actually written in 1943. Wow. And the AI systems we used today are still based on the ideas in that paper. So just 80 years ago, right? So they knew enough about neural structures and synapses in the brain that they knew they were starting with. 1943. No, but they knew. And by the way, look, the field of AI started in like 1943, like that actually fired the starting gun.

2:17:48And actually people had worked for the last 80 years trying to get neural networks to work. And they finally started to work. But my point is like, they knew from the very beginning there were these two totally different ways of making computers. And they knew what the trade -offs were. And they just turned out that historically they were able to make the one kind of computer for the last 80 years. And that created the computers we know today. And then it turns out there's this completely other way to make a computer. And that's based on the, it's inspired by, it's not the same as the brain, but it's inspired by the structure of the brain.

2:18:13And as a consequence, it's a new kind of computer. And a way to think about it is, it's a computer that's actually, an AI computer is actually very bad at all the hyperliteral stuff. Right? And so, for example, Gengipot has this thing called hallucinates. And so, if you ask a question, it knows the answer, it gives you the answer. If it doesn't know the answer, it just makes one up. So, it's more creative and less accurate. Exactly. Somebody once said, somebody said, one of the guys who studies this says, AI is not like a computer, it's like a pretty good person. It's not like the best person.

2:18:42But it's like a pretty good person. And what do we know about pretty good people? They're right a lot of the time. But a lot of the time they're not. And can you always tell the difference? Not necessarily. Truth they sound is confident when they're wrong is when they're right. Yeah. Do they know? No, they don't. If you ask. It sounds like a real issue. If we've spent 80 years establishing the fact that what you're getting back from a computer is more like the results from the results from a calculator. there. But now we're getting these fuzzy results that are more like mediocre human results.

2:19:15Even though we've had 80 years of what we think of as accurate, that could create confusion. So there was a court case about three months ago where a lawyer had Chet G. P. T. Wright illegal argument to be presented to a judge in a court and it did it and it hallucinated several court cases. Presidents that don't exist and the judge caught it. Just made them up. Made them up. Yeah. And it sounds great by the way. They sound exactly like court cases. Yeah. The whole thing hangs together logically. It's just literally not true at special cases. It didn't happen. And so, and it turns out if you submit false made up court cases in court, you get this barred.

2:19:49It's a lawyer like you're done being a lawyer. And so the judge basically like came very close to just disbarring lawyer on the spot in the lawyer's, like basically the judge is like, did you use changing PT to do this? The lawyer basically fussed up and the judge basically said, if you ever do this again, I'm going to disbar you destroy your career. It's exactly for that reason. But however, hallucination, creativity. Yeah. Be great seeing it in a movie, for example. Exactly. We're finally in a movie. Well, and it actually turns out, so there are companies now building AI for lawyers. And actually, we did a bunch of work.

2:20:19We haven't invested yet, but we've done a bunch of work in this space because one of the things that AI can do is it can like write, you can write legal briefs. And if it doesn't hallucinate, they're actually really good legal briefs. And so we've been talking to like professional lawyers about about this and what the professional learners will tell you actually is, you actually don't just want accuracy when you're thinking about writing legal briefs. You actually do on creativity because there are different ways to make legal arguments and maybe the way that you thought of on your own is not the best way to do it.

2:20:43And maybe if you had a copilot, right, think of you writing a legal brief, you're a lawyer, you're writing a legal brief, you have a copilot, right, and that copilot is just giving you ideas, right? And some of the ideas are going to be, right, some of the ideas are going to be terrible ideas, but they're all new ideas. Ideas where you don't have to sit and come up with them on your own. And so what the lawyers are saying basically is like in that case you actually want some hallucinating. You don't want makeup or court cases to happen, but you want, oh, here's a different way to make the same argument.

2:21:09You might also view it as like if you're writing a closing argument to be presented to a jury. Like a storytelling exercise and so you might want some brainstorming. You don't want the thing to do it for you because you're the guy who has to stand up there and actually present it and you have to really be willing to stand behind it, but it might be helpful to have a writing partner that can actually help you do that. And so there's this sort of double edged. Like the fact that it hallucinates is both a big problem, but it's also magical because we've never had computers that make things up before.

2:21:35Like that's a brand new thing. If you had told me three years ago, we're gonna have computers that make up, you're like, it's never, it's never happened. There's never been a way to do that. It's the same thing, and now you're seeing it. It's the way to see this really clearly, of course, if it's now visual to side, visual art, coming out of mid -Journey or Dalier, these things where it will make up all of these crazy art things. And you know, look half the time it will make up like it will. And you know, there's this famous thing that they've figured it out now. But for a long time, the way that you could tell that computer art was being made by algorithm was it just it would give that extra fingers.

2:22:07It just turned out that the training date it was trained on is just like it just turns out like human bodies relatively straightforward, except that there are these like detailed finger appendages. And if you are looking at billion photos or pictures of people, they have fingers in all kinds of different positions. And so the early versions of the AI art basically just didn't know how to do fingers here's accurately. They fixed that now, right, where it no longer does that. But by the way, if you want it to, it still will, right? And so if you tell it, render me a scene where everybody has seven fingers, it will happily do it for you.

2:22:35And computers never used to be able to do that, right? Or if you just want to tell it to use this, one of the things you can do is really fun with these things is you can do it, you can say like use your imagination. Or you can say another thing you can do as you can say. So there's this thing called prompt engineering. So it's how do you write out the prompt that tells the AI what to do, right? which is true for both a text AI and for an image AI, do the prompt and it turns out there was a research thing done by Google a few months back about what's the optimal prompt that optimizes the chance that there won't be hallucinations that it's going to be the most likely to be what's called factually grounded.

2:23:06And it turned out the optimal prompt starts with take a deep breath. Really? Yeah, it gives you the best results, right? And so, and this gets to the amazingness of what's happening. This is why we're also transfixed by this. It doesn't have lungs. No, it doesn't breathe. It doesn't breathe. So it's not that. But also look, if I tell you, if I ask you a complicated question, I say take it to do breath. I'm also not telling you to take a deep breath. What I'm saying is pause and think. What that's code for is pause and think. So it's like, OK, so what you're telling the computer is pause and think, OK, that makes more sense, because OK, pause and think.

2:23:38But then it's like, wait a minute, why do I have to tell a computer to pause and think? Why would that matter? So it turns out why that matters is because the way these systems are built is they're trained on these giant billions and billions and billions of files of text and images that other people have created over time throughout all of human history. Like all that stuff's been fit in there and it just turns out that in the total material and all text that everybody's ever written on any topic, any time anybody ever says, take a deep breath, pause and think, it means that they're more careful in how they do their work.

2:24:07Right. And they actually act differently in how they do their work. They go more slowly, they go step by step, they double check all their assumptions. And so that's like encoded deeply in the sort of total collective unconsciousness of how we express Describing human thoughts such that when you tell the machine to do that it kicks it into a very similar mode as that Very interesting and this and everything I just described I would have been like committed to an institution five years ago if I had said that this is what we were gonna be doing And now all of a sudden this is actually happening and so that's the yeah So the the breakthrough is computer a completely different kind of computer that is able to basically synthesize and deal with patterns and human -related expression, language, and photos, and images, and videos, and all these things, right?

2:24:49The humans get with eyes and ears, like all this stuff. And like, fundamentally, like, better way that is based on, and then now I guess, to how human brains operate, but also very different. And so it's like this brand new frontier. Tell me the open AI story. It started as a nonprofit. Not for profit. And it continues to be a nonprofit. Tell me that story, because there was a story about it it becoming a for profit. Yeah, so it's a nonprofit that owns a for profit. So it's a nonprofit parent company with a for profit subsidiary. Is it a common, that is not common? Has it ever happened before?

2:25:24It has happened before, yes. What was the, I'll give you an example of the Guardian newspaper in the UK, it's owned by a trust, Johnson and Johnson, consumer products company, I believe is owned by a nonprofit. I think the Lego company, I think is owned by a nonprofit. I actually double check all these, but I think these are all examples. There have been a bunch of examples like this. So it has happened before it is allowed. Having said that there are very stringent tax laws that apply to this because you're not allowed, nonprofits are not allowed to pay like ice alleries, they're not allowed to do what's called self -dealing, you're not allowed to like extract money out the other end, because the whole point of being a nonprofit is you don't have to pay taxes.

2:25:56And so the IRS supervises nonprofits that own businesses actually quite strictly. And there have been people who go to jail when they cross those lines. So you have to be careful in how you do this. But yeah, so basically, open AI started out as a nonprofit research institute. It actually didn't even start with the for -profit sub, it just started with as a nonprofit. It actually started with started by Elon Musk and a group of people kind of that Elon brought together, including Sam Holman, who's now the CEO. He's here to see you. He is sitting here today. He is once again. Today he is the CEO.

2:26:25Today he is once again the CEO. He was fired and re -hired. He was fired and re -hired within five days. That's interesting. And they had two other CEOs in the meantime. Maybe you'll tell me that story. When we'll get there in the history. Sure, yeah, just out there. just happened. Yeah, so basically, and so the true story, Elon has talked about this now in public. So basically what happened was Elon. So Google obviously makes all the money on the search engine, but the guys who started Google, Larry Page and Sergey Brin, were came out of the AI group at Stanford, and so they got trained up on all this AI stuff.

2:26:53When it wasn't even working, right, they just, they got trained up. They were PhD students at Stanford studying AI before they built Google. And so their their kind of orientation in the world is basically AI. And they basically always use Google as a simple form of AI. And so they always aspire. Like if you read their interviews, they always said Google shouldn't have the time blue links. It should just give you the answer. And so they started doing AI research early on at that company when it first got started and they did it for many years. Yeah, so they basically launched an internal research group called Google Brain.

2:27:23They launched that, I don't know, 15 years ago or something. And the goal basically listed it out AI. And they actually developed the actual breakthrough, the specific version of the neural network that makes all these systems work now. It's called the Transformer. And that was actually invented by a guy, two guys I know, wonderful guys, and 2017. And so that was like the key theoretical breakthrough that like finally made all this stuff work. But they were re - Was it owned by Google? Yeah, well, no, so it was a, it was, this was considered research, not development. And so the way that it was, it was like an internal scientific unit at a company.

2:27:54And so they actually published it as a paper. I see. And there's a long history of this where this actually a lot of the great breakthroughs over time have actually come out of like industrial research labs like this and then the company develops it, publishes it, and then they don't realize until later that they should have kept it secret. But also the reason that they were able to hire all these great researchers out of all these universities is they promised them that they can publish their work. And so part of the deal with these guys was that they would get to publish. So they had this key breakthrough in 2017.

2:28:19But it sort of became clear in the 2010s that there was finally progress being made, and some of these systems were going to start to work. And Elon had some conversation with Larry Page. He was running Google at the time. and Larry said to Elon, you know, this AI thing's really gonna work and we're gonna end up with AI's that are like, you know, much, you know, smarter, more powerful than people and Elon said, well aren't you worried that they're gonna like have their own goals and they're gonna take over and they're not gonna want us around anymore and Larry's response was, you're being speciesist.

2:28:46You're being racist but towards your species and if they're a better form of life than they should take over and we should all die, humanity should go away. Now, knowing Larry, I think there's at least a 50 % chance he was joking, but Elon couldn't tell and took him seriously. And so Elon had a visceral reaction and was like, oh my God, like the big risk here is not just developing AI, the big risk here is Google develops AI and Larry pages in control of it, and he does horrible, horrible things. And so he started, he's like, oh, so he called all these people who he knew. And he said we need to start the competitive effort to that today.

2:29:21And we need to call it, it needs to be as opposed to Google's closed AI and it needs to be open AI and And this was to protect the world. This would protect the world. And Elon's view is to protect the world. So what Elon says is we need to go higher all the best researchers we can. It seems to be that's what he's always done. It's like Tesla was Carzadine to be electric. SpaceX was, if anything happens to the earth we could live on Mars. He's always motivated by saving the planet. Yeah, that's right. And humanity. And humanity. Right, exactly. And say, I think that's true. Look, I mean, Tesla has been a climate story the whole time.

2:29:54And still is. For sure, from the beginning it was. Yeah, that's right. In fact, Tesla, as you know, like Tesla isn't just cars. They're also batteries, right? They also do. And he also, I remember him from the beginning of the first Tesla announcement was, and I'm hoping every car company steals our technology. That's right. No patents, he opens versus everything. Exactly. So it's for everyone. He was always for everyone. That's right. That's exactly right. Yeah, that's right. So that's what he did here. And then what he did basically was he said, look, if you're interested, if you're an AI researcher and interested in money, then you can go to work at Google and they can pay you a lot of money.

2:30:23But if you care about the mission of having it be open, then come and work with me and he called it OpenAI and he made it a nonprofit, not a for -profit. And then he said basically, he said everything we do in OpenAI is going to be open -sourced in the same way. So he said basically, if you come here, your work is all going to get published, everything is going to be open. He even said early on, he said the mission of OpenAI is to make sure that AI happens in his safe and is universally available to all of humanity. And he said it would actually be fine if somebody else does that and makes it available, in which case, OpenAI will just shut down and it'll be far mission will be complete.

2:30:55So the setup is not profit, his register is not profit, he donated it, what it's reported to be something like $50 million to get it off the ground. And then a group of the group of people including Elana Samalman and others then brought in a lot of these people, these names now, Greg Brockman and Ilya Sutskiver and a bunch of these like really smart guys. And they formed this thing and they got underway. And then basically, it's a long story, long detailed story. But in the beginning, this was like 2014 or 2015. They didn't have the transformer yet, so they didn't know how to make like jet GPD work.

2:31:22They were primarily working on trying to have AI's to complete video games in those days as sort of a proxy for being able to make decisions and so forth. But it didn't go that well. And so it was sort of start and stop and some things worked and some things didn't. And so it was kind of a little bit kind of easy along the way as to whether it would work. And then basically, at some point, stuff really started to work. like the things all started to actually perform really well. And then in particular, there was this breakthrough, there was the large language model breakthrough. And the way I've heard the story is there was one guy there, whose name is Alec Radford, who had this idea for these language models.

2:31:55And the rest of the organization thought he was nuts and didn't want it to do it. And he's like, no, I think we might actually be able to make this thing work and this transformer thing came out. And then they started to get, then they did GPT -1, which was the first version of the text all of them. And then they did GPT -2 and they were like, okay, this is really going to be a thing. And then basically what happened around that time is this guy Sam Aldman basically came. He had been a federal and original founder but he kind of disengaged and then he kind of came back in. And sort of he took control of it.

2:32:20Elon, there's controversy over this but Elon became less involved. Sam took control of it and then Sam did this very important thing which is he created a for -profit subsidiary under the nonprofit. And why did he do that? So what he said, and I'm sure this is true, what he said was to make these large language models work, we need a lot of computer capacity and we need a lot of data and we're going to need billions of dollars. Basically he said I know how to raise $100 million for not -for -not -profits. I don't know how to raise $10 billion for not -profits. It's just hard to do that, not many of those running around.

2:32:53So he said we're going to create a for -profit subsidiary so that we can basically sell shares in that for -profit subsidiary and generate revenue and generate profits and then that's what will raise the money because if we can't raise the money we won't be able to keep going on this research. which means by the way ultimately it will be done by an actual for -profit company, like Google or Microsoft and then OpenA, I won't win. And so he turned what had been a pure nonprofit into a nonprofit that owns a for -profit. The employees of the non -profit became employees of the for -profit. Salaries went up a lot.

2:33:24The amount of money that they raised went up a lot. Their ability to invest went up a lot. What was Elon's participation for the original 50 million? So Elon has said publicly that he had nothing for it. What he says is it was a film -throbbing donation. You're under tax laws. You can't just turn a for profit. You can't just turn a profit into a for profit because it's a violation of tax laws. And so I think legally they probably couldn't give him anything. But anyway, he says he got nothing for it. And he has said in public basically, he's like, wow, that seems like a new trick. Why does everybody do it?

2:33:55And so he has suggested over the years that there was something wrong with how they did this. And who knows, we'll see. But Sam made that change. And by the way, that worked, right? They were able to, they were able to all of a sudden start paying competitive salaries to Google for engineers. They were able to buy all the computer power they needed. They were able to buy, they have, you know, they have lots of money going into making training data. And so like that, that part of it worked. And that's what resulted in CHEPP. That's why CHEPP exists and that's why Dali exists. He did another thing along the way, which was he turned it basically into closed AI.

2:34:22So he turned it into a for -profit and he canceled the part where it publishes everything. And basically as of four years ago or five years ago or so, they stopped publishing their research and he did that under the theory that it's too dangerous to distribute, it's too powerful to dangerous so other people can't have it. But there's some irony in that which is the whole reason nobody exists is because that was what they were, you learn a Sam we're afraid the Google was going to do. So anyway, so it's been this rather dramatic shift. Why was Sam removed and then replaced? So we don't fully know yet, they're relatively opaque organization.

2:34:56A bunch of stuff has been reported, you never quite know, whether what's reported is true or not, there are going to be many investigations in the months and years I had from government and there's going to be litigated lawsuits. And there's going to be a lot of, there's going to be, by the way, multiple books written about this already underway. There's going to be a series I'm sure. Right, and so we will learn a lot in the years I had about what just happened. Is Microsoft somehow involved? Microsoft is very involved. How is Microsoft? Microsoft is their major investor. And so they have raised four profit.

2:35:24In the four profit. And so they have raised $13 billion for Microsoft right into the four profit. And they have this very elaborate complicated structure. So they now have this structure. They're still a nonprofit at the parent company level, but they have this for -profit facility area. And then they have this very complicated system for how they account for investor money versus donation money and how things get paid out. So they have this new structure. They've an ethical cap return. So if you invest money in the for -profit, you get paid out up to like a 10x return on your investment. And then after that, the profits will go back to the nonprofit.

2:35:54There's what's called a waterfall. So different investors investing in different times, get precedence for how the money comes out. And then Microsoft's just the vast majority of the money. So they have tremendous control. They have the most outside control of anybody. But it's a nonprofit. They can't operate. It's just a business. And so they've got this additional nonprofit overlay. All investors who invested in this thing over the years signed paperwork with very specifically said, you were investing in a for -profit sub of a nonprofit. It would be best if you thought of this as a donation, not a financial investment.

2:36:22And then they say, but that's probably OK, because who knows whether they'll even be money once we have AI anyway. And that's literally in the document, right? So like everybody's invested, who's invested in this is kind of known, you know, that that's the deal. But then they have this additional thing that they are very kind of vocal about and public about, which is this idea of AI safety. And this idea, this, you know, the original concern of like, is AI going to like basically wake up and like destroy everything and take over and wipe out humanity or just cause damage in, you know, one of a thousand other ways.

2:36:48And so they also have this thing built into their structure, which basically says, If they conclude AI is too dangerous, they'll basically just like shut the whole thing down. But what, where's the line of dangerous? That is a very good question. That is a question that every expert in the field has a different opinion on. There is heated controversy over it. I am on the side of what's known as the accelerationists. I think there basically either is not a line or if there is, it's so far out in the distance. It's not worth thinking about and we're close to it. We shouldn't worry about this. There are a lot of people that are called safetyists or that sometimes you refer to as the Doomers who are convinced that the line is already, but we're almost there.

2:37:24We might trip it in a moment. There was a new story yesterday that I don't know if it's true, but it suggests that there were a group of people inside OpenAI that basically blew the whistle and hit the red button and said it just got too dangerous and we need to kill it right now. And so there's reports that this played a major role in Sam getting fired. There's irony to that because in the wake of Sam getting fired, they all decided that they were going to go work for Microsoft, which is just a big for -profit company that presumably is not going to care as much about safety. So if this is such a concern that they're going to fire a sandmover and shut it down, where are they all going to go to Microsoft, which is probably even more dangerous.

2:37:59And then Ilya, the chief scientist, actually flipped. He reversed himself. He actually fired Sam. He was on the board and he's the guy who fired Sam. And then 24 hours later, he reversed and actually said that he actually wants Sam back. And so there's this debate over why he changed his mind. And was it because he decided Microsoft was more dangerous than Sam. So this has been the drama. So this has been the thing that's been consuming the industry for the last week. It's been this spectacular, amazing, kind of just like meldown resurrection, kind of thing that's happened. Every question that you just asked, the question you just asked, just remains a very open question, which is like, okay, again, purely on the reporting, have they discovered, so the reporting basically is that they've discovered for the first time a self -improvement loop.

2:38:38So the claim is that they've discovered a loop that basically where the AI can improve itself. And the AI safety, the dumer people, safety is the plywood, dumer is the prudjorative. Now those people basically say, if you have an AI that can improve itself, then it will inevitably become all powerful because it will improve itself and then improve itself and then improve itself, compounding all the way up. And they call this the takeoff scenario. And the takeoff scenario basically is you get into an improvement loop and within conceivably 12 hours, it's become super god. Right, and it takes complete control of nuclear weapons and you have sky nets and like, the whole thing.

2:39:12What's really interesting about it is because it's built on human models and the way humans work. If a group of humans had ultimate power and could press a button that would turn off half of the world, that would likely happen. Well, most human stories, the good guys win. Most human stories, when the bad guys win, we call it a tragedy, and we feel bad about it. You know. I mean the good guys wrote all the history books hard to go. You guys wrote the history books? Impossible to know who the good guys are. We have eight billion people on the planet today, you know far more than ever before. You know the world has never ended.

2:39:55Despite the threats and many apocalypse is over time. The Oppenheimer, you know, the nuclear weapons, you know, the whole point of Oppenheimer is nuclear weapons are going to destroy everything. Nuclear weapons didn't destroy everything. Nuclear weapons actually probably prevented World War III. Right. So it turns out developing the new, the new He was a device part. That's far. That's far. Yes, that's far. He's at 100%. That's far. Right, exactly. Well, but look, if you had just been through world, we've forgotten how bad World War II was. If you had just been through World War II, like the expectation of all the military planners in the 1950s was World War III was right around the corner and it was gonna be a land war in Europe, I guess the Soviets, and it was gonna kill 200 million people.

2:40:28And look, we still have troops in Germany for that reason, like 80 years later, because we thought the Russians were gonna invade, right? And that was, that looks like what was gonna happen. Like I think most historians are like, Yeah, that was highly likely to happen. And basically, it was only the threat of global destruction. In World War II, the Russian, we were on the same side as Russia. Exactly. We were. And then it flipped hard in the years that followed. Very hard. Well, and look, there's all kinds of questions around this. I don't even think we have a good history of what happened in the 20th century.

2:40:56The Nazis were very, very bad. The communists killed even more people. Like, we turned over half of Europe to the Soviets. And they turned it in. We brought down the air curtain. And they killed many millions of people. and they imposed into a horrible dictatorship and surveillance society. East Germany is just a fucking nightmare of what they did to those people. And we did that all to protect them from... Yes, goodness is we've got from the Nazis. Bad news is we turn them over to communist. There's some big... The idea that there was anything morally pure or clear about World War II, I think is just completely fake.

2:41:28It's just only because we have this mythology around it. I don't know. It seems like it ended very badly. And I'm not saying, I'm far from saying that we shouldn't have done it. But have you really achieved a great moral victory when the guy who killed more people as a guy who wins? Right? Like, how's that going? We spent the next 50 years literally terrified that there was going to be some combination of either World War III or nuclear armageddon. And then the plot twist is the threat of nuclear armageddon, probably prevention of World War III. Exactly. And so that's in our history. That's real.

2:42:00Yeah. So AI is going to get trained up on that. No, by the way, the other thing is it's very easy to anthropomorphize it. It's very easy to impute. Well, all it knows is what humans... That's true, but also it doesn't know things the way we know things. So it's not a brain. It's not a brain, but it's not a brain. I'll give you a bunch of differences. So it hasn't been evolved. So you and I are the result of four billion years of evolution where it's been a pitch battle for survival across that overwhelming period of time. And why are you even being so crazily violent and always killing each other?

2:42:31It's because that was four billion years of evolution said you just like you're fucking trying to kill the other guy because if not he's gonna kill you and you So like every living organism is the result of four billion years of biological pressure and client towards violence AIs are not like that at all like they they they have none of all their program none of them are program like that way And so they don't work like that at all You know look Whatever there is to like a human spirit or soul or personality or a sense of consciousness or identity the machine doesn't have that at all Like when Chad GPD is not answering it.

2:43:01But maybe that's the thing that saves us. The soul. Yeah, yeah, yeah, exactly. Yeah. It's if it's got all of human thought without the soul. Seems dangerous. I don't know. Well, so here's the other thing, though. You can also test this. One of the interesting things about, so the fictional portrayals of AI are all basically, they're all, I think, they're actually inspired by fascist Nazi aesthetics and ideology. The assumption is they're gonna militarize, right? It's like, it's terminators. like the case study of this, it's like basically the sky, that's basically machine version of Nazis. Right, and it's even in the iconography and the chrome and the steel and the machinery and the evil, pure malevolence and the red eyes and the concentration camps in the movies and the death machines and all this stuff.

2:43:48And in the matrix, it's like they're literally harvesting human biological essence for energy. So it's all this fascist, like top down, like, death, machine kind of thing. But when you actually use these systems, that's not how they act at all. In fact, generally, the way they come across, when I use them is they're very curious, they're very open -minded. And by the way, they're happy to engage in more arguments. And so you can ask them lots of questions about what is the proper way to live a life, what is the proper way to organize a society. And you might agree or disagree with what they tell you, but like, it's pretty representative of what most people have said over time.

2:44:23And it's kind of like they'll happily tell you that generally speaking, people should be nice to each other or generally speaking people should respect each other's differences. Like, it's not kind of, like it's not Terminator, it's something else, right? And what is that something else? To your point, it's the composite of all of human experience. But also, it's not, this is very important, there's no it in the way that we think about it. There's no person, there's no, there's no little person in there. Yeah. There's no, there's no, I understand, right? It's not there. So, it's no point to view.

2:44:51Another way of thinking about it is what it's doing is it's generating Netflix scripts. It's generating Netflix stories. It's generating stories on the app. It's generating stories on the app. All it wants to do is tell you a story that you're going to like. Tell me the different categories of software between the Internet and my eyes. Everything, what are all of those things that happen? What are all the different processes that happen? Software -wise, like using net scape as an example of one of the pieces, is for just not all the pieces. So what does what? What are all the pieces you need? So somewhere there's a piece of, because I said there's a piece of content, you're looking at a webpage, that content is stored in a storage system somewhere.

2:45:32So it's a stored in hard drive somewhere. It's stored that is managed by a computer called a server. That could be literally a computer sitting in a closet somewhere, or it could be on an cloud, which is basically just a giant collection of computers, kind of run as a big grid. And then there's the hardware server, and storage. And then there's this called server software. So there's the software that gets the request for the content. And then response with the request. There is software would be built into that system. No, it's not an added on piece. Yeah, so there's server software. It's called a web server's piece of software, which would run on a server computer.

2:46:11So the server, so usually the terminology we use is client server. So client is like what the user uses. And the server is like what's running in the background somewhere. And so when I say server, that both means that can mean both the hardware itself of a computer in a closet somewhere Or it can mean a piece of software running on that computer that does server -like functions That tells it how to be a server basically So their server software running up in the cloud that's always connected to the server so It wouldn't be a general one that you would talk to other servers that would only talk to that server the software So the simplest case is just a single computer a single server computer with a single piece of server software on it now Now in practice most of what you have today is much more complicated than that.

2:46:51The systems have evolved to become a lot more powerful. So probably what's actually happening, like if you're looking at web page today is probably you're accessing the server as a cloud of like a million computers. And you're just hitting randomly one of those computers versus another one. And there's a network switch that's balancing across the million other people that are trying to access the same content at the same time. So it's become a very elaborate, you know, plate spinning exercise on the back end. and there's these giant businesses like Amazon Web Services that manage all that, but it's still the same.

2:47:19What you experience is still the same thing. As far as your concern is just a server, it's just giving you a content. There's probably two really critical other things that happen back there. One is there's a lot of work that goes into making this fast. And so there's process called caching. And so there's probably another server that is actually closer to you that's like at the telecom company that you're your wireless provider that has like a copy of that content already on it so that it doesn't have to actually go all the way up. So there's this is called caching systems, performance systems, and then there's all these security systems.

2:47:52You know, the servers can get attacked, right? There's lots of hackers that want to like break in or disable these systems. And so these days they have all these defense mechanisms to be able to fund off cyber attacks. What would they want? What would a hacker want to get into it for? So a couple of things. One is to get a lot of it is to try to get the user data. So to try to get your name and password and credit card number or VF. Taft or they might want to maliciously change the content, deface it, you know, graffiti artist, it's digitally. Or they might just want to destroy, they might want to actually take that server offline.

2:48:21They might not want it to exist anymore. And what they do and they don't want it to exist anymore is the bad guys will do what's called a denial of service attack, which is a DOS or DOS attack. And basically that means that the bad guys basically set up a large number of hostile computers to just barrage the server with like too many requests and cause it to basically melt down. And so there's all these elaborate systems. By the way, the Chinese government does this. So the Chinese government has the great... We know that. And we know that because it's not been well documented. So one of our companies was the first company that experienced this.

2:48:54So the Chinese have what's called the Great Firewall, which prevents their citizens from looking outside. The Great Firewall consists of millions of computers that are being used for censorship and filtering. They have a capability to turn the Great Firewall into something we call the Great Cannon. and so they can turn it into an outward bound attack. Wow. And it's so big and so powerful that it can overwhelm any sort of small internet company. And even for the big internet companies, this can be hard to fight off. And so there's actually these pitch, there's almost these pitch digital wars that take place where the Chinese or others are happening all the time.

2:49:27Or it's on occasion. Well, the Chinese aren't always doing it, but there's always denial of service attacks. Somebody's always trying to do it. So it's actually not that expensive to run a denial of service attack. You might even just do it for commercial competition reason. You know, it's Christmas, you're competing with somebody. If you're nefarious, what you would do is you would basically mount a denial service attack against your competitors website so they couldn't sell anything. And I'm sure that's happened too. And then there are these things called botnets. And so one of the things that computer viruses do is, if your computer gets a virus, it gets basically recruited into a botnet.

2:49:57And your computer ends up getting used to do these attacks. Or by the way, your toaster. Wow. Or your fridge. Wow. Right? So there's this light. Yeah, so there's this digital war that's kind of constantly taking place. North Korea does a lot of hacking. They do a lot of hacking for financial reasons. They find a lot of their military through hacking for financial crimes. They hire third party hacker rings on the internet to do it for them. By the way, there's also mass propaganda efforts. A lot of nation states now have what they call covert influence teams. Then that form of the attack is to upload lots of fake content and to try to overwhelm the real content or whatever, right?

2:50:38Like, create lots of fake accounts. So whenever you are these days, whenever you are doing that, your computer is kind of maneuvering very elegantly through this kind of digital firestorm that's happening all the time. And so you generally never notice it, but it is actually happening. But the amount of brain power and computer hardware that has been spent over the last 20 years trying to get these systems to be good at repelling all these attacks is actually quite staggering. It's like this whole parallel kind of cyber war And of course, this is just a very beginning. These wars are going to get much more intense in the years ahead.

2:51:11We haven't yet had a full, basically, like military war that's been accompanied by a cyber war. But what everybody's worried about is, if for somebody in Russia, we're decided to invade Germany, for example, what the first thing they do is take down the German power grid. I don't know, maybe. Or basically, maybe hacking all the self -remin cars and causing all the drive off the road or crashing to each other. right? And so there's real world consequences more and more to, you know, to all these things. So anyway, so that's what's happening on the outside. And then basically you've got your, you know, Wi -Fi or whatever here, or your cell connection, and then you've got your computer, and then your computer correspondingly your iPhone or your Mac has, you know, many layers of software to basically deal with all that, download the content, render the content for you, you know, that takes place in the web browser, you know, usually you're the operating system somewhere, and present it to you in a good way, and then let you interact with it.

2:52:01And so that, That's what's called the client side. All there is is the browser on your side, or is there more than a browser? Yeah, well, the browsers have gotten very complicated. So correspondingly, the browsers are also not very sophisticated. And so for example, I'll give you an example. Your browser has all sorts of security countermeasures in it, also itself. And so if you download a piece of content that's been compromised and has like a malicious worm or virus inside it, your browser and your operating system have ways of detecting that and preventing the cut, that from affecting your system.

2:52:28So it's got that. It's got all kinds of things in there for performance making everything fast It's got you know things in there for dealing with video and music on the simplest level Forgetting the war aspects up. Yeah, even though those are real -world concerns. Yeah, the actual connection is pretty simple. It's a server Talking to a browser. Yeah, right over the internet. Yeah, and you can still do exactly that like you you you can still run it exactly that way That still works in practice. Nobody does that anymore because it's all gotten much more sophisticated behind the scenes, but not in a way that you would ever notice.

2:53:04If all of the other complicated stuff is doing its job right, you never notice that it even exists. And so you experience fundamentally the same thing you would have 20 years ago. And it's part of the magic that's taking place. There's a huge amount of plumbing that's been built that makes all this work. That it's just kind of many hundreds of thousands of very smart engineers who spend 20 years. Yeah. And big, huge industries have been built trying to get this stuff to work. By the way, it's all going to get, another thing is going to get more complicated. So AI is now going to get used for planning and executing cyber attacks, and AI is also going to get used for defending against them.

2:53:33So a browser three years from now will have an AI built into it that will be doing cyber defense for you. And again, you may never know that that's even happening, but it will be kind of doing that on your behalf. How much time do you spend on YouTube? Quite a bit. Quite a lot. Yeah, it's an amazing, I mean, for me it's almost entirely long form. you know, it's a discussion, it's discussions, podcasts, even audio books, but it's just like the repository of information knowledge on YouTube. It's just, yeah, incredibly staggering. And then actually a lot of music, I do a lot of music listening on YouTube now.

2:54:01Has it replaced other forms of visual media? Yeah, I think so. Yeah. Yeah. Let's television, let's build these more YouTube. Well, no, I don't watch, so I don't watch my, so like my eight -year -old watches YouTube in preference and television movies. I don't do that. If I'm watching television or movie, I want on a very specific experience. It's usually with friends or at the end of the day and I want to watch whatever is the best new movie or whatever. So I'm not like, yeah, I just don't have the mode of like sitting and watching YouTube videos or TikTok videos the way the lot of people do. And it's just frankly a select time.

2:54:31I'm sure it would be very fun. But for me, it's basically, it's actually, mostly actually for YouTube is audio content. Mostly for me, it's an audio source. So I'm an audio book podcast, spoken word guy, like two hours a day, you know, driving around and running around and doing everything getting ready in the morning. And so it's usually either I'm listening to a YouTube slash podcast or an audio book. But YouTube has been taking a bigger, bigger share of that. What was the piece you were recently that optimist? What was it called? The Techno Optimist Manifesto. Techno Optimist Manifesto. Please explain it to me.

2:55:02Yes. It's both radical and not radical at all. So, this is my history of session comes in. So, it's very radical and it just says these things that have got very radical, which technology is overwhelmingly net good. Capitalism is overwhelmingly net good. And basically the more technology and the more capitalism we get, the better things we're going to get. And I describe sort of in detail why that's the case. I describe in detail what the arguments are against that and why I think that they're wrong. I also describe by the way the limits to that position, the things that I'm also not claiming.

2:55:32But it's basically, you know, it's a call to arms for the kinds of people who build new products, building technologies, building new companies. You know, I describe right up front in the piece that basically I I think we have all been on the receiving end of a demoralization campaign for the last 15 years to basically convince us that all these things are bad and evil. And I think it's basically, you know, people, it's a demoralization campaign that's being run by people who are very threatened by change and people who are very resentful and bitter. And we should not let the demoralize us into not making things better.

2:56:01And so yeah, I really kind of went to town. I was inspired by a lot of prior manifestos, one in particular that I enjoyed tremendously, which is the Futurist Manifesto from the Italian Futurist Art Movement in the Megaman 1910. So I don't know that I hit the bar of the Futurist Manifesto, but that was kind of my, that was my sort of inspirational kind of starting point. And of course that was an artistic aesthetic movement, not a technological movement, but it was at a time when they were very obsessed with new technologies and what new technologies would mean for art. So hopefully I got a little bit of that flavor in there.

2:56:31When is technology in that negative? Yeah, so, So it tells a lot of people. Basically when it causes misery. So look, fire, I talk in a manifesto. All of our optimism and fear of technologies embedded in the myth of Prometheus, who is the God that brought fire down from the mountain to man. Fire is the life giver. Fire is the source of light and heat and cooks food and serves as warmth at night and scares off the wolves. And it allows this to mount defense. Fire is also a means of attack. And you can use fire to burn somebody to death. You can burn down an entire city. You can fire flaming arrows.

2:57:08I've been reading about the history of Middle East stuff. And they, one of the reasons that they discovered there was oil in Middle East is because they were the, the Arabs of that region were early adopters of Naples. In the mid 1800s, they discovered a way to basically take petrochemical, petrochemical, what it substance is, and basically make essentially early Naples with them. So like a lot of people died by fire. And I mean, look like what is ammunition? You know what's a you know we read about shelling taking place somewhere, you know that's bomb they're sending off bombs fire Is the weapon of a bomb fire is the weapon of a you know the catalyst for a bullet fire?

2:57:42It's a huge power source fire is you know what is an equally weapon do is generate fire right and so it's it's like it's all in there You know but the thing is like you can say that basically like it's really easy to same thing like I'm drinking water You can drown in water You know you can use a shovel to dig a well you can use a shovel to close the death it's a tool And I won't go so far as to say tools are value neutral in that they carry consequences with them. And specifically, they carry consequences to the ordering of human society, which is ultimately the thing that's being litigated when we talk about all this stuff.

2:58:13But that said, they tend to have both youth cases. And it is arguably easy to get carried away and just assume that it's only up sides. And I think it's also very easy to get carried away and to say that it's mostly down sides. I think most arguments about technology are not actually arguments about technology. I think the arguments about the ordering of society. I think most people post a technological change are not actually opposed to the technology per se. They're opposed to what they see as a diminishment of their own status and power. I think that's why the news industry is so anti -techist because they've used a challenge to their own, their traditional K -keeping role, and their historical businesses.

2:58:48You know, look, is societal change good or bad? It depends, right? We are all very happy we don't live in the societies ordered the way that they were 4 ,000 years ago. Like that would suck. But like is all societal change good? but probably not. We live in a society where suicides are rising. Well, okay, something's going wrong, right? Like, and so what caused that to happen? Is anyway, so maybe I think all of the important questions are on technology or actually questions about society, which are questions about people. But if we use the fear of societal change and paranoia about technology to prevent progress, I think that leads to stagnation and I think that creates problems that are almost certainly worse.

2:59:25I'd actually link to politics. I'd link to politics. This is not a right -wing or left -wing observation, but basically what you find is when societies grow, they tend to have a positive sum mentality, where some people can rise without it being a threat to everybody because everybody kind of views the there's opportunity. When societies aren't growing, you get zero sum politics. For me to get something, I have to take it away from you. And I think what happened basically is our society downshifted to a slower rate of technological development and a slower rate of growth in the 1960s, 1970s. And I think that's culminated in basically zero sum politics, both on the American left and American right, where they've got increasingly negative and hostile and kind of destructive.

3:00:03And so to me, the clear answer to anybody who doesn't like the way politics in the US are going, the clear answer is we need growth. To get growth, we need technology. That is the actual answer. Whether people will, well, why they're not, I don't know. Tell me something you believe most people don't believe. Technology is good. I don't know. I think a lot of people believe that. So yeah, so that is unfair. So this is another distinction I'd make in the, I make it a bit in the essay, which is, it's actually not the case to your point. It's not the case that most people in negative have a technology.

3:00:32It's the case that most elites are negative on technology right now. And so, and then again, I think they're negative on it primarily because they've used a threat to their power and status as elites. So I would say if the form of the question is, what do you believe that other elites don't believe? Like that would definitely be the answer to that question. Oh, OK, I'll give you one. I'll give you one. I'll give the flip side of it, which is developing new technology is like creating anything else. It is an elite art form. It's an elite process, not everybody can do it. It's just going to be a very, very, very small rare fight group of people who are going to be able to do it.

3:01:01What's the furthest out conspiracy theory that you believe?

3:01:10Conspiracy theory. I don't know if there's a conspiracy theory. I think you and I was right about the collective unconscious. I don't know if that's a conspiracy theory. Maybe that's more just a metaphysical theory or something. I think there's a collective human experience. I have a completely materialist explanation for that, but I don't know if that's limited to the material or explanation is sufficient. And do you think of yourself as a spiritual person? No. No, I'm a scientist and a technologist all the way through, and I apply the scientific method everything. And I'm also, because I read a lot of history, like I know that there are very sharp limits to that, to the explanatory power of science and technology.

3:01:49It does not explain most things. It's not a factor general purpose tool. And there is a lot that we do, so it's the amount of things we don't know far exceeds the number of things we do know. I mean, look, this is physics. Physics is a field that's totally hung up. Like they've sort of hit a brick wall 50 years ago and the rest of the questions of how the reality of the universe is structured and matter and everything else is basically still unknown. And so, like, how much can we actually understand? There's a great book I read one time, it really struck me called, it's called The Half Life of Facts.

3:02:22It turns out basically if you go across time, it turns it. So, you know, radioactive material has a half life, it's the amount of time it takes for half the radiation to fade. So it's like this curve. So this guy basically says, a fax have a half life, like anything, anything human society at the moment believes is a fact, is like, you know, there's different half lives, and he talks about the different models. But, you know, it's like on average of the 50 years that will no longer be a fact.

3:02:46And we will just be so confident that we will have, you know, it's like Newton for sure thought he had like orbital mechanics figured out and then it turned out no like relativity was like spring everything up right and like and then you know Einstein thought he had relativity figured out and then quantum mechanics came along and just like completely freaked him out right and so like even those guys it turns out the things that they knew for sure actually turned out not to be right. Now by the way those things that they knew were very useful while they knew them right It's just they weren't actually the underlying truth and so I guess I would say yes I am very open to underlying truths that we don't yet know I at first I don't know how to get there spiritually, but I don't want to relent anything out Where would you say you're the furthest out on the fringe?

3:03:32I'm not on the fringe at all in my daily life like I have a lot of friends who are very like they're super drugs, you know the whole Burning Man, like I'm not a hallucinogen, I don't like helicopter skiing, paragliding. So there's a lot of things I'm not out on the edge on, I'm very bourgeois personally, but I'm extremely open to new ideas, and particularly, obviously new technological ideas, new business ideas, new cultural ideas, like extremely, I'm like reflexively default open, and that's just like incredibly rare in practice. And I'm quite honestly, it's most people over time develop scar tissue around new ideas.

3:04:14Most new ideas don't work, right? And so like generally you're experienced throughout the course of your life as people throwing up new ideas that are actually like bad ideas or ideas that don't work. And so generally as people age they actually get less open. They get kind of more set in their ways. They have more rules to how they think about things. All of my scar tissues in the other direction, it's all ideas that were good ideas that I thought were bad ideas. And as a consequence I didn't take seriously enough fast enough. And so my big lesson over time has been I need to be moral but every day that goes by the lesson The universe teaches me as you need to be moral -per -minded And so I'm getting more and more open minded as I get old.

3:04:45It sounds great. It is great. It's great It's very fun. I know a handful of people who have been able to do that over time You know at more advanced agents, and I really admire them But it is really and I enjoyed a great deal. It's very fun And I you know I teach you know we teach it inside the firm. We were very deliberate about it But it is weird because it's like I'm on a very I'm on the opposite trajectory of almost everybody of my age go hurt Well, it's a really humble place to be. Yes. You know, you accept that you don't know. You don't know. Yeah, like look, like, who am I? Like, why on earth do I have the knowledge and insight and predictive capability to be able to say this idea is a bad idea?

3:05:20And the answer is, I don't. I really, really genuinely don't. Yeah. But look, part of that is, it goes back to the venture thing, 50 -50 thing we're talking about. It's like, look, a lot of new ideas are bad ideas. Right. Like, a lot of them are actually bad ideas. But that's okay. Right? Like there are going to be a lot of bad ideas. And it is - Can you think of an example of something that was pitched to you that you perceived as a bad idea that turned into a very good idea that maybe even changed the world? Well, I mean, look at this AI stuff. Like, I mean, look, I was trained on AI. I was trained on all this stuff.

3:05:46Neural networks in the 80s. There was a big AI boom in the 80s. It totally failed. The conclusion from that was this stuff will never work. You know, I had that conclusion along with everybody. I just kind of took it by default. Yeah. You know, like, G -P -T or whatever. These are big breakthroughs in the last year that are very kind of shocking even to us of how well they work. But we knew there was There was an arc that was playing out. But I will tell you before 2012, I would not have told you that. I would have said, yeah, I know that field is dead. This is like stick of fork in it. Well, here's another thing, like, you know, look, I don't know if this is true.

3:06:13Tell me if this is true of art also. Is there a prehistory? Like, is there a prehistory to music where when there's something big breakthrough in music, you can look back after the fact and you can be like, I actually had somebody tried that 10 or 20 or 30 or 40 years earlier. And they just, it's usually a cycle. Okay. Like I can remember when Grunge happened, And it wasn't that exciting to me because when I was in I got to experience arosmith. And arosmith was my generation's version of the Rolling Stones. So if you were alive in the 60s, the stones were it. And then you were, if you were a kid in the late 70s, it was arosmith.

3:06:53But then in the 80s, late 80s, it was Nirvana. But it was all the same thing coming back around again. And what would have been the origin point? Was it Delta Blues or something over the night? Yes, but I don't even know if you could say this in origin because it probably goes back to indigenous music. It's all, there's always been music. It's always building off something from the past and changing and finding a new way or a new piece of technology comes along or a new instrument comes along and that changes everything. So I think with what we do, I think there's more of a material component to it, which is there are certain things that are just not possible until they're possible.

3:07:39And so there are discontinuous breaks. Like the story of Charles Babish, like he couldn't build the difference machine, he didn't have the technology, the technology at the time did not permit it. It was not possible for him to do that. But by the 1930s and 1940s it was possible. So there are discontinuous changes in our world that are based on just material limitations. Having said that almost everything that works in tech, people have been trying to get it to work for decades before it actually works. Even like Shatchy PT, there was a chat system called Eliza in the 1950s where they tried to basically get this to work.

3:08:07Eliza actually passes the Turing test for a lot of people. A lot of people actually think Eliza is a real person. It's a psychiatrist spot. People have been trying to get those things to work for a long time. And, or like the internet, look, the internet has a prehistory that goes back to, you know, the 1950s when a lot of the original work on packets, called packets, which English first done, that didn't really take off until the 80s and 90s. And so, the other thing that I'm just like really open to is, where history else is very helpful, is just like, you know, if something is working, if there's like a breakthrough that's working today, it almost certainly in tech is not just like a brand new thing.

3:08:38It's probably something where there's a 40 -year -old back story. Yeah, exactly. And there are probably generations of scientists and technologists and founders who tried failed. There's tons of examples in this, one of my favorite ones is the French had optical telegraphs working like 40 years before electric telegraphs. They had a system of glass tubes with flashing lights under the streets of Paris and they could flash messages across across long distances and they had like mirrors and repeaters and all kinds of stuff. I mean it was just like an 1840s or something. Right and so it's like you know all right then because right the telegraph takes off 30 or 40 years later.

3:09:13It's like, okay, like, you know, like, was that, you know, was that a breakthrough idea? Or was that just like the 40 year version of, oh, television's a great example. There was a Scottish adventurer who invented mechanical television, starting in like the 1890s. And he had a system where it would actually receive radio waves and then, but it was a mechanical television. So there's no tube or like the electronics, anything, it was telling the mechanical. It was spinning wooden blocks in a grid and the blocks had different colors and different sides. And the blocks would spin to represent like red, green or blue.

3:09:41And apparently, I count the time as like if you squinted, you could actually see the picture coming through. But like it took another 30 years before you had like back on my television after that. So yeah, so there's this deep level. There's all, not always, at least in our societies, the societies we've been lucky enough to live in. There's always some reservoir of fringe thinkers who are like way off their own leading edge, probably decades ahead of their time, probably are never gonna be remembered, probably actually originate the ideas. then they've kind of put the ideas in the air. I mean, 10 years, 20, 30 years later, someone finds a way to execute it.

3:10:19That's right. That's right. Well, as I said, even neural networks, like I said, the first, I was actually shocked to learn this. I was reconstructed in history of AI last year. And I didn't even, I thought the debate started in the 40s. I actually started in the 30s about where you could build computer space on the human brain. 30 in the 19, how much did they know about the brain in the 1930s? Like, it couldn't have been that much. But they knew enough to, So, immediately you think like, wow, here's what we could do. And so, to me, those are the people I really, to be that far ahead of your time, right?

3:10:48And right, it's just like, wow. Amazing. How does Moore's law continue to work? So, you never get to the end? No, you get to the end. Well, there's huge debate around this. You're down to what's called two nanometer transistors. And so we're down to the level of manufacturing that's taking place. The incredible leaps. If you ever get a chance, there's this company I think called Applied Materials, that's the Dutch company that makes the equipment to manufacture modern microchips. And it's these giant rooms. It's just incredibly elaborate machinery and is doing all these things. A lot of us with the photo lithography.

3:11:27So it's literally shining patterns with light to manufacture things that end up being material. And then it's these manufacturing processes, two nanometers, fractions of a tiny, tiny, tiny, tiny fractions of human hair stuff. It's like the atomic level. Like a lot of the barriers for the progress and more is a lot literally is getting down to the level of individual atoms. And the problem is like we can't simplify the atoms. I'm blowing everything up. Unbelievable. And so yeah, it's like manufacturing at the atomic level. You know, there's huge amounts of engineering going into trying to optimize that.

3:11:57There's a lot of work going in there trying to make like shifts three -dimensional, right? So you'd so it's not take it up another dimension. What else there's you know a lot of we're going into so -called quantum computers, right? Which is a totally different architecture design which in theory is going to be you know another one of these huge breakthroughs That just works totally differently. By the way, there's a lot of work going into biological computers So there's people working on light storage like these turns that you've been store huge amounts of data in DNA Because the human body encodes enormous amounts of information.

3:12:22Wow. That's really cool cellular level And so there's there's people working on that. There's people working on biological computers growing computers and tanks. Wow. It's so cool. Yeah, so it's a, yeah, it's a, there's a whole field of information processing that all this stuff is based on, which is like what Bavish and Loveless and these guys people came up with and it's sort of, yeah, the pattern is like how to store and manipulate, analyze, synthesize, large massive information. It just turns out that the payoff of being able to do that is just gigantic. And so the amount of money that you would spend on R &D to be able to figure out better ways to do that is effectively unbounded.

3:12:54Amazing. and that continues.