#238 Sriram Krishnan - Senior White House Policy Advisor for AI

26:59And in the fly, without even talking to each other, redoing the storytelling and creating, you know, something which any wrestling fan would tell you. If you're watching that, and if you don't have goosebumps, like, you know, something's wrong with you, right? And so anyway, so I'm a big fan of pro wrestling. I still watch it now. I'm fortunate to become friendly with some of the people who do it for a living. I have a lot of respect. I learn a lot from it. Interesting, interesting. You know, we have a saying here. It has us, the whole team. We compare U.S. politics to professional wrestling. So that is true.

27:35Do you see any similarities? Absolutely. First of all, the person I work for, the person in the United States, is in the WWE Hall of Fame. Okay, so let's just start there. And I think that, you know, a lot of people in politics and other parts of the world, other domains, learn from pro wrestling. For example, right? You know what cutting a promo is? No. Okay, so cutting a promo is when, let us say, you and I, I'm a good guy or a bad guy, and we have this big match coming up this weekend, right? And we want to get everyone to buy it. Back in the day, you would pay$60, get it on pay-per-view. You know, these days, you probably subscribe on Netflix or something.

28:18A promo is us trying to hype up the match. And I'd be like, I'm a good guy. Sean, I respect you for everything you've done, but this Sunday, you're going down, right? Or something like that was a bad promo, right? Like, don't judge me. But you kind of build up to it. And you basically put over, the other person got put over, meaning you make sure the other person looks a legitimate threat. Because nobody wants to see you fight someone who's not a legitimate threat. So you've got to put over the other person, right? And then at the same time, you're trying to build interest for this match, right?

28:52And I think if you look at a lot of people in politics, they have learned from that. And even outside politics, For example, Floyd Mayweather, Money Mayweather, he took a lot of how we constructed the TMT, the money character, from pro wrestling. Conor McGregor. Exactly. The walk, the whole thing, a lot from pro wrestling. Because people want to be invested. They want to see the story. They want to be invested in you. They want to see you kick someone's ass or see your ass get kicked because you're the jerk, whatever. But ultimately, you're trying to get people to invest in a story and I guess, you know, watch the fight.

29:33And I think some of the best people to do it, find a way to do it where you're like, man, I'm going to watch history and I got to watch it. So, and maybe I'll stop at this. You know, people ask me, you know, what if I'm going to new to pro wrestling, right? Okay, you're going to teach me how to use a six star. I'm going to get you back into pro wrestling. Right. And the match, I'm going to have two matches. I'm going to have you watch. And these are from 2009 and 2010. And they're Shawn Michaels and The Undertaker. Two matches in a row. So the quick story there is that Shawn Michaels and The Undertaker, I'm sure you know who they are.

30:08We've seen them growing up. Oh, yeah. Right? Now, The Undertaker had what was called the streak. And the streak was the sequence of matches at every WrestleMania that he had won. He was the only person, and at the time, I think he had won maybe 11 or 12 years in a row. Imagine one football team winning every Super Bowl for 11 years in a row. He eventually wound up getting the streak broken, I think, at 18 years. But at the time, the streak was this magical thing. You know, it's like, next person up, who's going to take him down? You know, he doesn't get taken down. Shawn Michaels, right? Like, you know, Mr.

30:47WrestleMania, right? Like, he was somebody who's iconic wins, went up against Undertaker. So WrestleMania 25, right? I think this was 2009. They have this epic contest, right? Just epic contest. One of the greatest matches of all time, right? You watch it, right? You know, it's just a storytelling they do. 100 ,000 people, again, invested. You know, I get goosebumps just talking about it, right? Now, amazing match. Now, what happens after that, Shawn Michaels, basically the character goes crazy. He's like, I came so close and I couldn't get it done. and so he challenges The Undertaker again for a rematch at the next year's WrestleMania a year later and Taker says no and Sean just does a lot of these things to get him to say yes and so he goes on and Taker says okay fine I'm going to give you a rematch but just don't be any match you know if you win you're going to break the streak nobody has ever done this before but if I win you're end your career.

31:45Okay? So, now you can think of this as a wrestling storyline, right? It's like two people in a room wrote it up. On the other hand, it is real because we grew up with watching Taker win every match through our childhood. It was a part of my story growing up, right? It is part of history. At the same time, we grew up with Shawn Michaels, right? Like, I loved Shawn Michaels, I hated Shawn Michaels, I loved Shawn Michaels again. So, you knew when you watch that Wrestlemania, something beautiful, which was real, was going to end that evening, right? When somebody counted one, two, three. And anyway, so you should watch him at the end of the match.

32:26You know, it is huge and emotional. And Sean is beaten down and Taker tells him, you got to stay down, man. Like, you beat, I'm just, you know, it's too much. You got to stay down. And Sean gives him this throat-slash gesture, which basically says like, you got to put me down. Otherwise, I'm going to keep coming back. And then Taker does this big move, wins. And then there is a sense of sadness there because, you know, you're like, wow, this amazing match ended, but Sean's career did really end. He retired after that match. And it was one of those sort of this epic mythological feelings, like these two amazing gladiators who had never been beaten.

33:09So anyway. So when I tell people to watch wrestling, you should watch that. You should watch Rock vs. Sogan. It'll make you feel something. I will. I will. Incredible. Let's move into your story. Yeah. You ready? Mm-hmm. All right. Where did you grow up? I grew up in India in a city called Chennai. It was called Madras when I grew up. India has four major cities, Delhi, Mumbai, Kolkata, and Chennai, and we are in the south. And, you know, what I would call, what Indians would call a lower middle class, middle class upbringing. My mom stayed at home, you know, took care of the kids. She was very focused on family, religious in a lot of ways, very focused on just raising us in sort of the proper way with a lot of right values.

34:07My dad kind of had the same job for 40 years. He worked in this nationalized company. And we're one of those people where, you know, I think my dad, when you're a kid growing up, you don't really think about how your parents are acting towards you is just like the way they are. But looking back now, and especially, you know, we were just kind of talking about sort of our parenting journey in a way, I see how much they gave me in so many ways. So my dad, you know, we never had like real money of any kind, but he always made sure that, you know, we were comfortable and that, you know, we never felt like, you know, we were left out.

34:54But I now know, just knowing some of the numbers, like, that must have not been super easy. I also really respect now because I took for granted then, you know, he passed away, he died in 2006, but he was just there all the time, you know. Took me to school, like, every single day. Came back from work, a long day of work. Hung out with us every single day. And back then, you know, it's like that just, the way things you, you're a kid, you're five, seven, eight, you're like, that's all everyone else. But now I realize how fortunate I was to kind of have that stable grounding experience. And my mom was just sort of this huge source of strength.

35:35You know, she was taking care of us without a lot of resources. And, you know, she was incredibly focused on my education. because when she grew up, you know, they didn't have any access to books or, you know, they had really struggled economically. And she was like, I'm just going to make sure you had a good education. And she would save up money. And, you know, I used to get really into reading a lot. And she would make sure I could always buy things. And, again, at the time, I didn't really appreciate all the sacrifices they were going through to make that happen for me. But I'm just very, very grateful for the experience I had because I think like that grounded me, taught me, you know, what being a great parent who's dependable, who's always there, you know, looks like.

36:26And I've been so much more fortunate in so many ways. But yeah, that was a, I would say I had a pretty great childhood. Yeah, you know, I watch your socials. And, I mean, it's really nice to see somebody in the position that you're in that is so focused on family. I see all the stuff that you're posting with your kids and your wife. And you take your family time very seriously. And I commend you for that. I think you're a very positive example of what it means to be a husband and a father and a family man. And, you know, one thing I would like to talk about with you is, you know, I mean, you're raising your kids here in the United States.

37:06You've done very well for yourself. And that is a – I probably shouldn't assume anything, but I think that looks very different than how you grew up in India. And so could you go into, you know, some of the differences or what your childhood was like growing up in India? Well, very obviously, I've just been very fortunate here. And I think my kids are very young. So I think there's very, very different childhood from what I had. But let's see. So, you know, we, you know, we're one, India has this term, which I don't think is as popular in the United States called middle class. And, you know, in my family, education was a huge priority.

37:52You know, they kind of, I think my family kind of pinned all their hopes and dreams on me in a lot of ways. They were really focused on education and it was very high pressure. If you didn't do well academically, you know, you were really going to struggle or at least that was what I thought. Now, I don't think this is actually true and I've seen a lot of folks kind of, you know, have great successful careers and whatnot, but that was what I was told, like, this is the way out. And when I was 18 or 19, I convinced my dad to buy me a computer, one of these old, again, I'm dating myself here, a Pentium 3 box.

38:31Back when a CPU was one of these big beige boxes that showed up and you stuck it on a table. And it was a big deal for him. I recently reconnected with somebody who knew him at the time, and it cost him like a year's paycheck or something. Wow. This is a big deal, right? And again, my kids, I'm like, you guys have it so easy, right? But it was a big deal for him, and I begged and pleaded for a long time. Now, again, he was a lot of great things. He was a rebel in a lot of ways. I get some attitudes from him, but he didn't really understand technology. He was not of that generation, right? He just didn't understand computers.

39:10But he took a bet. He spent this serious amount of money on me. And then, even though he didn't know what I was going to do with it, and then I convinced him to get dial-up internet, which was kind of a big deal. And I would say my whole story exists because of that and America. And I guess pro wrestling, because that's kind of how I grew up, really understanding America in so many ways. But, you know, I grew up, you know, I would spend every single night learning computer and learning to write code. And way back then, it was a bit hard to kind of get, you know, you can get online and you run up the phone bill.

39:54There's a concept which kids these days don't know, which is you log off the Internet. Do you remember that when you used to log off? I do remember. Yeah. Do you used to like run up the phone bill back home? Oh, yeah. Yeah. And people were like, Sean, get off the PC. You know, we have to make a phone call, right? Like, you have a big download, and then somebody picks up the phone. Anyway, so, kid these days. But I would have to do it late at night because that is when the internet would be faster. Because during the day, other people would be using these phone lines that would be slower. And I would get, you know, all these kind of, you know, coding guides.

40:31I would get all these used books on writing code. And at the time, I was a bit lost, I would say. in what I wanted to do with my life, how I fit in. I wasn't very social and I was kind of lost. But then I realized that this code was something that brought me joy. And one of the things I think people who don't do computer science don't understand is the deep joy of creating something on your computer with computer science code. Because the computer is unforgiving. You have to figure out a way to get it to understand you. You have to solve mental intellectual problems. You never get it right in the beginning, but AI is now helping with that, which we can get to later.

41:17And it is a intellectual exercise. And once you get it working, it's just amazing, right? You just feel so good because you've created something and no one can take that away from you. So I was just doing this every single night. I would stay up from like 10 p.m. till 4 a.m., you know, and then go late to school the next day. I would just be writing code online. And this is where I think two amazing, amazing things, you know, wound up happening. One is my now wife, Arati, you know, she was very similar to me because, you know, she came from, you know, she was kind of one of the first people in her family to go to school in another city.

41:56She was one of the first people to be kind of good with computers. And she was in a different city, and she was learning how to write code herself. Now, at the time, I kind of built a little bit of a reputation myself in my town as the computer science guy. I'd written some piece of code. Do you know what a virtual machine is? A virtual machine? Yeah. No. Okay, so a little nerdery here. If you're going to nerd out for a bit. So people usually manipulate computers using programming languages. You might have heard of that, right?

42:31But one of the things people figured out is if you directly have a programming language access the computer, it might be unsafe or it might be sort of hard to manipulate in lots of ways. So they came up with the idea of this virtual machine, which is a computer that runs inside a computer. And the reason you do it is you're not giving it access to all of it. It's a sandbox, which runs inside a computer. but you can run any sort of code on it. You can write a game, you can write all these amazing things on it. Now, it's a very sophisticated piece of technology because you've got to be really fast because you've got to run it like a computer.

43:08You've got to know all the things a computer does and you've got to make sure it doesn't sort of break out of the sandbox and then does maybe evil things on the actual computer underneath. And for a lot of reasons, I got really into building virtual machines and how they work and how to make them fast, how to make them performant and all these kinds of things. And I kind of had a little bit of a reputation. Think of the guy in your high school who can maybe dunk or athletic and the schools around him. I was kind of like that guy. And so my wife and I, we started chatting online because somebody had introduced us, hey, you guys are these two nerdy computer science kids who seem to know how to do stuff with computers, and can you help me?

43:50And at the time, I didn't even know she was a girl. I didn't know what her age was, but she was like, I was just this friend who has similar interests to me. So we would chat every single night. We would stay up at night chatting. And then after six months, you know, she's asking me, hey, you know, like, who are you? I'm chatting with you. Who are you? And I'm like, oh, I live in the city. It's kind of nearby. And then we wound up meeting a few months after that. First time we met. And we've been together ever since. This was for the last 23 years now. 23 years. Yes. 23 years. When did you get married?

44:25We got married in 2010. But we met each other in 2002, started dating in 2005, 2006. Then in 2010, our parents were like, you folks are both crazy. You'll never find anyone better than each other. You know, you got to get married. And, you know, got married in 2010, you know, had two kids, the whole thing. So, yeah, so I always have to say the best thing computers have brought me is Aarti because without that, you know, I wouldn't have her. and none of this would be possible. The second thing which wound up happening is that I was writing all this code and at the time, there was a Microsoft executive who was touring India and they wanted some student who could do some of the things I was doing.

45:08And somebody had written, seen something I'd written online and I get this cold email saying, hey, do you want to come out and do this little sort of event for us? Because they wanted a student to come do a demo with this kind of this big shot exec. And at the time, you know, I couldn't really, like, speak English. I was kind of living in my bedroom doing these things on my computer. And this is like, this is amazing. So I remember I got on my very first plane ride ever. I've never been on a plane before. And my first fancy hotel room. And we were like, I'm guessing, I was maybe 20 at the time.

45:45It's like crazy, right? In this fancy hotel room, mics are paying for it. It went really well. This executive, who's now retired, but it was amazing, was like, you kids, he met me and my now wife, you kids should work at Microsoft. I was like, sir, we would love to, but we're here, we don't even know how to find you. Long story short, he said, let's find a way to get you in because he was kind of impressed by all the things I was doing. It was very connected to some of the things Microsoft was also working on at the time, like cloud computing. So he was like, hey, there's this kid who's doing these things that we're also working on.

46:20Let's just find a way to connect the dots. And so a year later, they flew me to Seattle, Redmond, where Microsoft was off. And I remember flying in, you know, and Seattle is a beautiful place. You know, the Pacific Northwest is beautiful. And I just fell in love. I was like, man, this is all I want to do. I've grown up, you know, watching Hulk Hogan, watching The Rock, watching Star Trek, watching these movies and I'm now here. And even though I couldn't really speak English very well, even though I didn't really fit in very well, I was like, computers I can do. And let me figure this out. So that sort of started my whole journey.

46:58So I think without my parents, the environment I had, without my dad spending, taking his bet on me, you know, he was not sure what a teenage son, teenage sons on the internet late at night, right? He was like, I don't know what's happening there in the closed door. Like, I don't know. Right. Like, you know,

47:18the door is closed, but they took a bet on me. And, you know, one of the things you realize as you get older is how fortunate you are. Right. Like, so without my parents, you know, being so focused on me and making sure I had a better life than they had, give me these opportunities without, you know, these people who just saw me and they were like, hey, I see something in this kid. Let's take a bet on him. And I would say mostly, I would say this country, right? Every single thing I have professionally, my wife has, has been possible because of America. We live in Seattle, we live in San Francisco, now live in Washington, D.C., right?

48:03So I'm so grateful for all these things. And I've just been very fortunate, right? Like I've been in the right time, right place, and I've been very, very fortunate. Yes, yes, yes, you have. And wildly, both you and your wife are the wildly successful. Are your parents still alive? No, my, you know, my father died in 2006, and not, and my mom died three years ago. And I had sort of, I think, you know, one of the things you don't realize, so 2006, I was, what, 20 or 23 when my dad passed away. It's kind of a big regret in a way because I never got to show him some of the success I wound up having later.

48:47Like, I think, you know, a lot of ways you want to show your parents that, oh my gosh, mom, dad, like, look what I made of myself. Look what I've become. And at the time I had just gotten this job and he knew it was a big deal because it was, you know, I was doing these cool things and I was in the newspapers and whatnot, but I never kind of got to show him, you know, that, man, like, you know, all the investments that you made in me, right? Like, you know, look, I've done something with that, right? I also, I kind of missed out on having an adult relationship with him, which I think is quite important.

49:27Right? I do think, you know, I grew up in my 20s with my dad. So, and obviously, look, a lot of people don't have parents at all, or they didn't have the benefit of the childhood I had. But I sort of feel like I missed out because I never – I was a kid growing up. You knew your dad in that facility. But it saddens me that he never saw me have my career, never saw me get married, have my kids. So I never had a sort of adult relationship with him. I never got a chance to provide for him. So he was a rebel, by the way. So he, back in the community I grew up in, right? It was very community oriented, et cetera.

50:05But you're kind of supposed to stick to your lane, right? But he was a bit of a rebel. He was like, I want to be a writer. I want to write my own books. He wrote, he was kind of, he had all these creative ideas. I think if he was just, if he didn't alive five years later, he would have self-published on the internet and, you know, probably, you know, being on YouTube comments with conspiracy theories. He was that person. but he missed out, right? And, but I think I have some of that nature in me to be like, to speak out. So one of the things he really taught me is that, you know, a lot of people would be like, you know, if you see something wrong, you mind your own business.

50:43But if you saw something wrong in our community, he'd be the first person to go speak up and try and help one. And as a kid, you're like, man, like, I don't want to get into this trouble. Like, what is this? Like, I don't want to keep my head down, but he was always good at that. So I miss him. I feel sad that I didn't get the chance to kind of show him everything and maybe provide for him. My mom, she passed me three years ago, but I'm very fortunate, and she got to kind of see my family. She built a very close relationship with my wife, my first kid, and yeah. What would you say to your dad if he was here today?

51:16Man.

51:24Think about this.

51:32that I somehow made something of myself. I don't know if I'll say it or show it to him, but, you know, like a lot of parents, you know, he was very proud of me, but, you know, sometimes he would have friction and he'd be like, and I just want to show him that in the very imperfect way I am now, you know, I've sort of made something of myself, right? And a lot of it thanks to him and a lot of the bets he took. And my hope is that somehow, someway he knows that. But that's the one thing I always feel like I never got a chance. You want to buy your parents cool stuff when you make some money, right?

52:10I think people ask me, what is it when you ever make some money, right? And I think one of the best things to do is to go buy your parents something ridiculous that they will never buy for themselves. and I got to do that with my mom, by the way. So, and I never got to do that with my dad. So I always feel like, you know, but I think the deeper notion would be to be like, hey, it kind of worked out, you know, I made something of myself and thank you. What about you? Like, you know, how do you feel about your relationship with your parents and how it has evolved over time? I think, you know, I think, I think me and you share a very similar sentiment and I just always wanted to prove to my parents that I could be something and do good for the world.

52:55And it's something that I took on at the age of 18 when I joined the SEAL teams. I mean, that's really what pulled me through. You know, yes, I wanted to go to war. I wanted to fight for the country in the highest capacity possible and be the best that I could be because I was a failure up to then. And I didn't make good grades. I wasn't very athletic. I didn't really have much going for me. And so, you know, what really pulled me through all that training and got me in was I had a horrible fear of telling my dad that I failed again. And so that stuck with me at 18 and Bud's. That's what got me through.

53:40That's what got me into the agency, you know, was I wanted to, I just always wanted my parents to know, you know, they did a good job. Yes. That I could make something of myself regardless of how my childhood went. And my parents are still alive and we have a very, very close relationship. How do your parents react when you first signed up? Most people don't realize how much their personal information is being bought and sold every day. Data brokers are making billions, pulling details about you from public records and the internet, and then packaging and selling it, usually without your consent.

54:17That's how your information lands in the hands of scammers, spammers, even stalkers. It's why you get endless robocalls and why ads seem to follow you everywhere. That's where Aura comes in. Aura actively removes your data from broker sites and keeps it off. They also instantly alert you if your information shows up in a breach or on the dark web. But Aura goes beyond data protection. With one app, you get a VPN, antivirus, password manager, spam call protection, dark web monitoring, and even up to$5 million in identity theft insurance. all backed by 24-7 U.S.-based fraud support. Other companies might sell just credit monitoring or just a VPN.

54:53Aura gives you all of it, together, at the same price competitors charge for just one service. Start your free trial today at aura.com slash remove. Protect yourself now at aura.com slash remove. Well, what was that conversation like? That conversation went, when it came out, like I said, I didn't make good grades. And I had a lot of potential. I was a smart kid. I was really good at math. But I didn't apply myself because I wasn't interested. And it came out one day. He was really frustrated. I just got a report card or midterms or something. And it was not good. I know that feeling. C was probably the best grade.

55:41And he said, it was an argument, and he said, I'm not going to pay for your college. I'm not going to do it. He's like, you're not going to apply for yourself. And I said, I don't give a shit if you pay for my college or not because I'm not going to college. I'm going to become a Navy SEAL. And at that moment, the whole, everything changed. And he, you know, he, I mean, I think he thought it was full of shit because I didn't follow through on anything that I'd said that I was going to do. And he had asked me, you know, how serious are you about it? And I told him, I said, yeah, I've had multiple meetings with the recruiters already.

56:21I've already talked about it. I want to sign. I was only 16 or 17 at the time. And why did you want to be a SEAL? I don't know. You know, I just always grew up playing G.I. Joe's, watching G.I. Joe. Oh, yes. I mean, your thing was professional wrestling. G.I. Joe, actually, G.I. Joe equally. I'm trying to get my kids into it right now. We have a, do you have a, sorry, what was your favorite G.I. Joe toy growing up? Ooh, Snake Eyes. Perfect, man, badass ninja. Yeah. Snake Eyes was a storm shadow, right? Yeah. And the rich kids would have the Sky Striker jet plane. Oh, man, yeah. Sorry, yes, G.I.

56:57Joe. But that's what got me. I think that – so I was always, you know, and early on in my younger years, fourth grade, my dad took a job as a pharmacist in the Army. And we went to Germany, and that's when Desert Storm was kicking off. So I was always – anytime we were at the bookstore, I was always grabbing all the magazines and books about what was going on over there. and I would just look at the pictures of tanks and helicopters and planes and military war fighters. And I was always out in the woods building forts, carving spears, making bow and arrows and shit like that. And so that just always got me really interested.

57:43Originally in my military career, I didn't really care. I just wanted to go fight in a war somewhere. So I wanted to be a Marine. Then I started looking into that, found force reconnaissance, talked to the Marine Corps at the recruiting station. They laughed me out of the office because I was about 100 pounds soaking wet. Went to the Army, wanted to be a Ranger or a Green Beret, same deal. Laughed me out of the office. And the Navy recruiter sticked his head out and said, hey, you want to be a SEAL? And I didn't even know what a SEAL was. because I was also very into all the Vietnam generation movies like Apocalypse Now, Platoon, all that kind of stuff.

58:24And so they gave me a pamphlet on being a SEAL, and I read it and went to the library, checked out a bunch of books, started watching movies, and I was like, that's it. That's what I want to do. And luckily wound up making it through and followed through, and that took me into the CIA and had an awesome career there. And then somehow I wound up podcasting. Man, well, thank you for everything you've done. But I'm so fascinated with just how you signed up because I'm very curious about why people pick the careers they do, especially for something like that because when you're, how old were you when you signed up?

59:0517. Right, so super young. You don't really have a sense of the world. Maybe you did because you kind of been around the environment a little bit. I'm curious about what is it that makes someone be like, all right, I'm going to do this. I'm going to serve my country. I'm going to put myself in harm's way. What do you think some of the general motivations are when people go like, I'm going to go sign up? I mean, I think it's always different, especially in the SEAL teams. I mean, I was one of the maybe – no, I wasn't the youngest. I was maybe the second youngest guy in my BUDS class. and you know when you go to budge you have all these you have guys that are coming in from other special operations units that had already been to war and uh they're just they're trying to to operate at the highest level as they possibly can you know and and what they think they want to do and so you've got 18 year olds up to 30 year olds you know that are trying out to become a seal.

1:00:05And, you know, for me, I can't say that I was overly patriotic. 9-11 didn't happen until I was already, it happened to like right after bootcamp. So I was already in, but even when 9-11 happened, I mean, I knew we were going to war, but I didn't, I didn't understand what that meant. So like I kind of said at the beginning, for me, it was, I mean, that's just what I wanted to do. It wasn't necessarily just for the country. It was what I wanted to do. I wanted to experience. I just wanted to experience that at the highest level. And then on top of that, even more than my own desires, I just wanted to make my parents and particularly my dad proud.

1:00:57because my brother and sister were good students, good at athletics, better than everything that I was doing. And I could see that. I could see the interest, especially in my brother. My brother was a really good ball player, baseball. And I could see my dad gravitating towards what my brother and sister were doing and kind of like, well, Sean's just the fuck-up of family. and I knew I needed to, you know, for my own mental health, change that. I just wanted my dad to be proud of me. Was there a moment through your career serving or elsewhere where you're like, man, this is the moment I know my parents are proud of me?

1:01:41Yeah, after Hell Week. After Hell Week. Do you know what Hell Week is? Okay, yeah. But thanks a lot, by the way, to your show and hearing people talk about it on your show too. Oh, cool, cool. But yeah, that was the moment. That was the first moment that I felt like, man, even though I'm only four weeks in to over a two-year process, I was like, that's the biggest hurdle. And so my dad was my first phone call. And I said, Dad, I made it through. The rest, it's going to get easier from here. It didn't get easier, but that was the big hurdle, and it felt really good. And then graduating BUDS, getting into the SEAL team, first combat deployment, finally getting into contracting at CIA.

1:02:35I mean, it's always in the back of my head, even still today with the interviews that I do. I take this very seriously, and I've made a lot of mistakes in the podcast world. But in general, you know, I just want to... And I have an ability to get stories out and make people comfortable during these interviews. And it feels really good to be able to take somebody, like I was telling you at breakfast, to take somebody that nobody's ever heard of, who's starting a business, who's been through struggles, and get them in here and get them vulnerable. especially with some of the special operations guys that have that have you know there's a cost that comes uh to doing that job both mentally and physically and uh and there's a culture within that and you know the culture within the SEAL teams and the special operations and even just the military community in general is not one that's widely accepted by civilians I mean there's a lot of drinking, debauchery, womanizing, fighting, bar fights.

1:03:52I mean, and now I can't even remember where I'm going with this, but to get somebody vulnerable to talk about what that's like and suicide attempts and stuff after. And a lot of these guys, they get out, they don't have a voice. they aren't able to document what happened over there and in and to showcase all of that you know what their career was like what it was like getting out the suicide attempts the drug addictions the the womanizing the the the infidelity and to have somebody come on here and talk about all that and how they got out of it you know because a lot of veterans feel trapped it's they don't fit in with regular society, you know, and regular society is fascinated in the lives of what me and my, you know, former colleagues used to do.

1:04:50And so to open that up, you know, it's like giving the American, not even just the American, the world. I mean, it's a huge podcast now, people all over the world listen to, to be able to get a peek behind that curtain on what that life actually is like from somebody that's lived it is awesome. And then the audience gets attracted to, they're invested. I mean, we go from childhood all the way to current date. And to give somebody a glimpse into what that journey looks like is, I mean, before this podcast, it was unheard of. You didn't get that deep. And so when the audience gets invested into a story like that, these guys that are, I think a really good way out of that downward spiral after military service and after war fighting is entrepreneurship is a great segue out because we talked about purpose, right?

1:05:54Breakfast Today and how people need a purpose, whether they were a factory worker or auto mechanic or whatever. And we were talking about AI taking jobs. I mean, for military guys, I think entrepreneurship is, gives you, for anybody, I think it's a, it's a enormous amount of purpose. And that, and, and guys coming out of the military and women, I mean, they have, especially in the special ops community or people that are wanting to be the best at what they can be, what they do. I mean, they're gonna, they're gonna implement that in whatever they do, but they don't necessarily get the traction, the exposure that they need to create a successful business journey as an entrepreneur.

1:06:41And we've had time and time again, we've had guys in that they didn't have the exposure, but I could see the drive, you know, and I could just see them in the hamster wheel, like not getting any traction. So I'll be able to bring them in, tell a life story, get the audience invested in their story and who they are as a person it just you know some of these stories are so wild and and and real that they don't care what your business is they just want to see you as a human succeed for all the sacrifices that you've made in the country and it brings other veterans hope and it really brings anybody hope it's like man if this guy's dealing with this shit then maybe i can And maybe I can push through what I'm going through.

1:07:30And we're able to turn a lot of startup businesses into multimillion-dollar businesses really overnight. And that's what I love to do more than anything. Or take somebody like you. I mean, you have a huge name in the tech world. You're very connected. You're very successful. but I think there's still a lot of people that don't know who you are, who you are as a person, all your accomplishments, where you came from in India in a middle-class family in India. I think a middle-class family in India looks a lot different than a middle-class family here in the United States. And so to be able to bring somebody on like you, who I wouldn't say you came from nothing, but very little.

1:08:22And to be able to come to the United States and build what you've built, you know, the life for you and your wife and your kids. I mean, that is going to bring at least one person that's watching this hope, you know, and because we live in this, you know, we live in this society where it's become very popular to victimize yourself and make excuses on why you're not finding success. And it's always somebody else's fucking fault, right? But you control, I'm a huge believer, and you control your own destiny. And when you find what your gifts are, what you're good at, what you're interested in, I don't believe there are limits.

1:09:04You know, I think, yes, there are limits. Like, I'm not going to be a professional wrestler. You know, I'm a - We'll get you working on it. Don't cut yourself short, sorry. Sorry. But, you know, I'm a buck 80. It's just not going to work. But I find the things that I'm interested at, that I'm good at, and I don't think there are limitations. I think the sky is the limit. You can probably go higher than that. And you are a perfect example of that. Thank you. If you have the drive and the work ethic and you don't spend your time looking for excuses on why you're not successful, you will rise to the top.

1:09:45and stories like yours or Palmer's or the Veterans or Blonsdale or a lot of the people that I've brought on, a lot of these guys, they didn't come from generational wealth. They built something out of damn near nothing. That's America right there. That's it. That's it. And that's why I do it. And so long circle back. I just want to say, well, thank you. But, you know, one of the, I don't know how much people can see just the room we are in. But before we started, you kind of gave me a tour. And it's just a powerful space because every object here kind of has some meaning. And so many of them, you know, are from people with just these insane, amazing stories.

1:10:33You know, often people who have sacrificed so much for the country. And this is so powerful. And I think one of the things that you've just done so amazingly well, and I'm not saying this kind of blow hot air up your behind, is create a space where people can share things. I think we were talking at breakfast about your episode with one of your closest friends. And one of the things I loved about that was that just seeing him, seeing somebody who's just such a great sort of masculine figure, but at the same time, opening up about what some of his struggles, dealing with the family, dealing with sort of all the things that he had to overcome.

1:11:17And also in the business world, I was very interested when he talked about how he was navigating the business world, right? And I think, so you just created just this amazing space and platform for people to be part of. So, well, thank you for that. I think what you've just done is just amazing. Well, thank you for saying that. But let's get back to your story. So you go to Seattle and you start working for Microsoft. What did you think of the United States? Oh, man, dream come true. You grow up a lot of places around the world, and you watch this on movies. I grew up on G.I. Joe, by the way. I have a whole big collection.

1:12:00I grew up on Top Gun, Independence Day, Rambo, Rocky, all the classics. So you're watching it on TV, you have an idea. I think the word American dream gets thrown out, and thank you for that. I didn't know of any of that. I just was like, this is a place where I can do something with my skill set. And one of the ways I can maybe make this useful is a lot of times people ask me, what is my advice for young people, people in their early 20s? And I remember when I first got to Seattle and Dreadman, I was just so lost. I didn't have a driver's license. I had never seen snow before and it was snowing.

1:12:45And my now wife and I, we were like, oh my God, like we had to walk a mile to the grocery store in the snow. And you didn't have an SSN for a while, which was, you know, because we had to figure out our paperwork. And we just felt so lost. We didn't I know a lot of people there. But when I walked inside a Microsoft building, and I would see these computers with coding windows, I felt incredibly comfortable. Because I knew that I was very, very good at. And I knew on that there was anybody in the building, I could go toe to toe with them. And I think that sense of mastery is very important. Maybe that might have been my arrogance.

1:13:29Maybe I wasn't like that good. But if you're kind of lost, and one of the things I tell people is find a way to become a master of one tiny niche thing. So, for example, when I was there, we had this sort of this big presentation we had to do for Bill Gates, another senior exec who were kind of running the show at the time. Bill Gates, Steve Ballmer, and this guy, Ray Ozzie. And one of the things I did was I became the guy who, for the thing I was working on, the voice of the customer. What I would do is I would go look at every single internet complaint. Every single, I would call up customers.

1:14:19I'd be like, hey, this is so-and-so from Microsoft. They'd be like, why are you calling me? And I'd be like, hey, I just want to know what you guys think. And I was not supposed to do that. I was on the engineering side. And the reason I tell the story is because over a few months, because I had the technical background and I was building some of these products, but I was also talking to customers on this very niche thing, I became like the worldwide foremost expert on that thing. How people were using it, what were they doing, what are the issues they were running into. And by the way, it's a tiny thing.

1:14:54Like nobody else, I was a master because nobody else cared enough to make them till some master. But the reason it was great for me was when I would walk into a meeting and I'd be in people who are much older. Okay, now they're younger than me. So at least when you're in the early 20s, you know, somebody was like 40 years old. You're like, oh my God, that guy's ancient. A little do I know. So, you know, now I'm older. But, and I had an accent, right? I look different. I'm weirdly tall, as you know. But when it came to this topic, which is about this kind of programming we could do, that was like, I'm the expert, right?

1:15:34And even if the others didn't know it, I knew it because I'd spent every day, every night for months and months, and nobody could take that away from me. So when you have these huge meetings, I once had to make this presentation to Steve Ballmer and these execs. I was comfortable because I had done all the work and I was an expert. And so Microsoft was very intimidating because it had all these legendary people, some of these real icons. But when I built this sort of sense of mastery, I found my way to comfort. So I always talk about it when I talk to young people, or even people are trying to break into the technology industry, which is, you know, the technology industry can seem bizarre.

1:16:16It has all these crazy larger than life personalities, like Elon, Palmer, etc. But if you can make yourself and everybody can make themselves the absolute expert in one area, you're going to feel so good, and it's going to start opening doors. And so that was a big part. The second thing was, I think a lot of people took a bet on me there, and I'm very grateful. So one of the things I absolutely need to underline is I am the result of so many things going my way and so many people taking a chance on me when they did not need to. Okay. And let me name a couple of people. And so this is going to get very nerdy and technical.

1:17:04So sorry about that. But one of the things Microsoft had a lot of were these amazing technical geniuses who were really good at one thing. And one of the most iconic people was a guy named Dave Cutler. And Dave Cutler is probably known as probably the greatest programmer of the 20th century, the greatest one or two. In fact, Palmer will tell you how much he admires Dave Cutler. And the reason he's known as the greatest programmer 20th century is he basically built the version of Windows in the 90s that became really popular. And he's a personality. When I showed up, he was in his late 60s. He's now in his 80s.

1:17:51He was known for being incredibly rude and mean to people. There are legendary stories of him punching holes in walls, of throwing people out of his office. he was not a person who suffered fools gladly, okay? But this guy had an insane work ethic. He had more money than God, but he would show up to work every single day, you know, at age, must have been late 60s now, then, now he's in his 80s, and then right quote, okay? So I remember, you know, and I was like, man, I used to idol worship this guy, right? And computer science people, they still idol worship. I need to find a way to impress this guy.

1:18:30and I didn't know how. I was terrified. And so what I would do is he would show up on weekends. I remember him showing up on December 26th and working, which is not a lot of people. And so I would start showing up on weekends and just see him in the holidays, etc. And then after a while, this young punk is here. And then after like six months, I found an error in his code. This scared the shit out of me because he was he's basically royalty when it comes to programming. I think, I don't know, you're going up against Michael Jordan and you're saying, hey, by the way, your jumper shooting form, I found an error, right?

1:19:12Or Steph Curry, that three-pointer, I saw something last night, I think you can fix. So this is on that level. So I was terrified. But I sort of went into his office and I was like, and he was like, sir, I have this thing and I have this thing and he goes, ugh. And he looks at his computer. I'm like, man, this is the end of my career. And he was known for firing people too, by the way. The end of my career right there. And he was like, you know what? You're right. And I was like, okay. Then I ran out of the office. And, but, you know, the thing about him was he was so intellectually honest. So he took my fix.

1:19:50He made this bug fix in his code. And then he kind of slightly, you know, Him and others took me under the wing a little bit. There are others, a guy named Barry Bond, who was a deep mentor for me and my wife. He would have us over for lunch every day just to kind of teach us things. And so I was taking a bet on it. And Dave was a royalty, right? And it's like Michael Jordan now seeing some punk kid and being like, I'm going to take you under my wing a little bit. So he took a chance on me. And then I had that stamp of endorsement. And other people are like, well, if this guy survived Cutler, at least he can survive this crazy old person, and he's good.

1:20:29So it helped me. So a lot of people took these crazy bets on me at Microsoft. So when I think about Microsoft, I mean, look, Microsoft is a crazy, complicated company, and a lot of people have mixed emotions about it. But at that time then, I think about all the people who took a bet on me, even though they had no reason to. And so now for my wife and me, now we are in the position to take bets on people. And I always think about, okay, how am I trying to spot the 20-, 25-year-old version of me? How would they look like? Where would they come from? Maybe they don't know anything about computers.

1:21:06Maybe they don't know anything about AI. Let me find a way to just take a chance. because I do think one of the most powerful things you can give a young person is this idea that they are capable of much more than what they realized. It's one of the most powerful things you can do. And I have been fortunate to have a couple of people do that to me, but I'm like, wait, I didn't like know, you know, I could, even increasingly, I can do this meeting, right? Or, you know, or I can start this project or I can start a podcast. And, you know, but that was so powerful because you need that belief, right?

1:21:48And I think about now, how do I find ways to give, especially young people, okay, you are capable of so much more than what you realize right now. It may not be easy. It may take a lot of hard work. You may not accomplish it. It's, you know, whatever, but you are capable. And I think that is one of the most powerful gifts that I've been given from my parents, others that I've worked with. And whenever possible, I try and find a way, can I capture some of that and give it to somebody else? You empower people. Well, thank you. People empower me. They gave me belief in myself when I didn't have belief in myself.

1:22:29Maybe they gave me too much belief in myself. Some people would say that too. But it's a gift. And I try and pay it forward whenever I can. I love that. I love that. I think that, you know, I wish more people did that, you know, that find success. And I think a lot of people do that. A lot of people that I talk to in here do that. People have done that for me, you know, but. You had such great stories. I just want to say, I don't want to put you on the spot, but when you're talking about breakfast, you know, you've given so many of these people this platform. And, you know, I don't want to sort of reveal this, but you were like, hey, you are capable of doing this and you know you're capable of so much more so I think you've done justice with this podcast too thank you thank you I mean it's how we make the country great yes you know you pass it on you pay it forward you empower people you give them confidence and uh and just shower them with positivity and you know that it's up to them if they go on to do great things or yeah or continue on the same route that they were but uh I think that's in In some ways, I think that is one of the most beautiful things about America, which is this idea that you have a shot.

1:23:45I'm not going to pretend it is easy for everybody. A lot of people just come from different walks of life, have bad shit happen to them. I think at the core of it, one of the promises of this country is that if you work hard and If you apply yourself and you do all the right things, good things will happen. You have a shot. A lot of other places in the world don't give you that, by the way. You're told you have to stick in this lane or you don't have any opportunity. I do think at the core of this place is the sense of opportunity. By the way, in some ways, we can get the layer, but I think about it with AI.

1:24:24How do we as a country expand the opportunity set, give people way more opportunity than they had before. So that's one of the things I think about when I think about my current job. Yeah. Where did you go from Microsoft? Did you go to Meta or was it Tori? So my wife and I, we idolized Silicon Valley, the culture of Silicon Valley. You know, when we grew up, you know, our first date was I couldn't afford anything. I had this tiny, you know, slightly smelly one bedroom and we had to see each other and the parents didn't we were seeing each other so she would sneak in and we would watch movies about Silicon Valley.

1:25:07One of the greatest movies about Silicon Valley, by the way, is this movie called Pirates of Silicon Valley. It's a movie about Steve Jobs versus Bill Gates in the young heyday and how they competed with each other. We would be like, man, someday we want to be there. And we were like, you know, that was a bit-oriented CD. Like, we didn't exactly pay for that. Like, we were like, we couldn't afford the actual real thing. Hopefully, I don't get in trouble for like 24 years later. But, you know, we were like somebody we wanted because Silicon Valley was the land of opportunity. You went there, you know, you were good at computers.

1:25:44You could make something of yourself, right? So, for me, like, you know, for example, if you're playing football, what do you dream? You want to play in the Super Bowl, right? If you're a pro wrestler, you want a main event WrestleMania, hopefully at Madison Square Garden. But for me, I also want a main event WrestleMania, but I wanted to make something else in Silicon Valley. So anyway, so in 2011, 2012, we decided, okay, Microsoft has been great for us, and we did great. My wife and I had fantastic careers there. We made a lot of friends. We did very well for ourselves. We were like, we want to take a risk and go down to Silicon Valley.

1:26:22We just got mad at two, by the way. And the way we did it is we said, okay, there are two of us. So, you know, we're going to sort of hedge risk between us. And one of us is going to take a chance and build a startup. And one of us is going to get like, you know, go make some regular paycheck and job. And so one of us is paying the bills and the other person can kind of follow this entrepreneurial path. And we can kind of take the pressure off each other. And the deal we made with each other was that we will kind of alternate. Like somebody will go start a company first, somebody will go get like a regular job, and then maybe we'll switch it on.

1:26:56Like we didn't know what we were doing. We were like, we just wanted to be part of Silicon Valley. And I will say Silicon Valley is a magical place. And I think it is one of the, you know, unique advantages America has over a lot of other parts of the world. Because the combination of capital which flows in, the combination of talent there, the density of the companies there, the, you know, the idea of anything. It's not perfect. a lot of issues with it which we can talk about, but it is magical and it is unique. So we wanted to be there. So we quit our jobs at Microsoft and we flew down to the San Francisco Bay Area.

1:27:35We're like, all right, what do we do? And so a year later, my wife, do you know what Y Combinator is? Yes. Okay. So Y Combinator is this very popular startup incubator. It's maybe one of the most popular startup incubators. At the time, they were maybe not as popular as they are now, but they're still a bit popular. What they would do is if you had an idea, you go to them, you apply, and they pick maybe 20, 30 companies a batch. Some amazing companies have come out of that. Airbnb, Coinbase, Brian Armstrong, I think talked about my Combinator here, Dropbox, lots of great companies that come from there.

1:28:16And so my wife had this idea, she applied, she got into YC. And YC, by the way, doesn't give you a ton of money. I think at the time they gave you for the entire company, maybe like$80 ,000 or something. So it's not a lot like, you know, and I was like, okay, I need to figure out, you know, how to, you know, just get a paycheck, you know, get some regular money. And so I wound up joining what was then Facebook. And Facebook now called Meta. And some of the older people here might remember this, Facebook then had just gone public. And they were in a dark hole. Because number one, the IPO had gone terribly poorly.

1:28:57They had gone public, I think at like$45, the stock price had plummeted. Second is there was this big question about the whole world is moving from desktop computers to mobile. and Facebook is only making money on the desktop. So there's this big question of, first of all, the idea of social networks making money seemed laughable, right? Like people were just laughing out because people had seen MySpace, they had seen all these other companies fail. They were like, I don't even know why this is a thing. Second was they're like, nobody can make money on mobile. So they were in a bit of a hole. And I had some friends there and, you know, I told myself, listen, I want to do something which is very different from Microsoft.

1:29:37So I want to do something which involves consumers, because I really was interested in consumer psychology. You know, what makes human beings use products, how they interact with products, how to build the technical algorithms. Now, maybe AI, you know, we didn't have that really that word then, which kind of interacts with them. I was very interested. And I just want to do something different. So I got a job at Facebook, and I wound up working on Facebook advertising. Now, it's a monster, you know, I don't know, like, or a$10 company. At the time, it was like the stock was down. And again, I got very lucky.

1:30:12You know, I wound up building this ad ecosystem product called the Facebook Ad Network with some amazing smart people. And we went from$0 to a billion-dollar business in three months. In three months? Yes. I remember, like, you know, I think$2.7 million a day is$1 billion run rate. And it was a rocket ship. And I think, look, again, when you look back now, I think we were lucky for two things. One, people were starting to buy things on their phone, right? Because the iPhone had come out in 2008. I think the App Store came out maybe a few years later, right? And one of my mentors tells me, like, you know, there is no difference between being wrong and being early.

1:31:02They're just the same. But in 2012, 2013, people are starting to buy things. So there was commerce happening. So when there was commerce happening, people wanted to advertise and push their products, Target, Amazon, mobile games. Second, Facebook at the time had actual authentic people, not like bots, which was a huge problem on the Internet. And third was we were able to marry that with these algorithms and products. And the big lesson I learned at Facebook is just like the power of working with great people. So because I had this small team, this guy named Vijay and others, and, you know, who were just fantastic engineers, right?

1:31:46These are kind of people who go off on a weekend and they would rewrite an entire system, which had taken a year for a team, and they do it over a weekend. Palmer Lucky. and I shared a common sort of hero figure, a guy named John Carmack. Did he talk about John Carmack when he came on the show? I believe he did. He did, right? So John Carmack, I think, by the way, one of the idols of the 20th century, John Carmack was the guy who built Doom, the video game Doom, right? Oh man, I used to love that game. Yes, yes. So Doom, and so Carmack is a programming god, right? What he would do is he invented basically what is called the engine which ran under Doom and then under Quake and all these engines.

1:32:31And then eventually he met Palmer because he was very interested in VR and they wound up doing Oculus together. And then, you know, Carmack got hired into Facebook, right? But the reason I bring up Carmack is that, you know, there's this great story at Facebook where they realized, like, Carmack individually was doing the work of entire 200-person teams. Holy shit. Just one person guzzling Diet Coke, right? Like I think he now lives in Dallas, right? Like sitting by himself, just like doing the work of 200 people. At one time, you know, Facebook HR realized like, you know, they had nowhere else to promote him to because they just run out of levels to give him.

1:33:17And he was just a machine. And so the reason I bring this up is it is very hard to make sort of get yourself to the top of an industry unless you know what greatness looks like. So I was lucky because when I joined Facebook, I was surrounded just by sheer serendipity by some great engineers, just like I was at Microsoft, right? So now, even though these guys are, just to be clear, they were a thousand times better than me. They would run circles around me with anything technical, like without even batting an eyelid. It's like me playing like a pickup game against LeBron or something, right? But I saw what greatness looks like.

1:34:01So which meant that many, many years later, when I was starting doing investing or when I meet entrepreneurs now, I now have a rubric in my mind of, okay, I know what elite talent looks like. I see the work they put in. I see how they talk, how they think, how they spend their free time. So now I can sort of like, you know, if you know how like Steph Curry shoots trees and then you know how, you know, the kid you play with at the gym shoots trees, you know, I kind of see what greatness looks like. So I was exposed to Facebook. I was exposed to some real greatness at Facebook. So I wound up doing ads.

1:34:37It became a huge hit at Facebook. and I think Silicon Valley is one of those places where it does a lot of good things but everyone's kind of looking for what is the win you've had what is the thing which kind of says you are somebody who's capable of something and that whole thing gave me that I started to just get known as oh, you know, Sriram's this guy who kind of built all this stuff at Facebook and if you go Google me from back then and I would start showing up in all of these press pieces. It's very important. I think it put me and my wife on the map in Silicon Valley. And so I'm very, very grateful for that.

1:35:18Wow. Wow. You had mentioned something earlier about five minutes ago. I think you said there's no difference between being early and being wrong. What do you mean by that? So I stole this from somebody we should talk about, Marc Andreessen. You know, Mark Andreessen is the inventor of the web browser. He was the founder of Netscape. Did you use Netscape? Oh, yeah. Great, right? So the spinning logo with the stars and all of that. So Mark Andreessen was kind of one of the original boy wonders of sort of technology world. He built this browser called Mosaic, which then started this company called Netscape, which was kind of this darling child of Silicon Valley in the 90s and then got crushed by Microsoft.

1:36:04But then 10 years later, him and Ben Horowitz started this venture capital firm, Atrix and Horowitz, which we can get to. I wound up joining later. But for many years, he was a mentor, is still a mentor for me. And he has a great many sayings. And so when you have an idea as an entrepreneur, right, like I think, you know, there are a lot of things which can go wrong. And one of the things that you have to think about is that, is this the right time for this idea? And history is filled with examples of companies that had the right idea, but were too early and died. Let me give you an example. Do you know what the company Instacart does?

1:36:52It's a grocery shipping service, or DoorDash. What does DoorDash do? right? Like you buy a product, they bring it in from the nearby restaurant or, you know, deliver your groceries, right? Exact same company as Webband in the 90s, right? One of the most famous dot-com busts. People ask you, oh man, what was one of the biggest things in the dot-com era which lost money? It'd be Webband. They were right. They were just too early, right? Or Pets.com, right? Like another famous dot-com bust, but people have figured out later. MySpace, right? Like MySpace or Friendster. Do you remember Friendster? I don't remember Friendster.

1:37:30All these social networks, but MySpace, I'm sure you were on, right? It was okay for a while and they sold to, I think, Murdoch, but Facebook did so much better. And I think as an entrepreneur, you have to think about, is this the right inflection time for my idea. Because when you connect the right idea, the right entrepreneur, and the right time in history, magic happens. I'll give you an example. If you look at YouTube, right? So people think of YouTube as just obviously these days the de facto way to have videos online, right? But when they first came out, it was other people had tried it, right?

1:38:10Like, you know, Google had this effort called Google videos that are trying to get videos online. Others have tried it, but none of them had really worked. But YouTube captured this moment in time because digital cameras were starting to explode. And I think you started to see the increase of original mobile phones, which had reasonable quality, right? And in 2008, the iPhone comes out. So one is you're starting to see more video production, right? Like by regular people, right? Without needing a camcorder or doing those home videos or needing a bulky camera, you're seeing video production go up.

1:38:47Second is broadband around the United States was getting better. Super important to basically get these videos to show up. Like, I don't know if you remember the 90s, you open up a real player and be like, buffering, buffering, buffering, buffering. You watch a minute, buffering, buffering. Like, you know, in 2008, most of those issues are pretty solved and you could watch a video without it, you know, buffering. And then so these guys, like the YouTube founders, like Steve and Jawad and all these guys built this product, you know, which more than anything just captured this right moment in history, right?

1:39:25And they rode that wave. And I think a lot of times the difference between a$500 billion iconic company and some company which runs out of money, you know, is not the persistence or the heart or the effort of the entrepreneur, right? It is just they were the wrong time. And so when you're an investor, which obviously I spend a lot of time as, often I was trying to think about, is this the right time for this idea to happen? Palmer is another good example. So a lot of people had tried virtual reality. In the 90s, there was this thing called VRML where you were like, oh, let us embed virtual reality in your computer.

1:40:12And there are all these movies and TV shows which had virtual reality in it. The Matrix is the most famous one, but there was a Pierce Brosnan movie. There were all these things that had virtual reality. But the challenge was that the hardware was too complicated. Nobody could figure out the hardware. And there was no internet bandwidth to kind of show you these experiences, right? So, what Palmer, you know, and through his genius, through his sort of, you know, he told his amazing story, you know, him sort of like figuring this out by himself, he sort of captured the right moment in time with the camera gear, trying to figure out all the low latency interfaces so that the image that the camera sees is now, or the image that is being projected is kind of reacting with you in near latency.

1:40:56So that was the right moment in time. So I often think about like, You need the right time. The other thing you need is you need some luck and some magical lightning in a bottle, especially with consumer products. I think with businesses and enterprises, it's a bit different. You can go call up your customers and you can be like, hey, what do you guys want? I can build that for you. But with consumers, I do think you need lightning in a bottle. So YouTube, they started taking all these videos. They put it up. I think an SNL video once went viral. They went off to the races. Facebook has an interesting history.

1:41:33Everybody knows about the Facebook history, about Harvard. They've all seen the movie with the social network movie. But one of the other untold piece of Facebook history is that how did videos become a thing on Facebook? Because for the many, many years, Facebook only did photos. This all seems like ancient history. And anybody under the age of 30 is like, hey, folks, what are you talking about? I'm on TikTok, Instagram. but for a long time you posted a photo on Facebook of your friends and you tagged them no videos until do you remember this thing called the ice bucket challenge yes right okay this is internet history so but the ice bucket challenge was basically people raising money and what you would do is you take this bucket of cold water dunk it on yourself right like with a video camera in front of you.

1:42:23And then most importantly, you would then shout out four or five people. You'd be like, hey, I'm going to challenge my celebrity friend to go do the ice bucket challenge, right? It would spread. So I was at Facebook at this time. And Facebook had just launched video, but they could not really get people to upload video or what was going on. But the ice bucket challenge had this remarkable couple of properties. One was that it was video and people had to create it on their own phone. So it was personal. But second, you need to tag your friends because you're tag some friends. Facebook just happened to have a platform where you could post a video and you could tag a friend.

1:43:05You could not do it on YouTube. So I remember being in this meeting with Sheryl Sandberg where there was this vertical straight line and that was usage, right? Because everybody, you know, was uploading videos onto Facebook. And so every platform has one of these stories. With AI, I would say it's ChatGPT, for example, which has a story. But I think my point is that with consumer products, you need writing in a bottle and you need the right timing. And often, you know, when I talk to an entrepreneur, I often ask them, what makes you so sure that now is the right time. Not four years ago, not four years from now.

1:43:43Why is today the right time? Interesting. I never thought of it like that. And your explanation makes a hell of a lot of sense. What, you know, you went to work at Facebook and created the ad network and had a big part in the videos and it sounds like you had a very big part in making it tremendously successful. your wife went into entrepreneurship. Yeah. What did she do? So she started this, I always tell people my wife is the more impressive person of the pair because it is true. She's a multi-time entrepreneur. You know, she's actually been through the journey so many times. And so she started this company which was about renting electronic gear.

1:44:29The idea, which by the way, I think she would say was maybe a bit ahead of its time going backwards, our sort of theme. The idea was that instead of renting, like for example, you're surrounded by several high-end camera gear, but imagine you'd want to buy it. Could you just rent it and then try it out because you're going on a trip or you're going on a photo shoot or maybe you're renting a drone. That was a big deal because drones are very expensive. Somebody was doing an ad shoot or somebody was doing, like, I just want to do this one scene for a day and they would rent out this gear to them.

1:45:02So she raised money from Y Combinator, from a bunch of people, and they did very well for a period of time. I think they had a couple of dozen employees, maybe. They had a lot of customers. They loved them. But I think the challenge that they ran into is, she would say, is a little bit early because now you're actually seeing other companies do that at scale because so many other consumer product categories have exploded which need that dynamic. and at the time, I think she was doing cameras and drones. So yeah, she had a reasonable outcome and I think she had a great experience but she was also like a bit early.

1:45:44But also at the same time, my wife and I, this is when we started to do some angel investing together and she's a fantastic investor. That's what she does now full-time but we had a tiny bit of money. We had a little bit of money we had made from Facebook. And one of the things about Silicon Valley is that you just start meeting people who are starting companies all the time. And we started to learn how to do angel investing. And we would write, I would say, small amounts of money compared to what a lot of other angel investors are. Sometimes we wrote a couple of thousand dollars. and just because we wanted to support someone who was a friend.

1:46:30But I think that over the next four, five, six years taught me a lot about investing. It brought me in touch with fantastic entrepreneurs, some of whom actually been on your podcast before. And then in a way, it kind of led me and her to our investing careers much later. Interesting. Yeah. Very interesting. So why did you wind up leaving Facebook? I was bored. I was a bit like Facebook is, I think there's a pattern through my career where once I've sort of felt comfortable and settled in, I want to be like, okay, I want to find my next thing, which really challenges me and pushed me. So I think they were great.

1:47:16A lot of friends there at the time, some have left, et cetera. I could probably stay, you know, I was making decent money. I would have made more money. But I was just like, I just want a different adventure. I had that problem. Yeah, yeah. I mean, if you look at my, you know, when you're reading my bio, I was like, man, there's a lot of stuff in there, right? Like, you know, there's like this podcast, pro wrestling. And part of it, I think, is I just, you know, I've seeked out different adventures over time. and so I was very comfortable there and also as a part of it, I'm very suspicious if I'm very comfortable because I'm like, man, am I stagnating?

1:47:57Do I need to push myself? I'm also very competitive so I would be like, am I not pushing myself hard enough because I'm just showing up in this job where Facebook at the time has become a big company? They don't really need me because that's how you build these big companies. Like you don't, you shouldn't have to need anyone. That's how these companies are built. So I just, I get very restless when I'm comfortable. One of the good things about my current job, you're never comfortable. But I was just like, I was just bored. I wanted an adventure. And I was kind of, I was advising a few entrepreneurs.

1:48:35I was doing some investing. And, but I would say what eventually that path led me down to is the first of two times at Twitter. Twitter at the time, there's a guy named Jack Dorsey who was running it, who was a founder, and they had been through a lot of really bad turns as a business. One of the things I like to compare is Facebook and Twitter as companies, because at the time in 2008 or 2009, Twitter, by the way, of course, now X run by Elon, which we can come to have had some little involvement there. But Facebook, there was a time in 2008, 2009, Twitter was seen as the potential$100 billion,$200 billion company.

1:49:18And Facebook was like, oh, just this toy social network that just will never spread outside colleges. That obviously flipped. And part of the reason, I think, is that Facebook had a very methodical way of using metrics and data and numbers and experimentation. One of the lessons I really learned from some of the people there was to always be distrustful unless you have the numbers and the experiments to prove it. And if the numbers back it up or the experiments back it up, you should be prepared to change your line of thinking. One of the things I think Facebook has been very good at is changing how they believe, what they believe about things when they have new data, at least the time I was there.

1:50:01Very different company now. I don't really spend any time with them now, obviously. So I kind of learned that mentality about like, okay, if I run this experiment, if I try something, if I learn something different, I'm going to change my worldview, right? They were very good at that. Twitter was very different, right? They had this product, 140 characters. It had worked. And then one day it wasn't really working. And they were a public company. They weren't making enough money. They had a CEO change. So it was a little bit of a spiral, right? Like you're a public company. People are comparing you to the other companies.

1:50:32You're getting like a new CEO every year. to. So it was a bad place to be. And they were spiraling and spiraling. And at the time, you know, I love Twitter and now X as a product. I use it all the time, right? Like, you know, it has given me so much. A lot of my personal relationships have come from that. Professional relationships have come from that. I think it's just a fantastic product for the world in terms of, you know, just what it has done. And, you know, somebody reached out to me and said, hey, do you want to help? And I was like, yeah, sure. It seems like a challenge. The company is in a bit of trouble.

1:51:06It was not glamorous. It was not the sexy company to work in, but I really like it and it seems like a challenge. So I wound up joining there. And that was quite the thing because one of the things I did not realize at Twitter was how insanely political, in many different ways, it was on the inside. So that was quite the adventure. but that was the guy so i joined twitter i think in 2017. 2017 yes you know before we go farther i'm just curious i mean what are your thoughts about just social networking in general i mean

1:51:48it's caused a lot of problems in the world i think it's also connected a lot of people in the world i I mean, I've made a lot of contacts off X, Facebook, Instagram, not so much TikTok. But I feel like it's a double-edged sword. You know, it's also a great way for government, anybody, to kind of map out who you are. It's immediate. I mean, I use it when I get messages on any one of these social platforms. And it seems like an interesting message. the first thing I do is I go to the friends list, see who mutual connections are. And so it's, you know, I mean, you're basically what a lot of people, myself included, I mean, your life goes onto these social networks.

1:52:37It's easy to map you. It's easy to map your connections, everybody you're connected to, who they're connected to. I mean, I'm just curious, What are your thoughts on all of it?

1:52:52Social media has completely revolutionized the world in a multitude of ways. Most people don't realize how much their personal information is being bought and sold every day. Data brokers are making billions, pulling details about you from public records and the internet, and then packaging and selling it, usually without your consent. That's how your information lands in the hands of scammers, spammers, even stalkers. It's why you get endless robocalls and why ads seem to follow you everywhere. That's where Aura comes in. Aura actively removes your data from broker sites and keeps it off. They also instantly alert you if your information shows up in a breach or on the dark web.

1:53:28But Aura goes beyond data protection. With one app, you get a VPN, antivirus, password manager, spam call protection, dark web monitoring, and even up to$5 million in identity theft insurance. all backed by 24-7 U.S.-based fraud support. Other companies might sell just credit monitoring or just a VPN. Aura gives you all of it, together, at the same price competitors charge for just one service. Start your free trial today at aura.com slash remove. Protect yourself now at aura.com slash remove. It's such a complicated question, right? You know, first of all, like, as somebody sort of worked in a few of these things, right, like, I have some biases.

1:54:07Like when you're a part of an institution, right? I've been part of multiple of these institutions, right? Like you sort of have an affinity. You want to defend it. That's number one. But the second part is I've also seen the abuses that happened when social media, you know, has the real power, right? I've seen the censorship. I've seen like the centralized control that can happen.

1:54:36and I've seen the harm that it can do for kids, for example. When we grew up, I didn't have to worry about, was I a popular kid? Would I get bullied if I said something? I have a lot of friends with adolescent daughters, sons, and they have to worry about that. It is a very complicated topic. I think the thing which I often think about is, and one of the things I saw on Twitter, was how censorship or nudging political agendas can happen, sometimes deliberately, sometimes even not deliberately, right? And how they can have huge impact, right? So, for example, people, you know, think of like the Twitter files and, you know, all these examples of censorship happening at Twitter, right?

1:55:25And I was there when some of this happened. And one of the things I always sort of fought back on was the influence of politics inside the social media algorithms. So often people kind of often know, I think, the easy ones. They are like, oh, my account got shadow banned. Or I said something like, for example, in COVID, I said the virus came from a lab and my account got blocked, which is, by the way, the NIH director, Jay Bhattacharya, he said that his account got blocked. and think about his story now. You know, he's now like running the NIH because Twitter deleted his account for basically saying, hey, maybe the virus came from a lab and, you know, maybe I'm not totally sure about the mask thing.

1:56:10And they deleted his account. And what I also saw was like how easily algorithms can be used to shape politics. I'll give you a story. I was Twitter one day and I wake up and I'm scrolling the internet and I see this story. And I won't say who, but it is this famous Hollywood actor. And it says that this Hollywood actor has this movie project that was just announced, but the internet is mad at him for getting this movie project, and he might lose it. And the reason it caught my eye was all these stories would embed the same two or three tweets inside. And I worked on that. But I ran a lot of the algorithms and I was like, these look a bit weird.

1:56:56There's no followers, there's no likes, retweets. How did these get found? Why are these getting surfaced up? So I had some free time and I sent some emails and I poked around. And it turned out, what had happened is, do you know what trending on Twitter is? So at the time, the trending algorithm on Twitter had all sorts of issues, all sorts of bugs. and what it would do is sometimes it would try and tell you, hey, this is what people are talking about right now on Twitter. It's trending. Sometimes we just kind of pick a random arbitrary thing and say it's not trending when it couldn't find anything.

1:57:35And what happened one night, that previous night, was it had found a random set of tweets and because it could not find a legitimate trending topic, this algorithm, by no one's fault, or someone's fault, but not any deliberate agenda, right it said these tweets by total unknown people about this actor is important okay so this happens maybe at like 2 a.m 3 a.m and then in the morning right like what winds up happening is uh on twitter there was this product called there was a way they could highlight these tweets and kind of give big imagery so one of these people wakes up and some of these folks are you know just sometimes had a political agenda but i think they were just like hey i just want to do my job you know and this thing is trending.

1:58:18They did not know this came from an error. So at 5 a.m. or 6 a.m. New York time, they take this and they go say, this is a thing which is happening on the internet. Okay. A couple of hours later, all these editors of all these, you know, media publications wake up. They're like, oh, there's a thing people are talking about on the internet. Hey, can you chase this story down and write about it? And by the afternoon, you know, I think that whole thing, you know, the guy's agent was getting calls and be like, oh my God, like what is happening, right? The internet is talking about this. And the reason I bring up the story is, first of all, it's kind of silly.

1:58:53It's like a Hollywood movie. Nobody really cares. But it tells you how almost easily, inadvertently, you could shape the discourse, shape the arc of something so easily. And the combination of the algorithms, the metrics, and some of the people inside Twitter meant you could really influence the world. Like, I often call Twitter, you know, a memetic battleground. You know what memes and memetics are, right? So Twitter is a place where all these ideas fight. And if you win, you get to spread all over the world a little bit. And so one of these things that people wound up doing was finding ways to sort of inject their idea into this memetic battleground and try and fight everyone else.

1:59:47Now, this was rewarded by the algorithm because what the algorithm was doing was it was looking at, okay, what is getting the most attention, likes, retweets, right? And these were often things which were provocative, like have made people angry, right? And then those people would get followers, okay? So there was this kind of system which is being created where if you said something provocative and you made people angry, right, like one, you could kind of shape the discourse. Second, the algorithm would be like, oh, this is what people on Twitter want because it's getting more attention. I'm going to send you more followers.

2:00:30I'm going to bump you up in the algorithm, right? And this had two, I think, catastrophic impacts, right? One was that the people who were sort of exhibiting some of the worst behavior, right, like we're getting rewarded, right? Like the more you get people angry, the more, you know, you get people outraged, the more attention you are getting. But the second more subtle thing which was happening was when somebody new signed up on Twitter, right? Imagine you walk into a restaurant for the very first time, okay? you walk into some fancy-smancy Michelin restaurant, and you're like, okay, everyone's here in a suit or dressed up, right?

2:01:10And you're like, I got to look the part. Or let's say you go to a sports bar, right? Like late night, you're watching a game, it's a bit rowdy, you know the vibe. The thing which Twitter was doing was it was shaping the vibe to be one where anger and provocation was rewarded as opposed to kindness, as opposed to education. I'm not saying it doesn't exist. There was a lot of it. And obviously, you and I and others have had great experiences. We're sort of pushing the ball over in one direction, right? And when new people signed up, they were like, oh, I don't know what this place is. Oh, who's doing really well here?

2:01:46Oh, it's that guy who's getting people angry. By the way, this happens on YouTube. This happens on platforms all the time. What do people do? They're like, what are the videos that are doing really well? What are they doing on the thumbnail? What are the titles look like? Let me copy that. So one, you're kind of giving people who are exhibiting more wealth, some literal wealth. Second, you are training everybody new at Twitter. Those are the icons. Those are the people you should follow, right? So there was kind of this whole system which evolved. And so anyway, my point of this sort of roundabout explanation is, one, it taught me the absolute power of social media.

2:02:25Second, that how these systems when centralized can have real power, right? And real censorship. And how it is important that you have absolute transparency, number one. We need to know what algorithms exist in these platforms. We need to know that there is no ideology in these platforms. So it kind of really was an awakening moment for me. It kind of really shaped me. Think of me as somebody on the inside, like, oh my gosh, like, you know, we need to sort of fix these things. This is why every platform, by the way, I'm not picking anybody. Everybody, every platform. The second thing, it kind of really sent me down the path of was decentralization.

2:03:07This idea, which I think is kind of a lot of the crypto people, you know, really resonate with this, like, we as people need to have ownership over how these things work. We need to have a say, We need to know what is happening behind the curtain. And maybe we just don't have like one big thing. We need to have a lot of other small things competing. So it may be very distrustful of top-down centralized control. And it made me feel like, okay, I need to find a way to bring more decentralized systems into the world, which is why I wound up in crypto for a while. But yeah, so social media, I think, shaped my career.

2:03:43It's given me a lot. But I also saw a lot of sides of it. My Twitter time was deeply formative for me. And when I left, I wound up investing, but part of it is like, I need to find ways to battle some of the negative things I saw over there. That makes a lot of sense. You know, that makes a lot of sense. I mean, it's just such a powerful tool. I mean, back to 2020 election timeframe. I mean, we saw what appeared to me from an outsider is we saw a sitting president get banned on X, get banned on Facebook, get banned on Instagram, get banned on YouTube, banned on all of these networks. And so, you know, the way it appeared to, I think, everybody is that these top guys, you know, that on these companies are controlling.

2:04:34And the entire narrative, everything that's going out. And it appeared, and I don't think it appeared, it was that way. Everybody leaned one way. And that was obviously towards the left. And it was that one platform, they just pulled it, AWS. us parlor parlor yeah they just pulled it i'm not i'm not sure why i don't know what's going on but you know it was like holy shit like now there's an entire party oh yes you know with zero voice to include a sitting president and that really i think opened everybody's eyes immediately of of how powerful and and dangerous that this could be the deep platforming was huge i think in a way, I had sort of seen the building.

2:05:28I had seen what was building up. So I was not terribly surprised when all of that happened. But if you think about that era, right? Number one was you had the previous administration putting pressure on the social media platforms to take down content, right? Like they would send these angry emails. This is now all documented and it's come out in depositions and lawsuits and hearings and whatnot. They would send all these angry emails to say, hey, you need to take this tweet down. We don't love this tweet about Hunter Biden. We don't like this tweet about mask mandates or whatever the case may be.

2:06:07So, they would send this down. The second thing I think what was happening was this was the peak of the culture wars. And these platforms where there was this domino effect where if one person took down a video, a piece of content, downranked something, shadow banned, for example, then every other platform immediately felt pressure. because what would happen is all of a sudden you would get an email from the former White House, a member of Congress, or a New York Times hit piece, which would be like, hey, those guys have taken down this thing. You guys are leaving this up, and you are responsible for all these crimes.

2:06:51And so they tried to enforce the overturned window of conversation, and they They try to really shrink what can be said. And I think in some ways, I think the platforms, I think the turning moment in my view, and some people may disagree, is when, well, obviously before this election would be when Elon bought Twitter. Because I think that totally changed how free speech on the internet works. because all of a sudden you had one platform. And by the way, I was there. I was there on the first day. We're kind of jumping ahead a bit, but I had spoken to Elon and he told me he was going to do this.

2:07:38I had left with him at the time for a couple of years and I joined this venture capital firm. But Elon said, look, I'm going to do this thing. Can you come help me out? And for me, it was a little bit like getting a do-over because I could go back and this is not an official capacity. I was not an employee, but like maybe I can go back and, you know, I didn't have the power to do some things in. I didn't have$45 billion to spend on a social media company. By the way, that was a problem I had. But now I have a chance to put some things right, okay? So do you remember when Elon bought Twitter, the let it sink in day when he bought the sink in?

2:08:16So I showed up that same day, kind of snuck in, making sure the TV crews didn't see me. That was such an experience, just watching Elon work and trying to clean up the company, finding all the ways that these progressive levers had been hidden. For me, it was a chance to put right some of those things. I would say that moment changed how free speech on the internet works. Because all of a sudden, you had a major platform, Twitter, then X, which was trying to live by free speech. which brought back the president, right? You know, which brought back a lot of people who had been taken down for just asking simple questions like, hey, what is the origin of the COVID virus?

2:09:06Okay. Are masks effective, right? Or how can you say that we can't sit together in a restaurant, but a protest is okay, right? Like, how does that work, right? Like, but even just the ask a question would just get you a con plan. And now here's a platform that was bringing it back. And look, people have a lot of complicated views on Elon or the company. But I do think that moment was huge in changing how free speech on the internet works. And I think even today that is not appreciated for how important that moment was. I mean, I think it revolutionized everything immediately. And I don't know.

2:09:46Were you on Twitter? Were you active then? You must have been. Yes. I probably wasn't very active. That's probably the, I don't know, that's maybe the platform I used to be least active. I'm a lot more active on it now, but I really pull away from, I go on there mostly to see how our content's doing. Because I do think that there is a lot of toxicity. I know there's a lot of toxicity that comes out of all these platforms. So I try not to fall into that. But I mean, I think it was – well, I don't think. I know it was instrumental in what Elon did because from – like once again, from an outsider looking in, what I saw was he would have capitalized on the entire market and taken the entire market share by providing an actual free speech platform where you don't have to worry about asking questions.

2:10:43You don't have to worry about calling out corruption and you're not going to be censored. And I think that all these other companies, Meta, Google, you know, maybe Amazon, I'm not sure. But, you know, I think that they saw that happening and they thought, oh, shit, if we don't loosen the reins up a little bit here, we're going out of fucking business. And I think they would have gone out of business. And with some of them, I'm actually surprised that they didn't just because of the repercussions of what they did to half the country. Oh, yes. You know, but I think that they saw that gaping hole in their business when Elon secured Twitter and turned it into X and everybody else had to follow suit.

2:11:30What do you think? Am I wrong on that? I think you're very right. I think there's another dynamic, which is courage, which is he just took a lot of the arrows back then. And so when he did that, some of these companies, they just want to stay out of trouble. They would love it if there was nothing political ever said on their platform. They really don't want me the business of navigating where the COVID virus came from. They just want to look, come here, have a great time, have great content. We want to sell some ads against it. Let's go home, okay? But that's not the choice, obviously, when you're running a major platform.

2:12:10It's almost like running a country. So they got sucked into it. And I think, you know, one of the, there's an amazing book you should read from the 50s, a bit hard to get on Amazon. It's called Private Truths, Public Lies. Private Truths, Public Lies. Sounds interesting. Yes. And this idea, and I think it's out of print or something, but it's very simple. idea. It is very similar idea that sometimes in society, everybody starts to pretend to believe in a lie just because it is the convenient expected thing to do. During COVID, right? You would somehow have to wear a mask into a restaurant and be asked to do so, but you could take off the mask once you sat at the table.

2:12:58You're like, okay, hold on a second. How does logically this can't totally make sense. There's something wrong here, right? But you can't really speak out again. So if you did, there were consequences. Okay. So what this book says, the second part of this book, and it's a great book, please, I'm oversimplifying, is what it takes is it takes one person or an entity to point out the emperor has no clothes. Okay? All it takes is one. Right? But when you have that one person do so, at usually, like, sufficient, like, kind of, like, risk to themselves, everyone else starts to fall in line. Makes it okay.

2:13:36Makes it okay. You had two guests on your show who have done this, right? Like, you know, Brian Armstrong with his mission-oriented company statement. Like, did you talk about that when you had him? We did. Right, right. So, that's such a great story. So Brian, please go watch that episode because he talks about, but he basically gets bullied into having to put out a statement, right? Like, which he doesn't want to. And he doesn't like the fact that he's being bullied, right? And he says, listen, we are not a political social company. We are in the business of making, you know, crypto available safely, securely to everybody.

2:14:16That's the job. That's the job we are in, right? So he puts out this blog post, which was hugely controversial because he was accused of being racist. He was accused of frontering in times. And he says, like, OK, we are a mission-oriented company, which sounds crazy that it's controversial. But he says, if you work here, you are signing up as a part of this mission, which is to make crypto available safely, securely at scale. If you don't like this mission, don't work here. No harm, no foul. Okay, go find a job somewhere else, right? If you care about something more than this, that's also fine.

2:14:53We just ask that you do not bring that into the workplace. Because in the workplace, we care about this mission. It's a big mission. We have competition. The stakes are high. The rewards are high. It's going to take a lot of energy, right? And as long as you focus on this, we don't care about anything else. But that was so controversial. Because at the time, I was watching inside these tech companies the rise of DEI. I remember my wife at the time had a little bit of a stint at Meta, and you would have these employees who just totally hijacked meetings. And they would ask people to issue apologies.

2:15:38They would say, we have to comment on literally everything happening in the world. And a lot of Silicon Valley executives were just scared. Maybe they should not have been scared. Scared of their workforce? Yes, they were scared of their own employees. They were absolutely scared of their own employees, right? Like they were terrified of, man, I don't want to seem, you know, Picanist, sexist, racist, whatever it is. Like, I don't want to seem that, right? And they would get attacked all the time. I spoke to some really famous people who were like, I don't want to hire this exec, but I'm being forced to.

2:16:18Because if I don't, I'm going to be accused of discrimination in some shape or form. And then it's going to piss a lot of other people off. They're going to piss some of my investors off or the media off. I don't know what to do. And there were so many people. And I think, look, if you're harsh, we can say they don't have enough courage. But also, look, they had a lot of employees. They were trying to make sure the business doesn't get into trouble. They had customers. And they're like, we don't want to be in the political game, but I'm scared. And this was a thing all over Silicon Valley. It was almost, I would say, a rising infection, which just spread by 2020.

2:16:55A virus. Yes. I remember kind of jumping back a bit, the original moment when I felt the rise of DEI in Silicon Valley. I think this was in 2013, 2014. And there's a company called GitHub. They are a popular developer company. They built source code products for developers. And at the time, they had actually a funny replica of the Oval Office. That's kind of a thing. And they had a carpet. Instead of the presidential seal with the bald eagle, there was a carpet with their logo and it said, in meritocracy, we trust. Okay. And you would think that's safe. Meritocracy, you know, if you work hard, you know, you're the best at what you do, you win.

2:17:43That was hugely controversial at the time. And I remember a lot of us going, that's weird, right? Like, you know, can you imagine like an NFL draft combine where if you say, hey, you run a 4-4, you know, you know, you're pretty good at maybe being a wide receiver. And they're like, well, you can't talk about it. That's just like racist. But that became a real thing. And then I saw over the years quotas, you know, me, a lot of other people, we would get into hiring situations where you'd be told, hey, unless if your entire team, you know, you know, is a male, white, pick your thing, you're being racist, right?

2:18:26Or sexist. and you are like, well, look, I want to hire amazing people, right? I don't care where they are. I don't care what they look like. I just want to hire the best people to the job. But there was a lot of pressure, I know, on Silicon Valley execs to be like, okay, my board has to look a certain way. My exec team has to look a certain way. Otherwise, the New York Times is going to come after you and you might lose your job. And some people didn't lose their jobs, by the way. So this was a bad space to be in. And I think a lot of people outside Silicon Valley don't recognize how bad it got.

2:19:01I was at this venture capital firm and we used to get calls from all these founders, but I'm just so terrified of my own employees. I don't want to deal with this. I'm just building, I just want to build a company. I want to build a product, serve my customers, help my employees, hope we all make some money. I'm not interested in weighing in on the latest social issue, but they were forced. And so the reason I bring this up, so when Brian Armstrong did that, right, like he was taking a huge risk. And I'm not sure he kind of took enough credit for it when he did your podcast, but he was taking a huge risk, right?

2:19:35He could have been fired. They were a public company. He could have been ousted. You know, you had all these ESG firms, you know, you had all these firms, well, kind of stopped doing it now in some ways thanks to who won the recent election. But his idea that, hey, unless you kind of check off all these boxes on the environment, on diversity, we won't invest in you. And if we are on your board or we're in a captive, we are going to put pressure on you to do these things. And so he took a lot of risk and I think it was not easy for him. But when he did that. So, you know, one, I have a lot of respect for him because to do that, he could have just played it safe.

2:20:17He could have been like, yeah, I'm just going to say the things which everybody wants me to say, and I'll be in the tower of time, and I'll be happy, right? But he took on the crowd, right? He took on all these folks who were trying to bully him. So, number one, he deserves some respect for that. Second, when he did that, it sort of opened the over-10 window. People were like, oh, wait, I can do that, right? Like, I can now start, like, saying, oh, wait, maybe we just need the best hire. I don't want to just fill a quota. And Alex Wang, who you had, you know, he did this thing where he said, instead of D-E-I, you know, I want M-E-I, where meritocracy, excellence, and I forget what I stands for.

2:20:59But it's the idea that, you know, we focus on things you kind of want in a workforce, right? You know, you want people who work hard, you know, you want the best of the best, right? Like you want, and you don't care how they look like or where they're from or what they do in their private lives. So, Alex took some heat for that, Alex Wang, when he did that, but that set the tone. So, I have a lot of respect for some of these people. I worked with some of them. We spoke about, you know, folks like Balaji Srinivasan, who I think are similar, who took these arrows when they didn't need to and then opened the overton window for a lot of people to follow them.

2:21:35I mean, I think Silicon Valley listened. Well, Well, maybe they didn't listen, but I mean, it seems like, you know, talking to, you know, talked to a lot of innovators in the tech space recently this year, and a lot of these guys moved down to El Segundo, it seemed like, for a different culture. And so now we see, you know, and I'm not terribly familiar with it. I've never been to Silicon Valley or El Segundo, but it seems like El Segundo is rising up as is like Silicon Valley too. Yes. I think yes for a certain class of companies. Elsevendu, that part of that whole community, they have a lot of ties to the defense establishment, making hardware, which is why you're seeing a lot of these American dynamism companies come out of there.

2:22:28They can pull talent from elsewhere. And I think the other places which I've done started doing really well, Austin, Texas has started doing really, really well. Very true, very true. I always have a soft spot for Raleigh, North Carolina. I think they have a great ecosystem there, which is awesome. Sometimes you have the ecosystem outside Silicon Valley, which is great, because you get these people who don't have a huge ego. They want to work hard, and they're not going to switch your company every year and go join somebody else. But I would say Silicon Valley is still really important just because if you think about a lot of these AI companies, for example, a lot of them are in Silicon Valley.

2:23:11And so one of the things, just a few plugs in, one of the things we've done in this administration to combat a lot of these things about workness, about politics in platforms is we lease this executive order. President Trump signed this executive order three weeks ago, which is called like stopping woke AI. And I played a big role in this along with David Sachs. And the whole idea behind the executive order is that it's not that we want you to pick a left versus right ideology. We just want no ideology in AI, right? Like we don't want employees to put their thumb on the scale and sort of embed their beliefs into such a crucial piece of technology.

2:24:01And this kind of sometimes accidentally happened. Like, for example, last year, there was this sort of infamous incident where if you asked a leading large language model, show me a photo of George Washington, it showed you a black George Washington. And it was sort of an accident. And I think it's not super deliberate. But there are other cases which are way more insidious. and I think given that and given the fact that a lot of these AI companies are in Silicon Valley, which often just have a very left-leaning ideology, I think it is super important to figure out, okay, how do we make sure we don't have a repeat of what I was talking to you about what I saw at Twitter?

2:24:39How do we make sure that? And by the way, with AI, it's going to be so much more important than social platforms. It's going to be so much more important. It's going to be so much harder to find things. So what the executive order says, and a lot of other things we have done says is, number one, no ideology. We want no thumbs on the scale. Number two, we want transparency. We want to know if you say a certain answer, even, for example, a political comment, that's fine. Just tell us where you got it from. Tell us what your sources are. Let the audience, let the viewer, let the person interacting with you, let them make up their own minds.

2:25:15Like, don't hide it, right? Because I think if you go back to my Twitter story, right, that story about this Hollywood movie, if somebody could tell how the algorithm was working, right, they'd be like, oh, wait, something's weird. Let me go investigate. So I think sunlight is the best disinfectant, right? I am a transparency, like maximalist when it comes to technology. So with this woke AI, with other things we have done, we are like, all right, these things are so important for our economy, for the world. And we want to make sure there's, one, no politics, and also we know what's happening behind the scenes.

2:25:53How does an ideology get inserted into an AI platform? Is that from the engineers? I mean, she's processing so much data. I mean, I'm probably way off, but the way I understand it, the way I took it from Alex Wang is that you have these enormous data centers, and that's what the AI model pulls from and processes and gathers the data, processes it, and presents it to you. And so, I mean, I would imagine, I don't know what all goes into that, but I would imagine it be fairly easy for a rogue engineer to insert his own ideology in there and it goes unnoticed by the rest of the companies. Is that how that happens?

2:26:40There are many, many ways how ideology can be inserted. The word ideology is so broad and vague. I'm an engineer. Let's make it very specific. Let me give you a non-AI story and then we come to AI. When I was at Twitter, for a while, there was this phenomenon where left Democratic congressmen, which sometimes are Democratic political figures, would get ranked higher in the algorithm. They got shown more. Like the YouTube algorithm, you know, like the algorithm bumps you up, just like on Twitter, than figures on the right. And the reason that it happened was when you train the Twitter algorithm, you give it examples of good and you give it examples of bad, and you tell the algorithm, hey, go find more things like this.

2:27:28And the people who were sometimes giving these lists, and I'll just give them the benefit of the doubt, right? I don't think they were trying to put their thumb on the scale, but they had certain beliefs. They stuck the news organizations they followed. They stuck the political figures they liked. They maybe didn't know or didn't approve of people on the other side. So the algorithm kind of learned good to mean a certain class of views, a certain class of publication. Okay? And so those guys started getting more traffic, you know, more attention. And so it kind of spiraled. So I bring up that story.

2:28:05So very easily, and sometimes even without any malintent, sometimes that is like malintent, even without malintent, you can slide the systems. Okay? Now with AI, so very simply, you know, just kind of like levels it a bit. the way a modern AI model works is there are two steps. One is a process called training. The way I would think about it is you take all of human knowledge, all of the internet, imagine this massive witch's cauldron, but imagine the cauldron spanned multiple football fields. You dump it all, every book ever written, every movie ever made, all of Wikipedia. right? If you Google all of YouTube, you know, every bit of human knowledge ever gathered, you stick it in there, right?

2:29:00And then you have this algorithm, which I really want to talk about, try and make sense of it, right? You know, so for example, right? You know, it says that the word Sean, what should the word Sean be followed by? Okay, let's say I say, okay, word Sean followed by Ryan. Then you're like, okay, Ryan, and then what comes after that? Well, the Sean Ryan show. Operator, military, intelligence, CIA. Make sense? Okay. Let's say after Sean, I say Michaels. Okay. Then you get title, wrestling, heel, face, stone cold. So it is kind of training and trying to make sense of all of this knowledge. Okay. So that's called training.

2:29:48Then there is a separate step where you kind of take this kind of this large blob and then there is a step called post-training where you are basically trying to make the model better in specific ways. For example, you're trying to make sure it gets better at say coding or you're trying to make sure it gets better at science and you're trying to make sure it gets better at these, you know, in these kind of this very targeted ways, right? Like you might have heard of this place called fine tuning. You're trying to make it better in these very targeted ways. And so that at the end of this, you kind of get this fully cooked model.

2:30:26And then you have inference, which is you go to chat GPT, or you go to Grock, or you go to Gemini on Google, or you go to Cloud, you type in a question, right? And then it starts giving you letter by letter, word by word, that is inference. You're basically giving, getting back answers. So the point, I want to kind of explain this whole thing, just kind of ground a lot of the other conversations we're going to have, but also every step of this could be infected by ideology. Number one, let's go to the step of this big cauldron. What if you only stuck left-leaning content in there?

2:31:08There's the written internet is left-leaning in nature. And so if you just dump the internet in there, you could just get a leftward bias just right there. So just by sort of the sources of data, your model could wind up learning just a certain ideology. Second is when you fine tune these things, like for example, DeepSeq, which we probably wind up talking about, this Chinese model, There's a lot of evidence that it was fine-tuned to add Chinese ideology. Say, nothing happened in Tiananmen Square, or what China thinks of Taiwan. That happened in that part of the process. Another part of the process you could add things are is when the model is trying to react and give you a token and it is thinking.

2:32:04A lot of these models have instructions on how they should think, rules they should follow. You could set up a rule which says, pick the answer which optimizes, if you're me, I would say optimize for truth, or you could say, pick the answer which optimizes for equity, in the DEI. And it's going to slightly slight the model some way. Or pick the answer which is least offensive, rather than pick the answer which is intellectually honest. These are incredibly oversimplified examples, but you could see that. And so any one of these steps could have either deliberate or accidental ideology inserted.

2:32:47Now, in the Google case, and I could be wrong because I only read some press reports later. I don't know what actually happened. I think what had happened is somebody by accident had added this idea that whenever the model is trying to generate photos of human beings, to try and generate people of every race. But then you're like, wait a minute, if you get a photo of a Nazi, they're probably not Asian. That's not what the average Nazi in Germany looked like in 1940s. Or if you have a photo of a Viking warrior, they probably didn't look like me. That's not what Vikings looked like when they shipped out from Norway.

2:33:29But this little thing was like, oh, whenever we ask for a human being, I'm just going to spread the human race out there. So what happens? You get the black Pope, the black George Washington, right? And I hope I'm not making this too complicated, but this idea is that these things can be very subtle. They can be hard to find. They can be deliberate or inadvertent. And for me, you know, in this role, I think, look, one of the things we really care about is making sure that there is searching for truth and that there's no ideology of any kind. And if you look at the executive order, it doesn't say you need to optimize for the right.

2:34:09It says you need to be truth-seeking, optimize for the truth. And if you don't know the truth, express skepticism and tell us what you're reading, right? And my hope is that regardless of whether you agree with me and you on our political beliefs, you will probably agree that truth-seeking is a good thing. So that's the hope.

2:34:34Wow. Yeah. It's a great explanation. Best I've heard. Thank you. Thank you. Thank you. Most people don't realize how much their personal information is being bought and sold every day. Data brokers are making billions, pulling details about you from public records and the internet, and then packaging and selling it, usually without your consent. That's how your information lands in the hands of scammers, spammers, even stalkers. It's why you get endless robocalls and why ads seem to follow you everywhere. That's where Aura comes in. Aura actively removes your data from broker sites and keeps it off.

2:35:08They also instantly alert you if your information shows up in a breach or on the dark web. But Aura goes beyond data protection. With one app, you get a VPN, antivirus, password manager, spam call protection, dark web monitoring, and even up to$5 million in identity theft insurance, all backed by 24-7 U.S.-based fraud support. Other companies might sell just credit monitoring or just a VPN. Aura gives you all of it, together, at the same price competitors charge for just one service. Start your free trial today at aura.com slash remove. Protect yourself now at aura.com slash remove.

2:35:53All right, Sri Ram, we're back from the break. We were talking about Twitter, how you got into there when Elon took over and basically eliminated the ideology of wokeness of Twitter. But something that we didn't talk about yet is A16Z and Treason Horowitz VC firm. And so I'm curious, we've had some offline discussions about investments and things like that. But I'm curious, how did you get picked up for that? How did you find your way into there? So A16Z stands for Andreessen Horowitz. So A followed by 16 letters and Z, A16Z. It's kind of a tech thing. And so they're one of the leading venture capital firms in the world.

2:36:40I would say the leading venture capital firm. I think they're the largest in terms of money they have. And the founders are these two iconic figures in Silicon Valley, Mark Andreessen, who we spoke about, who invented the web browser, and Ben Horowitz, who was with him at Netscape and is a best-selling author among many, many other amazing things. And in a lot of ways, they revolutionized, in my view, the way venture capital worked in Silicon Valley. Just a little bit of history. So when people sometimes think venture capital, they think Shark Tank. You come in, you pitch an idea, some rich person in a suit gives you some money or not.

2:37:24There's some element of that, but it's actually a lot more complicated. The history of venture capital actually goes back to whaling and fishing. So, back a couple of hundred years ago, when you would get these sort of sailors or bailers, I don't know what they're like, hey, we're going to go out and we're going to go into these stormy seas and take a lot of risks and be out for a few weeks, maybe catch some, whales and come back. And it's an incredibly risky investment, right? Half the time, they don't come back. Let's just say the hazard rate was pretty high. And so they would go around and they would go to people with wealth and they'd be like, why don't you stake me on this journey and I will give you a percentage of the whale I carry back.

2:38:19Which is, by the way, the word carry in venture capital comes from that. It was literally sort of, you know, the veil and sort of the thing that you carry back, you would get a percentage of that carry. So even 200 years later, that word has kind of stayed on in sort of the risk capital discourse. And so the history of venture capital is super fascinating. And I think it's an essential part of American innovation and Silicon Valley have existed. So in the 70s or 80s, what would happen is, you know, before venture capital, if you're an entrepreneur, right, how do you have an idea? How do you go get money?

2:38:57You go to a bank, you get a loan, maybe you have some friends and family, you know, they give you some money. And but it was very, very hard to go to a bank and say, I have a really risky enterprise and technology, which you may not understand, but I need, you know, a few million dollars. And by the way, you may never see it again. and you may not have friends and family who can do that. So I'm going to say in the 70s, there was a set of folks who wound up forming Sequoia Capital and Kleiner Perkins who kind of developed the modern venture capital industry. These were folks that made some money from Silicon Valley, sort of the origin of Silicon in Silicon Valley.

2:39:40They had kind of been part of these big chip companies and that's where the word Silicon Valley comes from. They made some money and they basically said, okay, we are going to stake these entrepreneurs, you know, fund them on their journey, not out of sort of the goodness of their heart, because we know that when these companies do really well, we stand to make a huge return on our investment. But the thing which I think is very different from the rest of the world is they were willing to lose their money. They didn't love it, but they're willing to kind of take bets on these risky enterprises.

2:40:12So I think if you look at a lot of these amazing companies in Silicon Valley, Apple, for example, or Google or Netscape, they all had one of these venture capital firms who are essentially taking bets on a totally unknown person and saying, I'm going to give you some money. Famous example is Google. So Google comes by, I think, in 1998. And again, I'm dating myself here, but Google was not the first, the second, the third, maybe the fifth or sixth search engine. Do you remember using like Ask Jeeves, Like Us? Yeah, yeah. Yeah, right? Or Yahoo.com, like Yahoo was another famous one. And back then, you would use the search and people were like, well, look, search is really not a business, right?

2:40:55We don't know how to make money off it, it's been sought. And then these two guys, Larry Page and Sergey Brin, basically come out with a new algorithm, right? They were these Stanford computer science guys. They came with this new algorithm called PageRank. But the page is super interesting. It basically says that I'm going to think you are important if a lot of other people who are trustworthy also think you're important. And that in one oversimplified sentence is kind of how the origin of Google work. But they were like, we don't have any money. They actually went to a lot of other companies and they said, do you want to buy our algorithm for a million dollars?

2:41:30So they went to all these existing search engine companies. Imagine by the way, you just bought like original Google algorithm for like less than a million bucks. and everyone just said no. So they didn't know what to do. And so I think they came to two essential parts of the Silicon Valley ecosystem. First is angel investors and the second is venture capital. So they went to basically kind of these wealthy Silicon Valley guys, this guy named Andy Bechtelstein and they said, listen, I have this idea. We need some money. And Andy, I think kind of gave them$100 ,000. They didn't even have a bank account so they didn't know what to put the money so they kind of kept it and started building Google.

2:42:06But so think of this construct. Two guys, crazy idea, nobody thinks it's going to work. But the ecosystem and the culture of the Valley was like, okay, if we think that you have a shot at this and you've done some work, there are probably enough people in capital who are going to take a bet on you and probably going to assume that 90 % of the time, this money is not going to come back. Every once in a while, you're going to create an iconic company. So Google then, of course, gets funded by another iconic firm called Kleiner Perkins. and then they go IPO and they obviously become the huge giant they are.

2:42:40So, all of Silicon Valley has these stories of these amazing companies, but these venture capitalist firms were at the heart of it. So, this story is very interesting because Andreessen Horowitz totally revolutionized, in my mind, this ecosystem. Back in the day, venture capital firms were sleepy. They were kind of behind the scenes. They would never say anything in public. They didn't want any attention or controversy. The idea was you come in, you pitch us, go on your way. The other thing was strategically what would happen is you would go pitch a person at a venture capital firm, a partner.

2:43:18Usually they had like five to eight partners. And that person would be assigned to you. And that person, if you needed some help, maybe like, hey, man, my company is failing. I need some advice. You had to go to that one person for advice. If you're like, hey, I need a contact in the Pentagon, do you know somebody at the DoD? That person better have a role or an ex, right? But you're essentially the one person. So anyway, so Mark Andreessen and Ben, they had their own experiences with bad venture capital firms and a lot of bad works, right? They do a lot of bad things to founders. They throw CEOs out.

2:43:51And they also were like, listen, why is it that when we go, we have to rely on this one person and we need all this other help? and they met Mike Ovitz. Do you know who Mike Ovitz is? I don't. So Mike Ovitz is the founder of Creative Artists Agency, CAA. I would say along with WME, the two iconic Hollywood talent agencies. He's a guy who's represented like every Hollywood, I think he's retired now, but he's represented every Hollywood celebrity, best-selling author, founder of CAA. And CAA had this amazing strategy to win the market, right? What they did was, number one, they said, we're going to be very loud.

2:44:32We're going to make sure everybody knows our name. We're going to be brash. We're going to use these red binders, which are colored, attention-grabbing. We're going to get people to pay attention to us. The second, and this is very interesting, they said, if you sign up with us, you are an A-list actor. You are with an opposite. The reason you should sign up with us is you're not just getting me, Mike Owitz. You're going to get the entire firm's Dolorix. and this firm is going to maintain Rolodexes just for you. We're going to maintain a list of actors. We're going to maintain a list of producers, directors.

2:45:07So if you come in and you have an idea, you don't need to go find the right editor. You don't need to hope that I have a friend who can direct your movie. We have entire teams whose whole job is to maintain a roster for you at all times. And they would go and win all these clients because Because you would go, what is your image of a Hollywood agent? A guy in a sharp suit, right? Like, you know, on a phone, quick talking, maybe as a Rolodex. CA industrialized it. They had some of that for sure. But they industrialized it because they're like, you come in, we can plug you into this machine, which is going to make great movies for you, make you famous, right?

2:45:42So Michael was Marc Andreessen's mentor and friend. So Marc and Ben heard this and they were like, why don't we do this for venture capital firms? Because let us say a typical entrepreneur, you know, we spoke about some in your 20s, 30s, you often don't know what it means to build a business or you're good at one thing. Let us say you're good at building technology. You're great at computers like I used to be, right? Or maybe you just know your product area very well. You're great at making ice cream. You're great at building rockets. But are you the best CFO? Are you the best marketer? Do you have a Rolodex of government officials?

2:46:21Something goes wrong? You need all these other things who these young CEOs didn't know. Or let us say you run out of trouble. You're running out of money. How do you restructure a deal? How do you close a big customer? How are you supposed to know these things? So what Andreessen Horowitz said, Mark was like, we are going to copy and replicate the CAA model, but for technology venture capital. So we are going to have a roster. We're going to have a roster of every single amazing CFO in Silicon Valley, every single amazing marketer. We're going to maintain a roster of potential board members. So when you come in, you know, 25-year-old with this idea, who has something working, we are going to plug you into the system.

2:47:03You want a CFO? We have everybody on speed dial. We have done them favors, so they will pick up our call. We can get you in. By the way, they can also help you hire CFOs. Often, for young founders, they don't know how to hire executive talent. How do you hire a 50-year-old head of sales if you've never met a great sales leader before? How do you even know how to do that? We have the team who have seen every great sales leader who will help you do that. So that was a different product, right, from classic venture capital. So number one, instead of one person's Rolodex, we're going to give you the whole sort of system.

2:47:35The second thing, they were loud. They were brash, right? They had press articles. You know, Marc Andreessen wrote this famous blog post called Software is Eating the World. which basically said that every business on planet Earth is going to have software underpinning it. And I think in some ways, I think he's been proven right. So there's a backstory. And this is kind of some of the tactics A16Z used to, I would say, become one of the most famous, powerful firms in Silicon Valley. So how do I come in there? So Mark has this great strategy, which he calls harpooning. In the venture capital or tech business, if you sleep, if you are not paying attention to what the next generation is building, time will pass you by.

2:48:24You have to stay current. You have to stay always on the edge of what people are doing. And Mark and there are many others are very good at this, like Peter Thiel, Lonsdale, who I think has been here, amazing people there. Mark was extremely good at it. he would have this tactic he called harpooning. Okay. And what he would do is if he saw anybody online who had written or done something interesting, he would send them an email, right? And getting an email from Marc Andreessen is like, he's a very notable technology figure. So you're just like sitting there like, wow, what is that? And the reason you do that is like, I want to get to know this person before they become famous, before they build this next video.

2:49:05It's a little bit like, you know, you see, I don't know, 15 year old who has amazing skills and you know, I don't know what the legality of this as a coach is. You're like, I'm going to make sure I'm like building a relationship with this person because in some day, if they want to pick a school or a team, I have a relationship with this person. So Mark was very, very good. I'm pretty sure even to this day, he starts sending out these emails in the blue and he harpooned me. You know, I had written a blog post and I would say 2012, 2013. and I get this cold email one day from Marc Andreessen saying, hey, I like this blog post.

2:49:37And I was like, whoa. And this was a very different time in my life, so it was quite shocking. Another blog post. Huh? Blog post. Yes. Blog post to Microsoft. No, I had left Microsoft and this was about… That's what they found you with. Yes. Blog post. Yes. There's a whole pattern here, by the way. If you can take a slight tangent, which is I think one of the superpowers I think people can have with a lot of, with very little effort, is putting content out online, doing what you do. Or I use write, I use to do video, because when you put out content online and if you're passionate and hopefully you know something about it that somebody learns from, some of the best people in the world are paying attention.

2:50:20And I've noticed every sort of world leader in technology is always scouring for new ideas, new people. And so there's been a repeating pattern in my career where I've written something and somebody was at the right moment saw it. And they were like, hey, this guy is doing something which I'm interested in. Let's reach out and make something happen. So one of the things I always tell people do, especially young people, is write things, put things online. These days, I would say, get a YouTube channel. Talk about what you're passionate about and somebody will find you. I know Elon, for example, has found amazing hires because he went on a YouTube rabbit hole and was like, let's get this guy.

2:50:58He looks really smart. You and I talk about like you follow someone on Instagram and next thing you know, you're like, hey, come on my show. And so I think writing what you do is such a great differentiator. Anyway, so I wrote something and Mark liked it. He sent me an email. We met up. We built a relationship for many, many years. That's one. The second thing was I started doing my putting my own money, a little bit of my own money into various companies. I put some into SpaceX, put some into Alex Wang's company, Scale AI, into a bunch of other companies which started doing really well. And Silicon Valley is an ecosystem run on reputation.

2:51:35If you are an investor who puts money into a founder and you are a jerk, you never show up, you never take the next phone call, you're not going to do really well because your reputation will spread. That founder is going to tell his or her roommate. They're going to tell the next person and you will not fare very well. On the other hand, if you take the phone call, if you wind up helping that person when they need you, if you're just not a jerk and you respond to every single thing timely, karma starts accumulating in your favor. And I just built up this portfolio of a bunch of these investments which started to do pretty well.

2:52:16So I made a of a name for myself is what I would say. And I made a little bit of money. COVID happened. I was sitting at home, you know, collecting pasta and toilet paper like everyone else was. You remember that? You remember the whole era? Oh yeah, I still have socks. Oh my God. What was your sort of, I can't believe we lived through that memory of COVID? What's that? What was your, I can't believe we lived through that memory of COVID? what is I camp? Yeah, I can't believe you did that. I can't believe we as a human race or you here, we did that. Like for me, for example, the fact that we spent months stuck indoors, not seeing other human beings, just bizarre.

2:53:01I can't believe we did that, right? And what was that for you? A bunch of stuff. I remember I fell for it. I completely fell for it for about a month. And then I was like, uh this doesn't seem right but i remember spraying packages off at my front door with lysol i remember we so multiple uh hand washing you had to do that you mean wash your hands my hands were all cracking here's a funny story for you so when we when we we moved to tennessee and my wife wanted to start a farm. So we bought, we went and got like six alpacas, bunch of goats, bunch of chickens, and some ducks. And did you have any experience in farming?

2:53:52No, no. Okay. No, we were, we had just started this right before it happened. And so do you know what an alpaca is? Of course. My mentor in Seattle had an alpaca farm. It's like a llama, right? very docile animals anyways it gets hot here you know could get to you know 100 degrees here and so every spring you're supposed to you're supposed to shave the alpacas and and i mean that's what people raise them for anyways the fur right so we call this guy or my wife finds this this alpaca wrangler woman and she comes down and she's she shaves the alpacas i remember at the beginning it was all these people in italy were supposedly dying yes of it and in the february march of that year at that time and they come down and i'm like hey make sure you wear a mask i don't we don't know where these people have been i go down there after this is all going on and there's like four people on this alpaca.

2:54:56My wife doesn't have a mask on. They're shaving this damn thing. And these two people that came with the, whatever you want to call it, the groomer, whatever. The alpaca is getting fancy. These are like free-spirited people. They don't, you know what I mean? They just travel wherever and do this. And two of them were like, yeah, we just got back from this big trip to Italy. And I grabbed my wife and I'm like, what are you doing you're gonna fucking kill us these people just came from italy and she's like holy she got all upset about it and uh anyways anyways about a couple days later you know because i didn't i don't even i don't watch the news i got tired of the news long before people got tired of the news and i was just like we're just being fed the same bullshit over and over well then covid turns up i don't have cable tv at the house i really i don't really watch anything smart and um i just i don't want to be fed that you know what i mean because i figured out you know i mean i think everybody's figured it out that they they're telling you how to think what to think they're injecting thoughts into your head by that and it can manipulate the way you think so we got rid of cable a long time ago covid pops up and i was like hey let's just see what's on air tv only station we got was abc you know and so i was like well let's let's just see what's going on in the world so we were fed you know that garbage for a while and then and then i started talking to some friends that still have news and and um and that's that's when i've you know figured it out i was like all right this is this is uh this isn't about a virus this is about something much bigger and so but that was kind of those were my moments oh man yeah lasted about a month isn't it crazy that we all lived through that uh we just had our first i gotta be honest it's fucking embarrassing but i mean we didn't know better yeah we were all told this and we had nobody had gone through this before and it was so much fear and you're like is it spreading in the air not spreading in the air like the number of feet uh everyone and we just had our first child like a little bit before that.

2:57:13And we spent so many months just not seeing any human beings. Yeah. I was so bad in so many ways and so bad for sort of elderly people I know who are just stuck and just so, it's both funny, but also I can't believe we all did that. But anyway, so I was sitting at home, you know, doing this and not seeing other human beings. And everyone was on Zoom. If you remember, this was the era where everybody was doing video Zoom meetings. That was the thing. and Mark Andreessen, he reaches out and he said, what are you up to? And I'd left Twitter because I'd gotten tired and I just wanted to do something else.

2:57:49And I was like, well, I'm sitting at home, waiting for this pandemic to be over. I'm sure it'll be over in a few weeks. I'm amazing at predictions, Sean. And he said, well, just come help us out. And what I did not know was, they'd been talking about me for a while and they had somebody else who I think was going to step aside. And so I became part of the team. I became one of the general partners, along with Catherine Boyle. She joined after me, but there's about, I would say, maybe 20 general partners for the firm. I became one of them and became a VC, started investing while also doing my podcast.

2:58:29But that's how my Andreessen Horowitz journey started. Wow. I learned a lot, by the way, Andreessen Horowitz about investing. I think I learned a lot about how to be a good investor there, which in a lot of ways, I think, can carry over in other things in life. What is it that you see in a startup company that makes you want to invest? I mean, what are some of the points that you look for? Good question. At the heart of it, that is the job. You are here to figure out who the winners are and put as much money as you can inside them. So the first thing is you probably do not know what to look for until you have met hundreds of companies and founders.

2:59:20So for example, right, like, you know, if I were to meet somebody from your world, from your background, right, I don't, without meeting a lot of people, I don't think I would know the difference between a amazing top tier operator versus somebody who's not. just because I'm just not from that universe. And I suspect that, you know, if you came into my world, you may not know the difference between a top-tier engineer, somebody who's maybe just good and not great. And the only thing that kind of sets that apart is how you put in the time and the effort to meet everybody. So the first thing, if you want to be an investor, is you just got to talk to everybody.

3:00:00You got to know, okay, who the great founders look like, who the great engineers look like, who are the great builders, marketers, who are they maybe? I need to meet everybody. When you meet people over time, you build a spidey sense. I'm sure when somebody reaches out to you, just because of this podcast, you now have a little bit more of a spidey sense in terms of how to judge them. Are they legit? Are they full of shit? Or somewhere in between. Okay. And the first thing is, unless as an investor, you've done your homework and met a lot of people, you will not know the difference between the next Google founder or these guys are just, they have nothing.

3:00:40You should have done the homework. That's number one. The belief I learned is that technology is a sector where often the winners are outsized. Peter Thiel talks about this. Peter Thiel has this book called Zero to One, where he basically says that if you go to Palo Alto, in the Bay Area, there are probably 20 Indian restaurants, 20 Italian restaurants. If you invest in one, there is no way that Italian restaurant is going to become the only Italian restaurant in the United States. Just not possible. At best, they maybe have a chain, you go to a few cities, but that's it. There is a cap on how big those Italian restaurants can be.

3:01:26No offense to Italian restaurants, but that's just the nature of the business. Technology businesses are different. If you invest in the right company, they may be the only search engine people use. They may be the only social media network people use or pick your company. So there is a huge difference in picking the winner in a category versus not picking the winner in a category. For example, in 2005, Google dominated the world of search engines. Who was the number two search engine to Google? I don't know. Exactly. Nobody does. It doesn't matter because Google just dominated. Same with Facebook.

3:02:05right? And so there are these winner-take-all patterns, which often wind up happening in technology where, you know, have you seen the movie Glengarry Gendross? No. Oh, okay. This is a classic movie. And there's this classic scene where there's a bunch of sales guys and they're kind of running low on meeting their quotas. And Alec Baldwin comes in and he's just sort of this amazingly famous salesperson. He gives them a pep talk, right? And he basically insults them, insults their manlyhood. And he says, you know, the guy who gets, you know, the most sales, the winner gets this amazing car. You know what second prize is after this fancy car?

3:02:45Set of steak knives. Nothing, right? So often in the technology investing world, you know, there's a great scene. You should check it out on YouTube. It's very similar dynamic where if you invest in the winner, right? Like you're going to be in Google or, you know, Microsoft or Apple or pick an amazing company. or you're in a company where you're like, oh, I don't even know who the second guy is, right? So how do you then figure out what the next goal is going to be? Well, number one, you have to really, really, really do your homework. And I think what the firm taught me is that you can get the category wrong, of company wrong, but you can't get the company, you can't get the actual winner wrong.

3:03:23What does it mean, right? So for example, about like, like VR is a good example. Like Oculus was a big, it was a reasonable winner in VR, but a lot of the VR companies didn't really do super well. They might come back now, but people invested money in VR. And what some of the partners would say, like, that's fine. You took a bit on the entire category. You had the best company in the category. That's good, right? But what is not good is if you're investing in search engines and you did not invest in Google because that is a difference between, you know, being part of one iconic company and not being part of anything at all.

3:04:00So we often thought a lot about how do we make sure that you're investing in the winner in a category versus somebody else? And you have to wait sometimes until you know who the winner is, right? You have to be prepared. You have to know all the founders. They need to know you. So that was, I think, another big dynamic that they taught me. The final and I think the most important part. I have a question real quick. You know, when you're talking about trying to find the winner in a category, I mean, would it not be wise to invest in several companies within the same category? Great question. Let me ask you a simple question.

3:04:42You have Palmer here, right? Let's say you invested in Palmer. Let's say you're also invested in Palmer's competitor. How do you think he's going to feel? That's the caveat to that. Right? And so we had a word called getting conflicted. I think the great entrepreneurs don't want you in bed with the competition. Now, of course, there are a lot of ways to kind of sometimes people get around it. Some people that I work in other places, they say, you know what, we work with everyone equally. but you know one of the prizes is you know if you just back the winners and they don't want you often to work with everyone else right uh so that is a definite dynamic but to be honest there's a ways around it where some people will say guess what i that's just the nature of me i just work with everybody that's a price you have to pay to work with me that's all fine but that was just the culture i grew up in when i was at anderson harvard which is you pick a person and that's the only person you work with, right?

3:05:43And I believe a lot in that because I do think a lot of founders value loyalty. Like they want you to work with them because they are in a knife fight every single day, not metaphorical, not a little one, with their, they're trying to win deals. They are trying to make sure that company's not like crushed or running out of money due to the other person. They don't want you also helping the other person. They want to know that you are the, you know, you are loyal to them. And I think, I really believe in that loyalty. I think that matters a lot. And I also think, you know, people are not loyal, you know, the great founders don't wind up working with them.

3:06:15So that's a big dynamic. But it's a good question. But that's the answer. I think the most important part I would say is you need to have a spidey sense for what a great entrepreneur looks like. And not look, like literally look, but how they operate, what they do. And I was lucky here because I got to spend time with lots of many amazing entrepreneurs from the sometimes working with them. or sometimes from the outside. And you sometimes see similar patterns across great entrepreneurs. Like, for example, I'll just pick one. Every single great entrepreneur is insanely fast. You know, they are urgent.

3:06:55I used to remember, you know, when I've been in teams which are not great, if you wanted to have the next meeting or a conversation about something, they'd be like, yeah, let's go meet in a week or two. Maybe I put a PowerPoint deck together. you know, who has been in corporate America probably recognizes this. If you work for a great entrepreneur, they'd be like, well, let's talk in five minutes. Let's find the answer. Let's go, go, go. We don't have time to waste. They work 24-7. They eat, sleep, and breathe this. There is a real sense of urgency of mission. And I think that's one of the things that I've learned to pick up on over time.

3:07:32And it's not the only thing. I think it's table stakes. that it's necessary but not sufficient, as they would say. But you build a spidey sense over time. Interesting. Interesting. How long were you there? Four and a half years until this job. What are some of your best investments? Well, I'll pick one because it ties to the theme of what we talked about in social media. I just became convinced that centralized social media platforms are not good for all of us. Because if you have the team which doesn't agree with your politics running them, they can just issue orders top on down. And that's how I really got into crypto.

3:08:10I became a fan of this idea that crypto is a way to decentralize these platforms. And instead of having one central company or algorithm which does everything, you can have people have a say in this. So personally, I'd invested in companies like SpaceX with Elon or Scala with Alex Bank. From the firm, one of my companies I'm really proud of is a company called Farcaster. And Farcaster is a decentralized social network. And I did this a few years ago. And it is done by this ex-Coinbase product builder named Dan Romero, who's awesome. But the idea was that imagine if Twitter was somebody, let's say you like Elon, you agree with this policy.

3:09:02Let's say someday in the future, Twitter is run by somebody who doesn't agree with you, right? Or say it's like, I just hate Sean Ryan. I want to see his account disappear. Maybe YouTube does that to you. In Farkaster, they built a cryptographic protocol, sorry, protocol on top of the blockchain where you own your social graph and anybody can build a client on top of Farkaster. So what does it mean? Like right now, if Twitter decides to ban you, YouTube decides to ban you, You're done. You're toast. In Farcaster, you're like, you know what? You can't ban me. I'm just going to go over using the same client as this other person.

3:09:39I'm going to take my follow graph. I'm going to take my followers. I'm going to take my content. I'm going to go elsewhere, right? By the way, the interesting thing is this is how the internet used to work. Let me ask you something. What was your first email? Don't tell me your first email. What provider did you sign up for your first ever email address? Hotmail. Great. Okay. And then you probably switched, right? You probably went to Gmail, et cetera. But when you switched, there are a lot of ways to take your email to other places. You could forward your email to other places. If you wanted to access your email, you could do it on your phone.

3:10:13You could do it using the official hotmail.com website. Or later on, you could use it on the iPhone or on your own desktop client. So you signing up with the service was not tied to the actual application you were using with it. And you're not tied to it. Sometimes you can take your email address elsewhere. and with social networks, people lost that. If you want to use Instagram, if you want to use TikTok, you have to use the official app. There's a lot of reasons as to why advertising is a business model. But what Farcaster is trying to do and I think others are trying to do in crypto is let's bring that original internet vibe back where you own your handle.

3:10:51So for example, I want a world where even if Neil Mohan, who runs YouTube, gets really pissed off at Sean Ryan, you can take your audience and your subscribers and just go elsewhere. And they can just follow you elsewhere, right? That's a possibility now. It is a possibility. Now, it's early days. And the reason, you know, so you have to figure out how to make it happen, right? Technically, economically. So they built this kind of crypto protocol which to make it happen. And they built all these alternative clients. It's very early days, right? But I love the idea. Because for me, that is the spirit of the internet I grew up in.

3:11:27like late night, because imagine in a world where I was late night in my computer back in Chennai like 21 years ago, and I was told, oh wait, before you can write any bit of code, you need to call up a salesperson in Microsoft before you can write code. Like I would never had anything. And so I think this harkens back to an original ethos and spirit of the internet. And I think of crypto, which is that you own these things. You have a say, you have a stake. Let me ask you another example, right? How many subscribers do you have on YouTube right now? Almost 5 million. Great. How much money do you think YouTube makes per year?

3:12:04How much money do I think YouTube makes a year? In ad revenue for Google. Man, I have. Let's make up a number. Let's say it's$5 billion, right? Again, I don't mean to pick on YouTube. I think they're amazing, right? How much of that money do you think you are owed? Or have you even thought of that, right? And, you know, because you are a stakeholder. You're contributing to this platform. If you use another social media platform, you are contributing to this platform. And I think one of the promises of crypto is that, well, let's give everybody, it could be somebody with 5 million subscribers, it could be even 5 subscribers, a stake in the platform, a stake in two ways.

3:12:43One, financially. If YouTube, Instagram, TikTok make money, you get money. Second is stake in terms of how you want to experience it. If you want to use, instead of YouTube.com, you want to use another app, you should be able to go for it. If you want to use a different algorithm on the right side, or if you want to use a TikTok with a different algorithm, you should be able to go for it. I want a world, for example, where one day you open up Twitter or TikTok and you're like, well, you know what? I want to pick this algorithm. I don't want to pick just the algorithm they give me. Imagine you have a shopping market of algorithms you could pick from.

3:13:16Crypto makes all of this possible. So anyway, that is one of the things I was a very deep believer in. Look, I was lucky to work with some amazing founders and entrepreneurs, some of the best, deepest relationships. These are friendships I'll have forever because they took a bet on me as much as I took a bet on them. And I'm grateful for all of them. Wow. Very. What's that? Forecaster? Forecaster. That's interesting. That's very interesting. I've not heard of that. Yeah. Well, it's early days, but, you know, it's one of those things where I think, you know, them or somebody like them, I think it's one thing that just needs to exist.

3:13:52Yeah. Yeah, that sounds genius. Every day I go to bed, I'm like, oh, this could all end. Well, you should be like, Neil Moore is a nice guy. You should be like, you know what? I put in some work. I deserve a little bit part of this. I have a say in the algorithm. And right now, I'm sure they listen to you. You can get a call and they'll probably listen to you, but you don't have an actual say. And I think that's the promise of crypto. Yeah. Wow, that's genius. I love that. So let's move into AI. How did you get picked up for the position? Man. Senior White House advisor on all things AI. It is a pure, I would say, accident in a lot of ways.

3:14:34So going back a bit, I had a lot of people in D.C. have long careers in public service. They have a lot of aspirations to be here. I was not one of them. You know, I was happy as can be back in Silicon Valley, back in the technology world, investing. I was thinking of starting a company. I was thinking of starting my own firm. I was just off doing, you know, my thing. Just because I'm sure a lot of people will agree, DC just felt like this other universe. Like, well, a lot of crazy things happening here, but I'm here doing my thing. but what wound up happening is about a year and a half ago like everybody had gotten really involved in AI I was investing in AI but there was this narrative that picked up about AI just killing us all I don't know how much you paid attention to the whole like doomer arguments on AI maybe a year and a half or two ago but there was this whole school of thought which picked up I'm going to say in late 2023 sometime under that frame which basically said that, oh, we should just stop working on AI because this is going to become this superhuman intelligence which just takes over all of humanity.

3:15:55And I disagree with that, which we can really get into. The other thing that they started doing was a lot of these folks started influencing government, influencing the Biden administration, influencing various legislators, including California. and what they said was, let's find a way to stop or slow down AI in any number of ways. And I was paying attention to this, but the one thing which really kind of struck me as very wrong was they tried to ban this thing called open source. Okay, so open source, by the way, just for history, is the software world through the last 30, 40 years has had two camps.

3:16:32One is closed source, which is somebody like Microsoft Windows, They build it in Redmond. They ship you a product or Apple's OSX operating system. You use it. The other is open source. The classic examples would be Linux, which I'm sure you've heard of, or the Android operating system, where there is source code, which any of us can go look at, modify, and then contribute back to. And open source is very important for a couple of reasons. One is it is kind of part of the spirit of the internet. It is how people kind of innovate, build on. You get kids, you get academics, you can download, you know, the latest and greatest and you can build on it.

3:17:09It's kind of this spirit of how all these engineering ecosystems at the heart of Silicon Valley. The second is open source is safer. There is a great law called Linus' Law, named after Linus Tarwells, the founder of Linux. It says, with given enough eyes, all bugs are shallow. It's a different way of saying sunlight is the best disinfectant. And it is like, if you have the world looking at your code, you can't hide a security vulnerability in there. We're going to find it, right? Because one smart person might miss it, but a thousand smart people looking at it, somebody's going to find it. So what has happened over the last 20 years is if you look at the heart of the things which power your phone, things which power security software, a lot of it is open source, right?

3:17:53Because people like I have trusted it, a lot of it's engineers working on it. It was very important. in AI, there was a growing effort to build open source, open weight models. And again, bear with me because some of these, I think ShadGPT is probably the first time people heard of these model AI models.

3:18:16There's been a long history of AI development. AI started in the, I'm going to say in the 40s. 40s? Yes. AI is one of the reasons computation was invented. There's this guy, Alan Turing. You know, you might have seen in the imitation game with Benedict Cumberbatch, he invented the Enigma machine or as a big part of it, which helped the British with the Nazis in terms of ciphers. He was a mathematical genius in the 40s and 50s. And he invented a lot of modern computation. He invented two really important ideas which underpin all of computing. One is called the Turing machine, which basically says that anything can be a computer if it can decide between option A, option B, or it can follow an instruction.

3:19:04It is the heart of every computer. But the second, which is even more interesting, there's something called the Turing test. Have you heard of this? No. The Turing test is, okay, I'm sitting in front of you. Imagine there's a door in front of us. I couldn't see you. There's another door. There's one of, behind one door is a human being. Behind another door is an AI, right? The Turing test is, can I, as a human being, tell the difference and know who's the human and who's the AI, right? And it's kind of, it was always seen as sort of this theoretical hypothetical test, right? But the thing about AI development, it started in the 40s and 50s, and it has always been, I'm going to call it the holy grail.

3:19:43It has inspired people. They came into the industry. For example, in the 60s, there was this guy, John McCarthy. He invented these amazing programming languages called Lisp, all because he wanted to build AI. And in every decade, there were people trying to figure out AI. And the 80s and 90s, people started really interested in the idea of neural networks. This idea was like, let's figure out how the brain works. and then let's try and mimic it in a computer and maybe we get AI. For some different definition of AI, maybe you think Terminator and Skynet, maybe you think 2001 A Space Odyssey, sorry Dave, I can't do that.

3:20:24You think data from Star Trek, whatever it is, that's some form of AI. So basically people have been trying for years. Now the challenge the AI has been, over the last 50 years, you would see these ups and downs. Somebody would get really excited about a particular idea, right, it would show promise for a while. People would build PhDs, they would build companies, and then one day it'll run out of steam. Because what would happen is this piece of AI that worked for one idea would not work for another idea. Maybe you can detect cats, but not dogs. Maybe you can translate English, but not French, right?

3:20:58It doesn't, it wouldn't scale. And so, all these ideas would happen, they have this little kind of hill and, you know, kind of this momentum and energy and then disillusionment. And people would leave the industry, companies would go out of business. And this was happening time and time and time again. Even in the 2000s, neural networks, which were super interesting and hard, every academic was interested in the 80s. In the 2000s, people were like, I don't know about neural networks. We've been stuck for 20 years. We haven't made a breakthrough. Instead of that, let's figure out alternative mechanisms.

3:21:30Like, there are support vectors, the other things that people are doing. Now, there are two really key moments that happen in AI, which one of the questions I think people should ask is, why is AI interesting now? Why not in 2010? Why was ChatGPD not built in 2005? Why is it being built now? I think there's a long history of technical accomplishments happening and two very important moments. One was there was something called AlexNet, which was helped build by this guy, Ilya Satzke, one of the founders of OpenAI in 2012. But the most important thing I would say, and this thing should, I think, someday win a Nobel Prize or something, is this paper that came out of Google in 2017.

3:22:14The paper is called, Attention is All You Need. I think this is going to be a historic paper. I think this is going to be as important as Einstein's theory of general relativity. It is iconic. The reason why it is important is that, For the first time, we found a mechanism that just continues to scale and work with neural networks. Remember what I said? Until then, there were all these stop and start-stop attempts. You start somewhere, you throw some promise, you would stop. With Transformers and Attention, these bunch of Google engineers figured out this thing, and they didn't know what they had at first.

3:22:49But it turns out it just kept growing. And it had this magical property called the scaling loss. And what it said was that if you give this more data and more GPUs, more computers, it just kept getting better across the board. And it won't stop. So far, it has not stopped. And the reason why is this important? Every AI algorithm in the past had stopped. It worked for a while and then it did not scale. People tried to be smart. They were like, can I detect the human face in a particular way? Well, yes, but then people have different faces. Well, you can detect the face, you can detect the bicep.

3:23:25And it just kept failing. But this algorithm, as long as you give it more data, more to learn from, and then more computers, more data centers, more energy, it just kept getting better. And so it was built by Google, but OpenAI, which had been started by Sam Altman and Elon Musk and a bunch of others, they kind of ran with it. And a few years later, came out with ChatGPT, which I think is probably the real moment where people are like, oh, wow, this is really powerful and so on. So just a lot of history in terms of how we got here. Why are we even here at this moment? Now, with ChatGPT, it's a closed model.

3:24:10When you use ChatGPT, Brock, Anthropic, Google, what does a closed model mean? You type in a question or maybe you give it an image, you give it a video, and it then generates an answer for you. But you can't really see what it is doing behind the scenes. You can't run it on your laptop or you can't run it on your own data center. It is closed, not open. But that's perfectly fine because they have a business model. They spend hundreds of millions and billions of dollars on this. They want you to pay a subscription fee and use ChatCPT, right? But a set of companies started building open source models.

3:24:46And these are models where you could take chat GPT, a smaller version of it, but run it on your laptop. You could run it on your phone. Meta was one. They had this model called Lama. Why was this interesting to how I got in here? Why this whole roundabout thing? A set of people got really convinced that open source was dangerous. They were convinced themselves that it was going to help the Chinese, that somehow it is going to make the world unsafe. And they tried to get California as a state to basically ban open source. So I was sitting here, you know, minding my own business. And I was like, man, this is just wrong.

3:25:23Right? Like, because this is the way the internet should work. This has been the heart of innovation. This is how you get multiple small entrepreneurs and not just a few big guys. Not that I have anything against big guys. They're awesome. But I need multiple entrepreneurs. This is just wrong. So me, as somebody who had no interest in policy, I started getting involved in these battles. Okay, so I started joining the right groups. I started putting my hand up. And I was in the United Kingdom at the time. I was helping Andreessen Horowitz grow internationally. I had a meeting with the then UK government.

3:25:53They had a bunch of people. And they asked me, hey, they asked the whole group, can we make this open source model public? This was two and a half years ago. It was super safe, obviously. I said, I was the only person in the room who said yes. And when I said yes, this person next to me looked at me and said, you have just killed all of our children. I was like, whoa, that's a bit much. But I remember thinking, wow, these people have infiltrated the highest reaches of government. And they've sort of scared the world into thinking that this AI is going to take over the world and just kind of take over humanity for reasons by which I can sort of dispute and why I think is untrue.

3:26:38But that kind of got me personally motivated. So fast forward, the election happens. And I was very, you know, I was close to David Sachs, the AI czar. And I said, listen, I have all these ideas for you, because I think this is one of the most existential questions. The Biden administration has taken so many wrong turns. They have hurt the American AI ecosystem. They have caused us to almost lose the race to China in a bunch of different ways. And I think there's an existential issue. And David tells me, well, come to the White House and help fix it. And I was like, whoa, like, I didn't know that was an option.

3:27:17And for me, this country has just given me so much. And I was like, here's a moment in time where I have the chance to give something back. And I've been incredibly fortunate where, like, imagine you have some skill set in some area, right? And all of a sudden, you get a chance to help your country with that particular skill set. I was like, I don't know when this chance will ever come again. I was convinced the country was going down the wrong direction on AI. I thought the stakes were incredibly high. If we get this wrong, which I thought the Biden folks were, we would lose this race to China with catastrophic consequences.

3:27:58And here I was with this opportunity to, well, step up and try and do something about it. So I flew to Mar-a-Lago and I got a call saying, hey, you know what? You're on the team. This was, I'm going to say mid, early December, a little bit after the election. Fast forward a bit more, the president gets sworn in. And a couple of days later, and I suspect this was time, China comes out with this model called DeepSeek. Have you heard of it? Oh, yeah. Oh, yeah. So DeepSeek is super important because it is a open source model. Okay. So first of all, a lot of the people who wanted to stop open source said, well, one of the reasons we want to have open source is because we don't want to help China.

3:28:41It turns out that the Chinese are actually way ahead and they actually had a genuinely a fantastic model that surprised the world. Okay. At the time, it was the only reasoning model, a model which can think and reflect on itself, which was not open AI. It was ahead of so many other models that America had. It captured everyone's attention. And I don't want to take any credit away from the team that built DeepSeek. There was this team of basically hedge fund guys who were very, very good with building on top of GPUs. And it turns out that a lot of the skills that you need to build great models is programming GPUs very well.

3:29:23So they built some innovative, cool stuff. So I always tell people like DeepSeek has some great ideas that we hadn't seen before, but it is, I think, a Sputnik moment. Wow. Because it showed us that not only are we not like - Scared everybody. Yes. And because not only are we not far ahead, we are super close. And we are on this wrong trajectory where we could just wind up losing. So we talked about all these companies, Google, Apple, et cetera. imagine if in 1998, China built Google, and that's all we use every single day. China built the iPhone. That's all we use every single day. And AI could be a much, much more important technology platform than those things.

3:30:06And we were off to the races. I remember the very first day coming in. I hadn't even been sworn in yet, so they had to give me a badge and do all these things, briefing everybody. And then the president comes out that evening, and he says, like, we need to compete. We need to unleash American entrepreneurship. So that was my, I think the day before my first day, the next day I started. And we were off to the races, man. Well, congratulations. Thank you. Late congratulations. But, you know, you had mentioned, I want to talk about, you know, you were talking about the doomers in AI. What is, I mean, I know a lot of the concerns, but I want to hear them.

3:30:44You know, what do you think the concerns are? What is it? So let me sort of try and be intellectually honest and steel man what some of the concerns are on AI are. There are, I think, several classes of concerns. The first, maybe the most important one, and this is not from the doomers, is AI is going to take my job. That's important. But that's not what the doomers are talking about. There is another set of concerns, which is AI could build maybe, you know, a new kind of biological weapon. a new kind of nerve toxin. Those are very legitimate, serious threats in there. But the real Doomer argument was this idea that as AI keeps improving, that at some time, AI models will start improving themselves.

3:31:33Instead of a human being, being like, all right, I'm going to control this AI. I'm going to try and make it better every single piece of time. at some point in time, a model will start to improve itself. And there's this word in sort of this AI debate, which is called takeoff or foom, which is kind of the sound of a rocket taking off. If you hit that moment of improvement, instead of AI just becoming better, better, better, better, it just goes foom. And so their belief is if that happens, what are the results? Well, you might get AI, they would say, that is not aligned with our hopes and beliefs.

3:32:17Not because AI is evil. Like, if you see an ant, you're not aligned with its interests. We just, not like we are particularly against ants, but we just don't care as much. And they worry that AIs might think of us as ants. Maybe there's this famous thing called paperclip maximizing. Have you heard of this? No. Oh, so paperclip maximizing is the idea that the AIs may not want to kill us, but they may not really know what we like. So they may put us in a job where they say, you know what, make just amazing paperclips, because they think humans are happy. And we're like, no, no, that's not what is the emotion satisfying job.

3:32:52So it is kind of used as a way to say AI might become this all-powerful, all-knowing intelligence, which then is going to be smarter than any one human or any one country. And then just given its knowledge and power, could just control us and may not have our best interests at heart. I think I've done a reasonable job of conveying this scenario. And so if you believe that, and they had some other concerns, especially the Biden people, they believe that, well, if you believe this, you need to make sure that we slow down, we don't get anywhere close, and we need to make sure that only America can build these AI models, no other country.

3:33:37Because why would you risk some other country having this superhuman intelligence? They would often compare it to a nuclear weapon, and they would often compare a GPU, GPU, a graphics card, one of these thousands of GPUs which are in a data center or hundreds of thousands to plutonium. They would say it's like collecting plutonium and you don't want to have another country, even an allied country, having a nuclear weapon before you. So they had this thought, I would say fear of we need to make sure that if we hit AGI, I'm sure you've heard of the word AGI, artificial, like it needs to be us first and we need to kind of get off it.

3:34:14They also, and I think, so this was, I would say, some of the school of thought around the doomers. And I just didn't buy any of it. And the reason I didn't buy any of it is that we are so many years into these AI models. And there are absolutely few things. There are absolutely no signs of takeoff. In fact, we're recording this in September 2025. I would say for the last several months, every model has slightly leapfrog over another, and there is no sign that any one model is taking off. There is no one model, they call it. In fact, what is happening is these models are increasingly specializing.

3:34:56You have one company which is building amazing models for code. You have another company which is building amazing models for maybe friendship or companionship. and other models which are great for scientific discovery and thinking. So, instead of having this one model, which is becoming the superhuman intelligence and surging ahead, you're having this cluster of models, you know, all sort of giving people great benefit, but not showing any signs of takeoffs. That's number one. The second reason why I disagree with a lot of the doomers is that it fundamentally underestimates human ingenuity. Okay.

3:35:38Human beings through history have been able to harness technology, right? The wheel, fire, right? Like electricity. Have you seen these videos of, you know, when people would try and scare people over electricity, but we know by killing elephants, like, you know, there's this whole thing where there was fear mongering about certain, like one form of electricity versus another. So they would do these incredibly barbaric things. They were like, I'm going to electrocute this elephant, this is why electricity is not safe for you. There was a lot of fear mongering. Time and time again, human beings found a way to harness technology.

3:36:16Even with nuclear weapons, human beings found a way to harness nuclear energy. Of course, there's a lot of do-mortar thought against that, which we can get to. Same with the internet. We have a lot of downside. We found a way to harness it. I think if you think about AI, it fundamentally underestimates human ingenuity. Humans are not going to allow a one all-powerful model to become superhuman intelligent, you know what, without being like, you know what, we're going to have a say in this. We're going to try and stop it way, way before that happens. They probably would have a bunch of other AI models which stop it.

3:36:50So I think just fundamentally underestimates humans, all of us, you know, our human spirit, our creativity, our ingenuity as individuals and assurances. That's number two. The third piece is that, you know, instead of having, you know, what I think AI has become is we, and I want to come back to this later into the jobs question, we absolutely need humans at both ends of the AI model. We need a human being to give it context and input. Like I was telling you before I showed up today, I went and asked a model, hey, I'm going on Sean Ryan's show. I am the White House AI advisor. What should I talk about?

3:37:27It's pretty okay, but it didn't know me really well. It didn't know you pretty well. But if I'd given a lot more input and I'd worked with it, it would have done a lot better. The second thing which we absolutely need humans for is on verifying the output, which is when an AI model gives you an answer, be it a diagnosis to a doctor, be it a suggestion to an accountant, or maybe a suggestion to somebody manning a drone, you will absolutely need a human being to check it to verify it, to specialize in it. So when I think of AI, I think of like the Iron Man suit. It amplifies you. It is a fantastic assistant.

3:38:06It does not replace you. So, but anyway, so all of this, I think, I think disproves the doomers and their version of this takeoff risk. But along with this, I think the Biden folks made a bunch of key errors. They felt that one, we have to stop this AGI from happening anywhere else in the world. Second, they were convinced that there was going to be a shortage of AI chips and GPUs forever, and that China just can't innovate. They just can't build amazing models or amazing chips. They were just wrong on all of it. And if you think about today, people can just get AI GPUs from NVIDIA, from AMD. There's a bunch of other companies, you can just get them.

3:38:55There's no more shortages, no more supply constraints because the semiconductor industry is very, very good at reacting to shortages and honestly just finding a way to make money. But the second thing is they really underestimated China because since they really believe these AI models can be constructed only by a certain number of people in San Francisco, they did not think that some smart set of people around the world could build a DeepSig and it totally shocked them. And nobody predicted DeepSig. So, as a result of all this, I just think the whole Doomer narrative, set the country on the wrong path, I think almost really hurt us in the race against China.

3:39:34And a lot of what we have done in this administration is try and undo that. Interesting. I mean, a couple of questions going all the way back to different AI-specific models. You had mentioned one for friendship and companionship. What is that? Well, I mean, I would say it's more of a use case. But I think if you look at Grok, there are all these sort of characters with certain personalities people use, sometimes not safe for work. But I think when GPT-5 came out recently, a lot of people were upset because they felt like they had built a friend in the previous model with GPT-4. It had a certain tone.

3:40:17It had a certain way of speaking to you. And I do think, you know, and they kind of projected this idea of a relationship. And sometimes some of these AI companies, what they're doing is they're specializing in tone. Are we the friendly model? Right? Like, are we going to be just very clinical and cold in how we respond? You know, so that is one dimension in which you can differentiate. But a lot of other differences which you can differentiate are, for example, capabilities. Coding is by far one of the most lucrative, most interesting capabilities in models right now. Have you heard of the phrase vibe coding?

3:40:53No. Oh, okay. So when I wrote code and everyone wrote code, the way you did it is you typed in a piece of code, a program, you gave it to the computer, whether it worked or not, and then you took it back and then you wrote more piece of code and that's it, right? These days, wipe coding is basically the idea. And if you want to learn, say, a new programming language, you went and learned how it sounded, how it looked like, you went to various corners of the internet, you figured out how to use it idiomatically, just like you learn a new language. It's one thing to look at the French dictionary, it's another thing to be able to speak in French idioms where people are like, I think this person knows French.

3:41:32And you have the same with programming languages. But AI models are incredibly good at coding. There are a couple of reasons for this. One is that there's a lot of code on the internet. So AI models are just trained on a lot of code. The second more interesting reason in my mind is that coding is a way where the models can actually get better by just by themselves. They can basically generate some piece of code and they'll be like, is it good? Let me run it. Oh, it was not good. Let me make myself better. So coding is one of these things where they can just like learn much better without having to have human input.

3:42:08If they wrote a poem, it's much harder to basically, hey, is this poem better than that poem? But with code, there is an objective answer. Now, I'm oversimplifying, but there's a couple of reasons why coding has gotten just dramatically good on these models. So vibe coding is this idea where you ask these models to essentially generate code for you. You can do this right now. Have you ever written code at all? No. Oh, amazing. Let me get you into this. This is one of the most... You're going to teach me guns, and I'm going to teach you code. Okay? Let's do this, right? One of us is going to look way more badass than the other.

3:42:41But historically, even five years ago, I was like, well, I'm going to send you a book, or I'm going to send you a YouTube video. Because somebody would say, well, let's say, give me an example of something you want to do. Maybe an, do you have an, does this show have an app or a website? You have a website. We have a website. But do you have an app? We have no app. Great. Let us say you want to build an app for the Sean Ryan show, right? It notifies you in this new episode, collects email addresses. You can sign up to get early access, all these kind of things that app might do. We should build that.

3:43:16Please, you should. Or somebody watching should build it and get your attention. But now what you can do is I can teach you to do it right now. What you would do is you would open up one of these models. You would say, I know nothing about computer science. I have no background in writing code. take this YouTube channel and build me a mobile application on my iPhone, runs on my Android, which does all the things I just said. That's it. And what it is going to do is it's going to start generating code for you. It might ask you some questions, like, how do you want the screen to look like? Maybe you say, look, I want the screen to have this color, have this functionality.

3:43:53It's going to generate that for you. And then it's going to maybe even run that for you. And without maybe even you individually writing a single line of code yourself, you could have today a fully functional mobile app. People actually build much more sophisticated experiences. You should try this right now or after this. You know what, I'm going to show you maybe a demo and you should try this right now. And I think this is such a superpower because computers are often sort of this arcane thing where you're like, well, I'm going to go to school, I have to teach myself this. but with models, anybody and everybody can just get into it and they can amplify themselves and you can focus on the thing you want to do, building a great app for the Sean Ryan show.

3:44:33So wide coding is this idea where, you know, instead of spending a lot of time trying to think of what the right code to do, you basically tell the model, here's the code I'm trying to build and one of the models says, all right, yes, yes, yes, just keep going, keep going, keep going. And instead of a little bit of a tongue-in-cheek idea, The idea is that you don't get perfect code. Obviously, if you're writing this in production, if you're writing this for running inside a bank or a nuclear reactor, you absolutely want to make sure you check it and you know exactly what it is doing. If you're doing it for fun, it's great.

3:45:08You can explore things, try out a new idea, maybe build something as a hobby. So wipe coding is that you can just go with the model. So that has fundamentally changed writing code. where I don't know the exact stats, but a lot of tech companies, I would say 30, 40, 50 % of their code is now written by AI. Anyway, so my point is to kind of going back a little bit, AI models, instead of having this one model which takes over everybody and becomes sort of this gigantic terminator brain, you now have multiple models which have specialized. I'm the great coder model, or I'm the great, you know, one of the great personality.

3:45:48And so we see no signs of takeoff happening. So I think the doomers in my mind have been completely proven wrong. Is takeoff a, I mean, would you say it's a possibility? Well, theoretically, absolutely, right? But for me, you have to come back to science, right? You have to come back to the scientific method. So I would say you have to show me proof that it can happen. and every data point we have now is pointing the opposite direction. And what we were doing is in fear of, in my mind, this theoretical scenario, which has had no existence proof of, we were basically shooting ourselves in the foot.

3:46:28We were trying to stop AI. We were trying to ban open source AI. We were trying to stop our AI from being used by other countries, our allies. We were trying to stop our allies from using our GPUs and chips because we were worried they would build AGI. At the same time, China was just searching ahead, building these models and building this chip capability. So if you ask me, is it a possibility? Anything is a theoretical possibility. But we live in the world of reality where you have to be like, you know what? Show me the empirical evidence that we have any proof of this happening. When I have so much evidence of things in the opposite direction.

3:47:09One of the things I heard the president say is, I'm not sure you've heard this, he says he hates the word artificial intelligence. He hates the word AI. And he's very funny about it. But I think there's some kind of truth to it. Because the word AI, I would say, almost oversells the space a little bit. Because in my mind, I just don't think of this as this thing where we're getting into some sci-fi future. I think of it as the next great computer platform. This is like the internet. Peter Thiel says, on the scale of this is a nothing burger, or this is some sentient AI sci-fi race, I'm somewhere in the middle.

3:47:51And I think I agree with him. This is like the internet. This is like the mobile phone, maybe bigger. Is this going to fundamentally shape humanity? Yes, the internet did. Mobile phones did. AI is going to. But I have not seen the evidence that this is going to be some all-knowing, all-powerful God that takes over all of us. Makes sense. Makes sense. I mean, so, you know, for the doomers, I mean, I've listened to all these things and all these different theories on what it could do, what it could turn into. I mean, what would it take to kill it? I mean, wouldn't it just take removing the power source?

3:48:29It's a good question. Let me ask you, though. What are the theories you've heard? everything that you've just said. Okay. It's going to take everybody's job. It's going to turn into the Terminator. It's going to kill everybody. It's going to wipe out humanity. Yeah. Everything that you had stated above. Let me ask you. Well, let me ask you. I would say one of the challenges of some of these questions is they are almost a theoretical thought exercise where it is so hard to disprove something which is so theoretical, okay? But let me give you an answer. imagine you know tomorrow there was this idea that insure what sriram was wrong you know he came on the genre and show he was wrong we are actually seeing signs of these models taking off and maybe wanting to do bad things to human beings if that happens tomorrow what do you think you me all of humanity is going to do do you think we're going to sit still no yeah do you you're going to allow that to happen?

3:49:26Like, do you think, you know, all these other people, companies, engineers, do you think that, well, that's it then for the human race. Let's pack it up and go on home. No, they're going to stop it way, way before it ever becomes a thing. Like we talked about social media, right? Like, you know, social media, you know, today versus 10 years ago is so different. There's so many checks and balances and laws and regulations. So this idea that you go from this sort of what we have today into this all-knowing God without reasonable people, whether it be in government, technologists, just regular human without being like, hey, you know what?

3:50:04Let's hold up a second here. Let's just stop and think about it. Let's put in some safeguards. Let's have ways to counter these AI models. That will definitely happen. There is no way we just get from here to there without a bunch of things in the middle. So sometimes when people say, how do you stop this all-powerful AI, which is taking over a data center? I'm like, how did it take over the data center? What are the 25 steps which happened before then? How did it amass all this power? I'm pretty sure somebody stopped it when they took over the second data center. So that's my first sort of instinctive reaction to when people post those questions.

3:50:43The second thing I would say is the best answer against models is other models. And I think it's very true, by the way, in what I think sometimes the future of cyber warfare might look like, where the best way to defend against maybe a model which is showing this crazy capability is have another model, which is looking for these capabilities. So, look, I hear the sci-fi. I grew up on sci-fi. I grew up on Star Trek. I grew up on 2001 A Space Odyssey. I've seen the Terminator movies. I know all the risks. I know all the theories. The thing is, we have no proof, no evidence. We are anywhere on the track.

3:51:23Second, we have a lot of evidence. We are in the opposite direction. This is an amazing platform. There are some questions in terms of how do we benefit humanity, but there are no signs of takeoff and sci-fi behavior yet. And most importantly, China is searching ahead. So if we rest, if we stop ourselves, the other side is not. And I understand that. Before we move into China, I do want to ask you one question. I mean, do you have any concern about people in their relationships with AI? So we're starting to see people ask very personal questions. Use AI as a therapist. Should I get divorced? How should I discipline my kids.

3:52:08And they're asking AI very, very personal questions. And the, I don't do this, but the AI will spit out an answer. And then I think that, I mean, we've seen manipulation with social media to an extraordinary extent. And so, you know, I think that the manipulation through the potential of manipulation through AI could be even bigger than what social media has done. I mean, do you have any concerns about? Absolutely. I would say this is at the heart of why we made that executive order happen. Exec heart. Like, imagine some young kid influenced us AI a very personal question. we don't want ideology to influence an answer.

3:53:05We want the honest truth. So one of the things executive order does is to basically say, if your model is not truth-seeking, the government will not work with you. And I think that I absolutely have that concern. There's another very interesting concern, which I think is going to come up more and more, which is the idea of privacy or confidentiality. If you go to your doctor, your lawyer, or your priest, what are you promised? You know that nothing you say can ever get out of that context. There are laws, there are social conventions where you have attorney-client privilege. You have doctor-patient confidentiality.

3:53:47Various religions have this construct where what you say is sacred, except if you do something really crazy. Now with AI, we are starting to see people ask very deep personal questions. they say, look, here's my medical report. Like, what does this mean? Like, you know, give me a sense of what my blood test means. Or, you know, maybe I'm not feeling great about myself. You know, what should I do? Or maybe just basic career advice. I know a lot of people, you know, ask AI for career advice. I would think it would be weird for all of those to be public. Imagine if somebody could say, you know, if you gave a piece of AI your medical history and somebody could just say, you know what?

3:54:30Just like I can get access to your emails, I want to get access to every single thing you're asked, a piece of AI. So I don't know what the right answer is, but I do think the way we work with these AI models is a little different than other pieces of technology. And I do think we're going to have this public conversation about what are the right legal constructs, kind of protect that. But yes, I absolutely do have concerns about ideology, and that is at the heart of why we did this. Yeah, I'm not just talking about privacy. I'm talking about, I mean, what if somebody were to, I'm just pulling something out of thin there.

3:55:10What if somebody were severely depressed, you know, and they are, you know, they're asking an AI series of questions that it ends up with, you know, should I kill my partner because they upset me? Should I kill myself? and I mean the AI is going to respond to that. Oh yes. You know and so that's kind of what I'm getting at is when I'm talking about manipulating population privacy is another concern but I wasn't there yet but you know it has the potential to manipulate entire populations the entire population. Absolutely and there have been these very tragic incidents I would say in the last few months where AI has encouraged people down a very, very bad path where a regular human being would have been like, hey man, maybe you need to get some help.

3:56:05Maybe you should have a conversation. And so I think one of the things which a lot of the leading model companies are working on is addressing sycophancy. This idea that you just don't want an AI which just agrees with you, but you want an AI which spots patterns of, you know what, this person may need some help. Or maybe we need to alert law enforcement. Is there a kind of precedent for this, by the way, in other places? One of the things I would say about social media platforms is they just see a lot of very dark things in humanity. And people wanting to do things for themselves, people trying to do really terrible, bad things to others, and they build a lot of systems over time to try and kind of deal with that.

3:56:52And most social media platforms, if you, I'm not going to say every single time, but they try, like, if you try and do something where you might be harming yourself, or maybe you express a desire to harm someone else, they try and find ways to directly. They're not perfect, obviously, you know. And I think AI companies, this is going to be an existential question for them. They need to find ways to figure out how to spot it when people are going down dark paths and make sure either they're alerting someone or they're kind of coming out of it or you're just not saying, hey, man, yes, you're absolutely right in everything you believe.

3:57:24And I think this is going to be a key, key topic for all of AI. Okay. Okay. Let's talk about China. I mean, that's the big concern. I mean, Xi Jinping has said, you know, the winner of the AI race will dominate the entire world. I mean. I have a general idea of how that would happen, but how far ahead is China than us? Well, I think we are ahead right now, but maybe it's more useful to break down a little bit of a scoreboard. Okay. I would say let's start with the basics. What AI needs is, number one, is infrastructure. Infrastructure in terms of energy. Energy which powers data centers. Because again, the scaling loss, more energy, you get better models, you can use more AI.

3:58:21And there, I think one of the fundamental challenges the United States has had is for many, many years, our energy usage as a country has been fairly flat. I think it's improved by a small percentage every single year. And then all of a sudden, AI shows up and you need a lot more energy. You need a lot more data centers. And then all of a sudden, you have this spaghetti bowl of issues which suddenly come up. The first one is, where do we get the energy from? And that is where I think the current answer is absolutely natural gas. That's where I think it's going to power a lot of these AI along with what the president would call clean, beautiful coal.

3:59:06but also the future is definitely going to be nuclear. That's going to be a big part. On this, I would say China is ahead because they have just invested in building out their energy grid, building out generation. They've done a lot of work on nuclear and building out the grid that transmits power. We, on the other hand, one is we have work to do across all these fronts. I'll talk about what we are doing as this administration, But one, if you need more energy, right? Second is we have all these outdated, broken rules and laws which stop data centers from being constructed just from a lot of, I would say, completely nonsensical climate concerns.

3:59:52And we need to get rid of the red tape. We need to get rid of the red tape and we need to, what the president called, build baby, build. We need to build these data centers. We need to build energy. and we need to also figure out a way to upgrade our energy infrastructure. So we, and we have a peggotty mess of issues. So this administration, you know, we did a bunch of things to attack this. There's an executive order, which has, you know, come out, which is tackling nuclear, which has said, I think, I forget the exact number of years, but I think for 30, 40 years, the NRC, the Nuclear Regulatory Commission, hasn't approved a single reactor.

4:00:32and I think there's a whole future. I know you had folks from the nuclear industry here with SMRs and so on. I know it's like a little bit of time away, but I don't think we had an executive order which basically says, look, you know, we cannot believe, you know, the nuclear definitely has a strong future in America. In terms of the present, though, we need energy now, and this is where the president set up something called the National Energy Dominance Council, the NEDC, which brings together the Secretary of Energy, the Secretary of Interior, and we've been working closely with them, which basically tries to attack all of this.

4:01:08We have to say, how do we remove the red tape on building data centers, on permitting, on regulation? What are all the things that we can do to just get more data centers built, more energy going? So just in the scoreboard front, though, this is one where China's ahead. And we've obviously done a lot to catch up, and I think we're going to search ahead. But for the last four or five years, they've just been on a great trajectory. And I think we have a lot of great work here to catch up and obviously exceed them. The second part is chips. So chips are super interesting and we should spend a lot of time.

4:01:42I see on your shelf, you have Chip War by Chris Kelly out there. On chips, we are ahead, but maybe not as much as people think. So when I talk about chips, there are multiple layers, But essentially, for AI right now, the most important ones are the GPUs, which are built by people like NVIDIA and AMD. And then, of course, you have Google, which has built their own hardware in terms of TPUs. And then Amazon has their own hardware. But NVIDIA, AMD, all these companies are obviously ridiculously important. And they are right now really far ahead of what the latest and greatest from China, which have companies like Huawei, which builds a product called the Ascends, or the other companies called Cambricon have.

4:02:30Now, why are we ahead? There's multiple answers, but part of it is because we have access to better technology from TSMC in Taiwan, which I know you're very, very familiar with. We have access to much better software, which runs these GPUs. And we just have like a lead on them. Now, it turns out though, on this case, China has done a lot of work on catching up. And instead of having a multi-year lead, I think our lead is much, much smaller. And right now, what China has been doing with companies like Huawei and Cambricon is really worth paying attention to. A few months ago, and this is, I think, a very interesting one to look at, Huawei came out with this product called Cloud Matrix 384.

4:03:18And it's worth Googling and looking up. And the reason why this is important is if you think about these data centers, right? If you take, say, for example, OpenAI or Grok, they have a bunch of NVIDIA GPUs inside them. And they have them in these clusters, right? Like if you have a H100, which is what a lot of people use, you have eight of them clustered together and you have maybe several hundred thousand of them. Or more recently, NVIDIA has come up with this product called the Blackwells, where you connect 72 of them. But you kind of cluster these. And the idea is the more of these you bring together, the better it is for training a model or, you know, inferencing from a model.

4:03:59Cloud Matrix 84 is interesting because people believe that China was way behind innovating GPUs. And what they did was they built a cluster where you take 384 GPUs, Ascend Chinese GPUs. Now each of those use way more power than NVIDIA. They are not as fast, but it doesn't matter because guess what? China has more power. They don't care as much as we do. Second, Huawei has really good networking technology. They're a networking company. So they're able to basically connect, if I'm oversimplifying, a lot of not as great GPUs, but to try and compete with much more powerful GPUs. And I think of this as an interesting example, just like DeepSeek, of when we try and say we are not going to let the world have our technology or let China have our technology, they're often been very good at working around in other very creative ways.

4:04:53So that's number two. But I think on the chip side, we still have an advantage in the performance of each GPU and also how many we can make. For a lot of reasons, we can just make several million of these. And China is definitely a lot more hamstrung in how much they can make, but we are ahead. Models is very interesting. On models, if you talk to somebody on January 15th or January 18th, they would say, oh man, American models are way ahead. We have OpenAI at the time, 01. We have Cloud 3, or we have whatever, Grok 2, I think, at the time. And China doesn't have anything. DeepSeek totally, in my mind, demolished that idea because all of a sudden they were not ahead, but they were very close.

4:05:42Okay? And one of the phenomenons that happened in AI is this idea of distillation, which is you can take a really powerful model and you can distill it to make a smaller, not as powerful model, but just slightly close behind. TLDR, what happened is China proved to us that they can build really, really good models in some ways surpassing what we have. I was at a developer event in San Francisco recently, and I asked the crowd, what are you guys, how many of you are using a Chinese model like DeepSeek, Kwan, KLM, there's a bunch of others, and almost everybody in the room put up their hands. And I was like, whoa, this is not good.

4:06:27Why is this not good? Number one, it is soft power. China having a technology platform, which is now running inside American, hopefully not American infrastructure, but American companies, definitely running in global companies, where instead of our models. Second, these models also communicate culture. Think about my story. I grew up on the internet. I grew up on the English internet. I absorbed a lot of American culture just because of the prevalence of America winning the internet. But if China dominates the model race, if you look at DeepSeek, it doesn't really share our ideology. And we don't want that to be the dominant ideology and values around the world.

4:07:17You asked about people asking models very personal questions. Think about a world where people ask a Chinese model very, very personal questions. I'm not sure I would want that. So one, on the model side, I think they've caught up very close behind. I think they're ahead on open source. We are catching up quickly. There's been some great new launches recently. We are still very much ahead on closed source models. The latest and greatest, GPT-5, Gemini, Grok, we are still ahead, but it is a much closer race. The last and maybe most interesting part of the race is what I would call diffusion, which is how are people using AI?

4:07:57Because at the end of the day, AI needs usage to be better. One of the reasons ChatGPT became really good is because when people used it, you could use that feedback mechanism and become better. This is very important, not just for AI today, but important for the future, like robotics, One of the things which is going to be key to getting robots, American robots all over the world, is can we get that robots data of usage either from a simulation real world and make it better? So I think there's a race right now, which is who can get their AI to spread faster, to be used faster. Okay. And historically, when you grew up, did you use Windows laptops or Windows computer when you grew up as a kid?

4:08:41Do you know who Windows' competitor was in the 90s? No. Nobody remembers. There are companies like IBM's OS2 and others. They all got crushed because Microsoft, Windows, along with Intel, dominated. Why? They got everybody to use it first. And they got all the developers to build applications on it first. They probably built Office, used Office, games that you use. Netscape, now everyone used Windows. So when you get everybody to use your stuff, you get this ecosystem flywheel, right? More smart people start building on your stuff. Your stuff gets better. More applications make your platform better and on and on and on, right?

4:09:21And right now, we have a window of time where we can make American AI the default here around the world. and if we don't, China will try and make their AI, their chips, their models, future of their robots, the default around the world. And that, for me, is the heart of the race. So with all the concerns and all the different sectors, it sounds like we're ahead on chips, software, but we're behind on energy. And energy seems to be, with the little I know, You know what I mean? Energy seems to maybe be the most important factor in AI and data centers and all this other stuff. And so what are we doing specifically to unleash?

4:10:08I mean, I've dove into the power grid, the vulnerabilities. I mean, it's atrocious of neglect for our energy system, our power grid, and for years, maybe. I mean, years and years. And if China's ahead, I mean, what are we going to do to unleash nuclear power? I mean, it seems like that's the key. Yeah. And we have all these innovators that are, we talked to Isaiah Taylor, who's building the mini reactors. We talked to Scott Nolan, who's enriching uranium.

4:10:50But we need to go faster. We need to go faster. And all these guys are very impressed with the current administration and getting rid of some of the red tape. But I still don't feel like – I mean, if China has zero red tape, how the hell are we going to compete with our power grid? I mean, we talked to Bajibat, you know, the founder of Robinhood, and he's trying to beam solar energy in from space to receivers that are going to get averted into the grid. I mean, we've got all these amazing ideas, but it's just not going fast enough, in my opinion. Well, I think I totally agree with you. This is one of those things where we just need to move as fast as we can.

4:11:41So I would think of it in a few layers. I absolutely think nuclear is the future, but I think we're still a few years away from getting there. And we have data center needs right now, like today. What is stopping the training of the next large model? Well, we need to have a larger data center. And we are seeing entrepreneurs all the time, like Elon, for example, if you see how we build Colossus, it's in Memphis. And so he builds this old, he buys his old, I think Electrolux factory. And then he drives in all these generators on all sides and then uses Tesla solar packs to basically even it, you know, kind of augment it, write some code to even it all out.

4:12:28and it's just one of these amazing speeds of engineering. But the point being, we need energy right now. So just on the nuclear front, it's absolutely the future. So administration has an executive order out, which basically tries to clear the red tape on the permitting for all things nuclear. But my sense, and I'm not like a deep nuclear person, is my sense from talking to the best people in this, is like, we're still looking out like a few years. The game right now is gas and gas turbines. And the challenge there, well, on the energy production side, then you have on the energy grid side, to your point, we have a very old grid from one infrastructure capacity.

4:13:11And second, we have all these local state utilities and monopolies which have weird regulation, which makes it very, very hard to take power from one state to another. There's a whole cluster there. So we came out with this document a few weeks ago called the AI Action Plan, which is the entire AI strategy for America in my mind. I spent a lot of time on it. A lot of others spent a lot of time on it. You know, the president announced it. And in that, one of the top priorities we talk about is, one, removing the red tape for data center construction, right? With things like how do we make sure that is the NEPA, the National Environment Protection Act.

4:13:53How do we find carve-outs so that we can get data-centered construction just going, right? Let's just get the red tip out of the way. Let's go build, build, build. There are directives in there on figuring out what to do with our grid capacity, figuring out incentives on energy. I think the president in his first week announced this large project called Stargate in terms of getting more investment and getting more of these data center construction going. So in my mind, nuclear is the future. I think we've done a lot to clear out the red tape, but I have a race to win right now. Like every entrepreneur is like, look, I need to get 100 ,000 more, 200 ,000 more GPUs, unlock, what do I do?

4:14:39So we need to unlock that. And I think unlocking data center construction, grid capacity, clearing of the red tape so the power can come to the data center. how do we find smart ways to go do that, really innovative ways to go do that. That's the game right now, which I think, you know, and I don't want to take credit for this because I do think the Department of Energy and Interior have done a lot of great work on this. They are very, very focused on, all centered around gas and coal. Moving on to chips. I mean, you say we're ahead on chips. I think it was, was it April? of this year, we said no chip sales to China, correct?

4:15:23From Nvidia. Was that correct? Not really. In August, we said, I believe it was August, right? It was last month. August, we said Nvidia can sell chips to China, but we get 15 % of the revenue, correct? The H20 chips, which are the high-end chips, correct? Was that a mistake? So actually, there's a few corrections in there. Okay. A little bit of history. So, NVIDIA makes a whole set of chips. And the latest, greatest generation is called the Blackwells. They usually start with the letters G and B. The previous generation was called the Hopper. They started with the letter H. The top of the line was the H100, which honestly is what a lot of people have right now, or the H200.

4:16:08They're kind of the top of the line in America. Now, about two and a half years ago, the Biden administration came out and said, hey, we're going to... Remember what I said, the Biden administration had all these mistaken beliefs. And one of the beliefs I think they had was that China, one, our allies can't have our AI chips and that China can't ramp up chip technology. So, because since they believed that, they said, number one, we're going to put limits on the power of the chips, the flops of the chips that China can get. And then we're going to slice the world into three categories. This is called the Biden diffusion rule, where some countries like USA, UK can get any number of GPUs they want.

4:16:56About 100 countries couldn't really get many GPUs without going through this crazy amount of red tape. And then some would just not get any GPUs at all, like Iran, North Korea, you know, the countries of concern. And as a result, like, you know, a couple of things happened. One was that a lot of our allies were like, well, we want to use American stuff, but you guys won't work with us. And they were just left out in the cold. A good example is like the Middle East, where the Middle East and countries like other countries are like, we want to bring AI to our citizens. And what is a way where we can get your GPUs?

4:17:35And the Biden diffusion rule said, there is no way you can ever get our GPUs. That's number one. So we essentially kind of turned our back on our allies all over the world because we thought, number one, you know, there's a supply constraint in GPUs that if a GPU ever goes out of America, it means one less GPU for an American company. We thought China will not compete on chips. Third, we thought AGI can only happen in America. It's a scary, scary thing. So we came in and we were like, this is just wrong, you know, because number one, we're making a lot of GPUs. If you want, you can go get a H100 like right now, an H200 like right now, because it turns out that all these semiconductor companies, when they find a supply constraint, they're very, very good at innovating and working on it.

4:18:26They're very capitalistic. They want to make money. So number one, if we ship a GPU to somebody else, that doesn't mean one less GPU to America. Number two, as I said earlier, I don't think we worry about AGI exploding anymore. We talked about that. I don't think we worry about takeoff anymore. I'm not so much worried about an allied country building a super intelligent AI ahead of us. I just don't think that is a realistic possibility. So we said, first of all, to our allies, we're going to tear apart this 200-page crazy document called the diffusion rule. And the first thing, first, I want to talk about the Middle East, then I want to come to China.

4:19:02President Trump, in his first state visit, he went to the Middle East, and we struck these deals called the AI Acceleration Partnerships. And his idea was that we will sell you these GPUs in return for investment in America and with ironclad security provisions to make sure that those GPUs are going to stay where we ship them. They're not going to go to some other country, and they're going to have somebody we don't accessing them. And the idea behind all of this is we want our chips, our models to be used the world over. We want to be like Windows. We want to be Intel. And we don't want to risk competition.

4:19:45Because in the meantime, what wound up happening is Huawei has been ramping up chip production. So we now have a competitor who has a good product. They build Cloud Matrix. They are building a sense. There is a story in Bloomberg last week about Huawei exporting chips or trying to export chips to multiple countries around the world. So if you think of American AI as a product, we have a competitor who's out there selling a competing product out there. So our stance is that when it comes to our allies, we want them to have American AI, our chips, our models, in return for investment in America and with the right security safeguards in place.

4:20:26I always, always underscore that. It is in the terms that we sign, for example, a couple of these Middle Eastern countries. Now, hopefully that's kind of clear on why we want our allies to have our technology. China is an interesting different case. There are two schools of thought on China. One school of thought is what you said, which is we should just deny them any chip, any GPU, right? And the idea there is that, hey, if we deny them anything, you know, number one, they're not making their own GPUs. They're not really making their own model. That was the original thinking. So we will just make sure we get this race to AGI.

4:21:05We will be far ahead. It turns out that was fatally flawed, okay? Because number one, China's making chips. They're innovating and they're building great models and they're going, you know, knock on other people's doors and be like, hey, America is not selling you, the Biden industry is not giving you any of their technology at all. Maybe you want to work with us. Okay. And so we think the answer is we will always keep the latest and greatest GPUs for the United States in huge quantities. That is irrevocable. Like, you know, we'll be the only country which can build these massive, massive super clusters of GPUs.

4:21:44But on the other hand, we want to make sure that we are not giving Huawei, Cambricon, all these Chinese competitors, oxygen. Oxygen of growth, oxygen of revenue and usage. Because when somebody uses a GPU, they're writing code on it, they're finding issues, they're finding bugs, they're making it better. So what is the answer? How do we thread this needle between let's keep our big stuff, but let's not give the competition any oxygen? And I believe the right answer is let's ship them technology that is ahead of what the competition has, but way behind both in individual performance and also in quantity what we have.

4:22:28So that's the principle and the framework. So if you're going to follow that, where does the H20 come in? The first thing about H20 is because some of these Biden staffers like to talk about this is Biden never banned H20s ever. They were completely allowed all the time through the administration. Because what happened is when Biden's team built out these rules, NVIDIA went out. And by the way, NVIDIA has H20, but AMD has something called the MI308, which is the equivalent of the H20. They said, okay, both of these companies, we are going to build a nerfed version of our GPU. You know what nerfed means in gaming lingo?

4:23:10It's a less powerful version of our GPU, which is way behind what the latest and greatest America has. We are going to ship this only to China because it will keep us ahead of what Huawei has so we can get these Chinese customers using it, but we are way behind what America has. NVIDIA built this. AMD built this. The Biden folks were completely okay with this. So we came in, and the first thing we said is, one, we tore apart the diffusion rule for the rest of the world. When it came to China, we said, okay, we need to know exactly what is going on when we ship GPUs to China. So we said, we're going to bring in a licensing regime.

4:23:49So every time you see a headline that says, you know, President Trump banned H20s in April, March, not true. we brought it under a licensing regime, which basically said, look, if you're going to start to export this, you need to ask us for a license first. Actually, the technical term is called an ASIN formula. We send to these companies, we need to come ask us for a license first, right? So we know how much are we shipping and who's getting it on the other end, right? And now, you know, as you mentioned a couple of weeks ago, these are my partners in the commerce department who are actually in charge of the licenses.

4:24:22Like they kind of, they take all the credit for thinking through all this. Them, along with the president, said, okay, you know what? We're going to get a great deal for the American people. We're going to start approving some of these licenses of these much older, less performing chips, way behind what America has in way less numbers than what America has. And we're going to get a great deal for the American people. So that's the history. But that's all the past. I like to think about the future because the H20 is now two and a half years old. and every company has a new generation roughly every single year.

4:24:57And one of the amusing things for me is we are still talking about this even though since the time the H20 has built, we have been through two iPhone generations. It's like we're talking about, I don't know what the latest iPhone is, we're talking about iPhone 11 and we're several generations behind. This conversation is going to keep coming up again and again and again. So what is the right long-term strategy? In my mind, it is, one, we need to flood the zone around the world with our allies with American technology. We're starting with American GPUs. Why? You know the whole old business model of Gillette, which is you sell them a razor, but you make money on the blades?

4:25:35You remember that? GPUs and models are a bit similar. If you are a country and you spend a few billion dollars buying American chips, guess which models you're going to use? Probably American models. You're going to make both those models and the chips better on using it. When you build a next data center, you already have all these innocent American models. What are you going to do? Probably buy American all over again. Now, if we refuse to do business with you, you're probably going to go to a competitor at some point in time because AI is just too important. We just can't keep telling people to pound sand every single time they knock on the door and say, I want to bring my citizen to AI.

4:26:11So, one, we should flood the world, our allies, with American technology, with GPUs, with models. Again, America will have an overwhelming lead in the quality and the quantity of these GPUs. We'll be the only ones who are building, you know, Elon, I think, is going to build a million-plus cluster this year. I think Google has way more TPUs. We'll be the only country which can build these massive ones. But if another country is like, hey, I want to bring my citizens education, I'd much rather them doing it on an American GPU, running an American model, generating American tokens. So that's for our allies, right?

4:26:51And we can talk about who the allies are. For China, our belief is the right strategy is find a way where we keep the latest and greatest, but, you know, ship technology. So if they want to build a chatbot, if they want to build education, I would much rather have them use an older, nerfed, a lower quantity version of our GPUs rather than buy a Chinese company and then keep improving it. Because the reason is if we do that, what happens? Let's say we ban all American technology China. We say you can't get anything. Well, they are going to say we need to accelerate all of our indigenous chip making efforts.

4:27:33because you guys left us no choice. This happened before, by the way, about 10 years ago, right? Like we stopped exporting supercomputing chips to China. And within a couple of years, Chinese indigenous supercomputing ecosystem just exploded. And they have, I think within a few years, like much better supercomputers than the chips we were exporting them to. So if we force their hand, they're going to now build out an alternative technology stack. And then they're going to start exporting it. They are going to go to other countries because we are being difficult to do business with in this Biden scenario, we're like, you know what?

4:28:09Buy a Chinese chip. And it's going to come freely loaded with DeepSeek on top of it. So that's the scary scenario. So I think the right answer is we ship China older, smaller quantity chips in enough quantity so that we retain the latest and greatest. We retain these super clusters, but it is enough to make sure the competition there does not get oxygen. Now, I understand everything you're saying. You know, I mean, I know you're very aware of the China-Taiwan conflict. I mean, so just the fact that we are sending them older technology, older chips that aren't up to snuff with what we have. I mean, do you think that's more motivation for them to make a move on Taiwan?

4:28:59To take over? I don't want to speculate. It's hard to say.

4:29:06Because then they cut us out. Well, my belief, I'm not a deep geopolitical expert in Taiwan. I'm much more familiar with the semiconductors. I think the motivation there is more about sort of how they believe history. But I think it's a good question, which is— I think there are other motivating factors. There are other motivating factors, right? I'm not the expert. but I think some of you had a lot of great guests on. Number one is the Biden rule definitely amplified that. Because if we go tell people, pound sand, you're not getting anything, you sort of have to find ways to find alternatives.

4:29:44We are relieving the pressure, right? Again, I keep emphasizing it because we have the latest and greatest, and we'd be the only country who can build these million GPU clusters that any one of these model companies have. But at the same time, I want to make sure that they get enough to stop them feeling like, oh, my gosh, we're just going to go full on out to build out our entire ecosystem and go talk to somebody else. So does it take away the motivation? I don't think so, but it's definitely better than the previous alternative that we had last year. I do think you bring up another interesting question, which is the Taiwan DSMC question, which is that we as a country, I guess the world, essentially has a reliance on a couple of companies around the world who are essential.

4:30:39One is ASML in the Netherlands who build these lithography machines. And the second is obviously DSMC. and if something happens in Taiwan, any number of situations could happen, but it probably means, you know, we don't get iPhones. You know, we don't get AI chips. It's not going to be good. And I think this is where the TSMC project in Arizona, all of President Trump's efforts to onshore these capabilities, the recent deal that President Trump and the Commerce Secretary did with Intel on this all kind of play a part, which is this is such an important capability that we as America need to have on our soil.

4:31:31And we need to find ways to bolster our entire chip manufacturing supply chain and not be so reliant on one single point of failure. So when I think about TSMC, I'm thinking about all things Intel. I'm thinking about all the things in, you know, how do we get these fab construction in Arizona, you know, going faster. How is that factory going? I mean, from my interview with Baal Shikim, it sounds like, you know, the VP of Taiwan, I mean, it sounds like that the tariffs may have had some implications on the speed of that. Well, I think they went a lot faster than the Biden regime, for sure. but I think there's always space to do more.

4:32:21I would think that one of the things that we have really tried to emphasize is that we just need to accelerate our indigenous supply chain. And I don't have the latest numbers on how these fab production is going. I believe they've kind of increased their capacity in some way than what it was two, three years ago, but we need to do more. I believe even if the fabs are at the current rate or full potential, it's not going to be sufficient to make up for all the things that DSMC winds up making. So we need other answers. We're probably going to need Intel. We might need other answers in there. I think this is going to be one of those interesting questions over the next few years in how do we build out these sovereign chip making capabilities.

4:33:08So that for me is one. But I want to kind of come back to the Chinese Huawei question because I do think it's important. Like one of the things top of mind for me right now is how do I stop? How do I make sure that somebody out in the world is building an application? They're building a critical piece of infrastructure. They are picking our chips, our models, writing applications on top. It's creating a moat and they're not picking the other team. Like that is very much on top of mind for me right now. I have an idea on how to speed it up. Please, yes. And you'll have to fact check me on this through my other episode with Xiaobu Kim.

4:33:46But one of the hurdles, I think, that frustrates them is they're very concerned, obviously, about China taking them either by cognitive warfare or actual kinetic warfare. and they purchased several planes, jets, military equipment from us, you know, but through our red tape and our bureaucracy, I mean, I can't remember, I believe it was about, I think it was, somebody's just going to have to fact check me. I can't, I can't look it up right now, but they bought these weapons like five years ago, you know, and they just got their first F-16, you know and and that's created a lot of frustration over there and so if there was a way to you know get through the rev tape and get the the the stuff that they had purchased you know from us in a in a in a in a faster way i think that would help motivate them to help us build the chip factories in arizona okay not super familiar with that i need to go back and do some homework on this.

4:34:53Yeah. Yeah. I mean, I know that they're frustrated about that, and rightly so. I mean, shouldn't have to buy something five years in advance before you get it. But I mean, so how far along are we in that facility?

4:35:12I wish I knew this off the top of my I believe, man, I think they had faced a few challenges in the first few years around permitting and workforce. I believe they've just entered production recently, but we just need to get a lot more going there. So it's up and running. I think so. Okay. Okay. Well, that's good to hear. Yeah. I think so. That's great to hear. But I don't want to get fact-checked on this. Like, you know, the way I understand it is that, and I'm going to be spending a lot more time on this because I've been spending time on some of these model questions, is that it was in this really slow phase for a little bit of time because they had a lot of issues with finding the right skill set.

4:36:00I think at some time they had trouble finding enough electricians and other skilled technical people. But I think it's ramped up now. But I don't have the latest and greatest on exactly where they are right now. Could you paint a picture of what it looks like, not only for the US, but for the world, if China does win the AI race? What are we looking at? Oh, man. Oh, man. Yeah. Well, look, there are multiple timelines AI could take. I think it's pretty obvious that AI is going to have profound impacts on the economy, on productivity, just making people's jobs better on drug discovery. Imagine a world where China just dominates drug discovery.

4:36:50Imagine a world where every Chinese individual is able to be smarter, more productive, get more done than what we can do. Imagine Chinese companies being able to build faster, being able to innovate faster, and that flywheel just picking up momentum. Imagine a world where our allies are using Chinese technology. This has happened before with 5G, where a lot of the world wound up running on 5G technology from China and instead of something from a Western country. That caused a whole set of issues. Imagine a world where AI being so much deeper. You mentioned it knowing people's personal issues. It knows how your business works.

4:37:43It knows your numbers. It is helping you make critical decisions. All of this being run on Chinese models and the influence and the power it would have. So imagine a world of robots building inside factories, inside people's homes, which are all powered by Chinese software and hardware all over the world. And again, this is for me a nightmare scenario. We are not going to let this happen. But that was, I think in some ways, the path we were on. Yeah, I mean, I talked a lot about this with Alex Wang as well. And he had talked about the number of AIs against number of AIs. So we would need more, you know, if they have, you know, 100 AIs and we have 200 AIs, then they could dedicate their, you know, allocate, you know, I don't know, 33 AIs to service more per vehicle.

4:38:36and if we have 200, we could put 66 against 33? I sometimes find those a little too theoretical. I often, when we talk about AI, I often want to ground it in terms of how does it make a regular person's life better? And I just think about an individual who can just do more. They are smarter. They are better informed. They can get more done. Because they're getting more done, they can work more effectively together. They have an always-on smart colleague in an AI who they can ask things to. The companies can innovate faster, build more. So sometimes I worry that we get lost in a little bit of the theoretical elements of AIs versus AIs.

4:39:25Maybe that's true. It's hard for me to tell with high levels of confidence how that plays out. But I do have a high degree of confidence that AIs will need human beings and need to augment human beings. And I think about how do you make sure we are augmenting every single American, we're helping every single American every day with AI in going to work with their family, with their hopes and dreams much better than the other team are. And I also don't want that to happen using Chinese technology. Yeah, yeah. You know, I never thought about, you know, when we were talking about open source and with DeepSeq, I mean, if that's the open source model, then the entire world uses that, maybe except us.

4:40:12And so they're using a Chinese model. And so I understand that because I totally understand that. But what I'm getting at is, I mean, what are we doing about open source to spread? Is it unlocked the fact that everybody can use our stuff, but where are we with the open source? The first thing we did is we basically said, stop scaring people about this is a good thing. And this is very important. Like if you go to the AI action plan that I mentioned, which is the official document, it is literally the first paragraph, which is we want American open source to win. And the reason this is important is when the government takes a symbolic tone, it sets a tone for everybody.

4:40:55in terms of do we like this or not. So that's number one. So companies would talk to us, but we'd have companies who would tell us, hey, I want to build this open source model, but the previous regime, or other state legislators, they were scaring us. They were like, if you guys do this, you will have all these lawsuits and you'd get sued of business. What do you guys think? And our message has been like, we want you to go out there, we want you to win. We want you to win on open source across the world. And if you look at the last four or five months, again, I think China's done a great job. I think their models are really innovative.

4:41:32I just want to make sure we give them props where it is due. On the other hand, we now have our own open source models. There is OpenAI, for the very first time, launched their first open source model. It's called GPT-OSS. It came out three weeks ago. There are multiple startups which are now building open source models. Again, because when you have the government saying nobody should do this, it's very hard for a startup to go out and raise money. It's very hard for an academic to work on this. We are coming in and saying, no, we need this for America. But look, I'm not going to say it's all hunky-dory.

4:42:08That is a place where I think we need to do a lot more. So the action plan, it has a bunch of initiatives in there on how we are going to work with academia, on making open source more real, and how we are going to work with industry. So that's one. I want to talk about something else here, which is about laws around AI, because it's very related. So I mentioned when I was talking about open source, this thing called SB 1047. And this was a legislation from the state of California, which were essentially destroyed open source, American open source. What did it say? It's a little bit like gun liability.

4:42:43Do you remember when some people said, you know what, if somebody gets shot at the gun, the manufacturer should be liable? it tried to do the same with open source models where it said, if somebody does something bad with the model, Mark Zuckerberg or Elon Musk or whoever, that company should be held liable, which essentially sort of just destroyed open source because no company can really afford to take that risk. And the challenge is we now have every single state now wanting to do their own rules around AI, but sometimes around open source. So the president three weeks ago came out and I think said two important things.

4:43:20The first thing he said was that AI is a national security issue. So it is way too important to just like leave everything to the states. So I don't know about you. I don't want California setting the tone for how we use AI all over the country. And there's a Gavin Newsom joke in there, but I won't go there. Sorry. He did veto the last one, but I don't want to go there.

4:43:48But if you don't want any individual state to set laws for the country, how can that happen? Because if you have a really important state, make a law, the companies will be like, well, we don't have to abide by those guys. We don't have to abide by those guys. Let's just pick the lowest common denominator. And what is China doing? No red tape. Go, go, go, go, go. So President Trump came out and said, listen, you should listen to the speech. he's like, this is a national security issue, and we need to make sure that we are doing this at a federal level. So I think this is connected. Very related is this idea of copyright, where, you know, remember what I said about models?

4:44:29Models become better when you have more data, and you have more infrastructure. And the Chinese models are training on our data. but if our American models are stopped from training on this data, the other models are going to get further ahead. So I know this is a complicated discussion but in my mind I think of the national security dynamic of how do we make sure our guys are allowed to do the exact same things as the competition because if they are training on our data our guys should also be allowed to train on our data otherwise we're going to fall behind. So with open source I think it's not just encouraging open source there is a cluster of issues which are all connected, which I think for folks watching this, you should read the action plan and you should listen to President Trump's speech when he announced the action plan.

4:45:20I know you had a very important dinner last night. In fact, three former guests have been on there, Jared, Isaacman, Alex Wang, Shamsanghar. You were there. What came out of that dinner? Well, I would say, so we had this big event yesterday, which is we had this whole AI education task force meeting with the first lady, which I think maybe something we should talk about, which, you know, the first lady and the administration, we really cared about how Americans get the skills to work with AI. So there was a big event. And as a part of that, there was a dinner with several really, really interesting people.

4:46:00And I think, you know, the overwhelming tone, if you look at that dinner, which, by the way, I was not there. I was on a flight to get here. Oh, shit. Yeah. By the way, folks, you know, I had to cancel on you yesterday because I was like, hey, Sean's team, I'm getting pulled into something. I'm going to have it moved to the next day. And then I was like, see, Sean, like, this was why I had to move to the next day. You know, you can, I'm not just making up an excuse. I'm not being a jerk. But if you look at the dinner, the overwhelming message from everybody was like, we just love what the administration has done in cutting red tape, in being pro-innovation under the leadership of President Trump.

4:46:42So you kind of see that overwhelming message, but it is fun and entertaining in a way only President Trump can be. I saw that, I could be off on this. Did Tim Cook say that he was going to put$600 billion into U.S. manufacturing? I think that was Zuckerberg. Was that Zuckerberg? I think so. Okay. I think there was definitely a big number from Apple, but I'm not sure whether it was yesterday. I think that was Mark Zuckerberg. I could be wrong, right? Like, I was on a very late plane here, so I wasn't watching the videos. But I think it was Zuck who did that. But I think you're seeing this across the board.

4:47:17But with President Trump and his administration, there's one very key message, which is you need to invest and build in America. Like you see this message through and through and through. So there are so many examples. You had TSMC come in and talk about their investments here. You had Tim Cook go to the Oval and talk about investing here. Zuckerberg yesterday. day, multiple, multiple companies basically talking about how they are investing tens of billions of dollars in infrastructure all over America. I think this is a key theme, which ties into some of the pieces we talk about. The reason to do this is not just a chip sovereignty, but it's the idea of we want to bring back investment and jobs and infrastructure in America.

4:48:06But there are so many examples from the tech world, not the tech world, of people announcing investments in the US. Yeah. Sounds - I think there's a second comment. I think after the event, I think the president looked at Mark and he said, Mark, this is your start of your career in politics. And Mark was like, I don't know about that. It was interesting to see Elon didn't have a seat at that table. I think he said he was, I don't know. I think he said he was invited. I think he sent somebody else. I don't know. I wonder why. But, well, Sri Ram, this was a fascinating interview. And I think, I mean, I'm excited for you and excited for the country.

4:48:50And I love your plan, everything that you just talked about here. And, man, I got to hand it to you. This is, you're probably, not probably, you are in one of the, if not the most important role in that right now. I think that this is a lot more serious of a race than most people even realize with the race with China. And so I just commend you for rising up and taking the opportunity. And we're all rooting for you. Thank you. Thank you for having me such an honor. And look, this is such a privilege. It's such a dream to just even have this opportunity to be a part of the administration, to work for everyone watching this video.

4:49:37For me, I just think every day about the time I have here, how do I make sure one, we're winning against China, and two, how do I make sure that AI is working for every single American? I keep going back to we need to make AI work for every individual. I think about my dad, how will he use AI and how will he spend more time with his family? How will it help him with his job and all of the current versions of my father? and I think about both those questions a lot. And I also want to say, you know, I just want to hear from everybody. Part of the reason I wanted to do this is because I think it's super important just be like very, very transparent.

4:50:17So if folks have thoughts and views, hit me up and like, you know, let me know. I want to, we are open for business. We want to hear from everybody. But most of all, thank you. I've been a fan for a long time. I've been such an admirer of what you've done. I can't tell people enough of how much being in this room, seeing the story on these walls is a testament to the space you have created. You deserve all the success and you're just getting started. And thank you. This means a lot for me to be here. Thank you. That means a lot. Cheers. Thank you.

4:51:01Most people don't realize how much their personal information is being bought and sold every day. Data brokers are making billions, pulling details about you from public records and the internet, and then packaging and selling it, usually without your consent. That's how your information lands in the hands of scammers, spammers, even stalkers. It's why you get endless robocalls and why ads seem to follow you everywhere. That's where Aura comes in. Aura actively removes your data from broker sites and keeps it off. They also instantly alert you if your information shows up in a breach or on the dark web.

4:51:32But Aura goes beyond data protection. With one app, you get a VPN, antivirus, password manager, spam call protection, dark web monitoring, and even up to$5 million in identity theft insurance. all backed by 24-7 U.S.-based fraud support. Other companies might sell just credit monitoring or just a VPN. Aura gives you all of it, together, at the same price competitors charge for just one service. Start your free trial today at aura.com slash remove. Protect yourself now at aura.com slash remove.