Uncapped #1 | Shaun Maguire from Sequoia

[00:00] All right, Sean, thanks for doing that. I'm really excited to talk to you today. Thrilled to be here, Jack. What's going on, man? All right. So I want to go right into it. So the first thing I want to talk about is sort of like the current state of the media and social media. I'm not sure if I have no opinions on this topic. We can move on. I'm just kidding. I don't know if I have the timing exactly right, but the way I experienced sort of you on Twitter was October 7th happened and you got super vocal and you started sharing a lot of opinions on X [00:30] ones that I agreed with, but ones that many people were sort of afraid to say in those ways. I wanted to start there of just like, what happened in that moment? And did I read that right, that there was some change in you or the way you thought about how you were going to show up around that moment? For sure. You nailed it. It's a pretty long backstory on this. I'll try to keep it brief. But [00:50] I'm a child of the internet, born in 1985. First time using the internet was fourth grade at my friend Nick Palchikoff's house, AOL. We went in AOL chat rooms. We realized that we could impersonate being a professional baseball player, Mike Piazza. If you're listening, Mike, I've never met you, but I did a great job pretending I was you in fourth grade. You know this too, but there was just really something about the internet in those days. [01:14] point so it was very new and the social elements of it were brand new and this is a weird tangent but like in physics [01:21] when you study some system, you start with very simple toy systems that you can actually understand. You study... [01:27] a single ball on a spring or a single ball on a string. Or you study with some simple system and then you try to build up from there.

[01:36] When you understand a simple system deeply, it makes it much easier to understand the complex systems. I feel like I got lucky with when I was born where I was... [01:43] on the internet in the very beginning of the social element. And so I think I was able to understand as a participant, like how... [01:51] manipulation works, how misinformation works, how it spreads. And I'll give you an example on Mike Biotts, and I've talked about this maybe twice before, but my friend and I, we would A-B test. Can we convince people as fourth graders that we're a professional baseball player? And we basically learned a tactic accidentally that [02:11] worked almost every time, call it 80% of the time. You know, if you just come out in a forum and say, I'm Mike Piazza, no one's going to believe you. [02:19] But if you join some chat room and you're talking to people and you seem normal and oftentimes they at the time would then go into a sidebar, you'd be talking to someone one on one. You know, after talking to someone for a little bit about random stuff, you know, the weather. [02:34] whatever their hobbies are, you know, they'd ask, what do you do? [02:37] And if you're not the one that starts, you know, with that and they ask you, what do you do? And be like, oh, I play baseball. And they'd be like, what do you mean you play baseball? Like, you know, like I get paid to play baseball. And they're like, so you're a professional baseball player? It's like, I guess so. It's like, who do you play for? Play for the Dodgers. Like, you're lying. Like, what's your name? Like, my name is Mike Piazza. Like, shut up, you're lying. You know, and they'd ask some question about you. And you'd be like, you know, you have to know all the trivia on the person. And so then you answer some basic questions.

[03:07] I basically lead with, look, I know it's hard to believe, but it's really hard for me to meet people in the real world. I like to come on AOL and just talk to people. You know, like no one's judging me of this. And that would click with people at least in 1995. Yeah. Or sorry, call it 1998, something like that. It was actually shocking where if you kind of did it that way, at least in that time. Today, that wouldn't work. But back then it worked. So anyways, I really grew up in the early days of the Internet. I started building websites. I built a lot of prank websites. [03:37] Thank you. [03:37] and just gain intuition for [03:39] how information spreads, what would go viral, all of this. And I basically followed... [03:45] information warfare from an early age. And then I actually worked in the U.S. government doing some of that work. But I was basically quiet online. That's what I was going to say. So you figured this out when you were like 12 or 13. And then, you know, 20 years go by. Well, you just know this and you're probably learning more. Then you started exercising it. Well, I did some things. [04:06] anonymously. That some very high impact. You're one of the anonymous accounts? I was running some anonymous accounts. I've had some. And I did some very high impact stuff. And so I was still practicing a little. It was just something like that I understood, but was not taking that seriously. Why didn't you do it? Do you want to and you were held back? Or were you just like not interested in doing it? It's a great question. Um, [04:28] In some ways, I was being held back and not by any one employer, not by that, but it's [04:34] I mean, can't really one thing is...

[04:37] before you're 18 and after the rules are very different it's ill-advised to take risk with some of these things well i mean i think we all kind of hold back a little bit you know like speaking for myself like you know i wish i could be a little bit more fearless online than i am i think a lot of people are and then i think you know people see what you do and there's a you know there's like a sort of a jealousy admiration thing of like i wish i could kind of do that and so i think most of us feel whether it's from our employer our friends or you know we're just afraid of [05:07] I think the default is people feel hugged back from going all the way open. Yeah, look, I felt that way for a while. And when October 7th happened, it was a moment in my life where, for a variety of reasons, I felt like I... [05:23] have to speak. And there's a lot to that. One of the things is... [05:28] I had followed basically every... [05:31] conflict in Israel in my lifetime, you know, and in the region, and had learned some of the [05:38] information warfare tactics from prior flare-ups. So that was one thing is I felt like I understood what tactics were going to be used in the future and so therefore had something to offer the conversation. Another was I was just to be very blunt like I was at a point in my life where I had sold a company for a billion dollars and I didn't obviously make a billion dollars but I had made enough money to where if I got fired and [06:02] I wasn't going to starve to death. That's a very lucky, privileged position. But you were unafraid. I was unafraid. You get to a point where it's okay if you...

[06:12] you know, [06:13] Thank you. [06:14] get fired. And so I was very lucky to be in that point. So it's like, I felt like I had a lot of context and stuff to say. I was at a point in my life where I was willing to face any consequences. As crazy as it sounds, like even death. Really? Really. You actually felt if I get killed, that's okay? Yeah. Wow. Why? How did you get to that? Because that's such an uncommon place to [06:34] to get to. When I embarked on this endeavor and I've [06:37] quieted down on the Israel front some in the last six months. [06:42] And I'll get to death in one second, but just one basic philosophy I have with information warfare is basically the first two weeks are by far the most important. There's basically three phase transitions, the first two weeks, then the next six weeks, so first two months, and then basically everything after. And that first phase, you have... [06:59] 90 plus percent of the world's attention. Almost everyone is... [07:03] you know reading the news trying to understand what's happening then the next six weeks probably 80% of [07:09] People drop off. [07:11] other than the people that have vested interests, like people that are Palestinian or Jewish or... [07:18] live in the region or whatever they're still going to be paying close attention but a lot of like [07:22] And random people that work in, you know, Fortune 500 company, 80% are on to the next thing. But some still pay attention, especially a lot of students still pay attention for those next six weeks. And then you get a big drop off and you're basically left from there just with either the people that have a vested interest in the conflict or people that... [07:42] our students to where it's like their job to be learning right now or people that have unlimited free time or who in those first two months became obsessed with this conflict to where now it's part of their identity. And so my philosophy is like you want to put.

[07:55] I put 80% of my time into those first two weeks and then, you know, a lot of time in the next six weeks. That's on any topic. On any topic. This is like a universal truth of information warfare. Do you need to be a prepared mind on the subject for that to work? 100%. Like you have to be willing. You have to basically already know what you're going to say the second that the thing happens. Because you can't get up to speed fast enough, especially if you're on the weaker side, like the side with fewer people. [08:25] that because... [08:26] if you get a single talking point wrong, you're going to get piled on and basically lose your credibility very early on. If you can't anticipate what's coming next, then you'll also get [08:36] kind of embarrassed. And if you also, if you don't already know what's interesting to say, or what's important context, and it won't get attention, especially when you're on the weaker side. And look, the reality of this one is, you know, there's a hundred... [08:48] times more Muslims than there are Jews and [08:51] you know, [08:52] Obviously, Muslims are not a unified... [08:55] point of view, but the vast majority of them side with Palestinians. And it creates like a giant asymmetry in terms of the support on this issue and the amplification of information on the literally thinking through like I'm willing to die on this topic. [09:12] There are many people that have been killed by Islamic radicals in the last... [09:18] few decades and you know you have there's a certain line like i'm not even going to say it but if you criticize a certain person you're

[09:28] almost certainly... [09:29] going to get killed. And you have things like the Charlie Hebdo cartoon in France, or, you know, there's like, and that's, that's a line I would never, ever go to. And I don't think is, [09:39] I wouldn't feel right going there. [09:43] But... [09:44] When you say things like there were certain things I was willing to say that, [09:47] were a line below that did lead to lots of death threats. Like, I've gotten... [09:52] many dozens of credible death threats in the last 18 months. And the... [09:58] The line that I felt was... [10:00] Very important. I think the two lines that I was willing to go up to or a little over are one, and I'll say here, but... [10:10] Like, [10:11] If you mention... [10:13] children, like you get, it's just such a sensitive topic. And the reality of this conflict is that like child soldiers are used, like the UN is not on the side of Israel, but the UN [10:27] The UN's own data says that Hamas has used child soldiers as young as 11 years old for suicide bombings and the official recruitment age of Hamas, according to the UN, at least before this conflict, because they're now starting to scrub their data from the past. But before October 7th, the official, according to the UN, the official recruitment age for Hamas was 15. I think it's important to. [10:47] be willing to say those things to people who have context that when they read, you know, like a 16 year old was shot. [10:52] Um, [10:53] It can be an innocent 16-year-old, and in that case, it's an absolute tragedy, and we should all be very sad. But it could also be a 16-year-old that's armed or on the front lines, and that's just something that has to be thought about differently. And there's a different—it's not always a Western standard that's being applied in some of these conflicts. And I think a lot of people lose sight of that. In some ways, I feel like these issues, maybe to your point, are just so tremendously complicated that for most people—

[11:20] in the West, we don't even have the frame to start thinking about these things. And so that's probably part of why people are silent on all these issues is they have a certain feeling, but they they know that they don't have the prepared mind, they know they're not ready for the next argument. And so a lot of people just kind of, you know, if they kind of feel something, they'll sort of clam up, perhaps. [11:41] Look, there is this fundamental asymmetry, which is because... [11:46] the pro-Azul side is so outnumbered in the information warfare. And it's not just... [11:51] civilians. I think people don't understand how much of a nation-state involvement there is here. And so on the simplest level, there's Al Jazeera. Al Jazeera is... [12:01] you know, a mouthpiece of [12:03] Qatari government, they have strong Muslim Brotherhood ties. [12:07] Al Jazeera is... [12:09] Another lesson of information warfare that I've picked up over the last few decades is... [12:16] The most sophisticated thing [12:17] Actors, they do something smart. [12:20] which is, so take Al Jazeera, Russia today, [12:23] RT. [12:25] They're the most accurate source of news you can possibly find on 95% of topics. If they don't care about the topic, it is literally the most accurate source of news you can find because that way they get a lot of credibility and someone will watch it and be like, wow, this is... [12:39] really amazing reporting. Like it's more nuanced, better than the BBC, better than CNN. But then they leverage that credibility on the 5% that they actually care about where they want to present this information to you. And it's not only Israel, like one of the topics for Al Jazeera that's heavily...

[12:56] misinformationized is India. [12:59] Um, [13:00] And so anyways, when you have this lens that they're trying to gain credibility to launder it on the things they care about, [13:07] Anyways, it just changes how you... [13:11] perceive all these things [13:13] And so in this... [13:14] information battle. [13:15] there's population asymmetry, but there's also [13:18] 55 Muslim-majority countries, probably five of whom have spy services that are literally for decades been gaining information credibility to try to use against Israel. And so you're going against that. [13:32] You really have to have your facts straight because if you get anything wrong, you're gonna get absolutely destroyed. As somebody who used to trust the media, like when I was younger, I was just like, "Yeah, it's the media." And I, you know, now I trust it very little. I think that's like, you know, common. It's not that I think everything's wrong, but it's this 95/5 thing. It's like, I don't know which thing I don't trust. And then once that happens, it's very hard. But so then we're in this kind of like Fog of War-y moment where [13:56] You know, we don't have these media outlets that we fully trust. You've got this like citizen journalist thing, which I think is good. It's good, but it's also flawed. Yeah, exactly. You don't exactly know who to trust there. And, you know, people have the same kind of flaws. And somebody who's very trustworthy on one topic might be completely untrustworthy on another and all this stuff. And so it does feel to me like we're in a moment in time where truth is hard to come by. I'm hoping at least that this is like a transitory moment and there's some resolution here. [14:26] see your read on the situation and like where does this evolve into because it feels in flux or something. I couldn't agree more. I think we're at

[14:34] basically a low point in the truth in our lifetimes. And hopefully we come out the other side [14:41] in a stronger place. But I think there's like at least two really important factors here, if not three. And just to calm my eyes, I think it's important to know what you're dealing with. [14:53] One is I think that we basically have had a... [14:56] phase transition in Western media in the West that I think really started in the late 90s. And candidly, I think Fox News, like, defected in Prisoner's Dilemma. [15:07] And they realized that they can start putting out, you know, highly editorialized things. It's a good way to put it about the Prisoner's Alignment. It was literally Prisoner's Alignment. I think until like through the early 90s, I think both sides were for the most part, you know, trying to do their best. But Fox News defected. In my opinion, they were one of the first. Once they defected, CNN defected and then everyone defected. And that was happening at a similar time that, you know, [15:32] before social media, but then a decade later, social media came and then that changed the incentives where it's not worth it. You should go for the headline. You got to go for the headline. So that created a second incentive, which, you know, just in this negative feedback loop, we defected in prisoners dilemma. And then once we were already out in the lower quadrant, [15:50] we were incentivized to basically put out fake news and, [15:54] tell stories very quickly. So this is like one big... [15:58] The second thing that I think not enough people really understand is just how much nation-state involvement and

[16:06] like intelligence operations are involved in the misinformation on all sides, including the U.S., like massively. You're saying like the U.S. is influencing like the media outlets. Everywhere. Everywhere. And but so is China. So is Iran. So is Russia. All nation states are doing this and have been doing this for a long time. But I think it's gotten much more extreme recently. And just a couple of points on this. If you go back to the 2016 election, there was there were all these fake [16:36] People don't ever stop doing what they were doing. They just get better at hiding it. They get more sophisticated. And Russia was not sophisticated enough at the fake news in 2016 to where we were actually able to catch them. It was relatively obvious as an American citizen. But they got more and more sophisticated over time. And now it's at the point that you can't even... [16:55] tell what's happening. And it's not just Russia, it's basically everyone. But I mean, the US has [17:01] forever as well. Even during the Cold War, we would start Radio Free Europe and start radio stations to put out pro-democracy content on the airwaves, which... [17:13] My interpretation of it is that's great, but everyone does this. [17:16] Living in America, greatest country in the world. I am a maximum pro-America person. [17:22] Patreon. [17:24] Thank you. [17:25] But freedom of speech is... [17:26] Um, [17:28] is kind of a root vulnerability for America in that our adversaries get to use freedom of speech as well. [17:34] And we don't get to use that. We don't get to go into China or Russia and, you know, say whatever we want. Our news platforms, everything gets shut down there, but they get to use it here. And so we're in this maelstrom right now of the news defected.

[17:49] in the 90s and fully deteriorated. Social platforms changed the incentives, you know, to go more for clickbait, have very fast responses. And our nation state adversaries got more sophisticated. And especially in an era of social media and all that, it was easier to, there's less of a check. Like New York Times in the 60s, 70s, 80s, they knew that Russia was trying to infiltrate their newsrooms. And so they were actively monitoring for it. So just like all those [18:19] It's just an extreme... [18:21] soup of [18:23] lies and fake news and misincentives. And I think the best thing that anyone could do is just really assume that everything you're consuming is wrong or a lie. And I do think there's some optimism that we can improve these things over the next- [18:37] few years. I think we're close to rock bottom right now, but people at least have to have this awareness. [18:43] One thread I'm curious to pull on there is... [18:46] When you think about decades ago, [18:49] And what the U.S. was like in the world then and us sort of, you know, maybe slightly paternalistically or very paternalistically sort of like trying to, you know, push democracy and Western ideals through the world. But we were like sort of like a hegemonic power at the time. And now it's a little bit different.

[19:19] us that's kind of how on some level you know when i think about you know russia ukraine when i think about like all these dynamics a lot of what's at stake here is u.s sort of pushing for its own ideals in the world where in the 50s or 60s i would have been like of course the u.s would do that and now it's like less clear when there's more in the balance and so what's like the optimal strategy there so i guess i'm just maybe quickly your read on that dynamic and is that such a good question um [19:47] And obviously a big and tough question. I have to call out that I'm coming from an extremely strong pro-America perspective. I love America. I think it's the greatest country in the world. But every country has... [19:59] flaws, and I think we have to be realistic about where we are today. [20:05] You know, next point. Do you ever play poker? Yeah. When you play poker, you have to adapt your strategy based on your chip size. Like when you're the big stack and it's like it's only two people left at the table and you have... [20:18] 100x the chips of the other guy, you should just make them go all in every hand basically because the probability is on your side and you'll almost certainly beat them. Whereas... [20:29] - Small you gotta pick your spot. - When you're small you gotta pick your spot, or when there's 10 people at the table, and everyone's relatively, [20:36] similar chip size. You have to adapt your strategy based on the chip size people have. And you use the word hegemonic. America, 20 years ago, 30 years ago, we just had such a big chip

[20:50] play a very different game. Or if you go to the Cold War after World War II, like after World War II... [20:56] Europe had been [20:58] pretty decimated. It was really like a bipolar world at the time. So the strategy when it's U.S.-Russia and both are U.S.-Soviet Union, like, use a different strategy there than where we're in right now, which is it's a [21:15] sadly eroded our chip sack over the last 20 years through many things through [21:21] Iraq War, through sending our debt load to $35 trillion, through... [21:26] outsourcing manufacturing to China specifically, Europe, in my opinion, being in a relatively weaker position than they were in 20 years ago. And as our top ally, she's like, we have to adapt to where the world is right now. And sadly, I don't think America is... [21:44] strong enough right now to fight every battle. So we have to pick our battles. We have to pick our spots, like to use that poker analogy. So do you think the right like 30-year strategy is to like... [21:56] hunker down, build the chip stack back, and then go back to aspirationally getting back to that kind of place. [22:02] The right strategy for the next 30 years is a lot more... [22:06] hunkering down than it was the last 30 years. I think we made a big mistake of being too loose with our chips the last 30 years. I don't think the strategy is to go all the way. I think it's strongly in our interest to still choose some strategic spots where we exert a lot of leverage around the world. But it is to pull back some. But I don't think we should ever go back to that

[22:30] like as hegemonic as we were. Every battle's our battle kind of thing. Every battle's our battle. I think that's the wrong strategy. And I think it cost us a lot. [22:38] If nothing else, I think that [22:41] That is [22:42] approach didn't properly account for the second order effects. For example, Iraq War, first order effects are, you know, [22:51] There's a lot of first order effects, but, you know, ran up our deficit, depleted some of our arsenal. Um... [22:59] Some of the secondary effects is it just created a lot of resentment for America in the world, all over the place. And that weakening of soft power... [23:07] has all sorts of ramifications around access to resources, around, you know, like, Europeans voting, like them having resentment that they sent troops, you know, to support us in foreign wars and, you know, and not voting with us in UN. There's just all these things. And I don't think we've done a very good job of keeping track of the second order effects. And so, look, I do think we need to have [23:29] much more of an inward focus the next... [23:32] two decades, three decades than we've had before. Yeah, I guess it's just tricky because as you unwind that [23:38] poker position, your opponents at the table start to see that you're not going to push, and then that changes their behavior, too. 100%. They see that if you're not going to, you know, if not every battle is your battle, well, like, maybe this battle is not going to be their battle either, and I'm going to go maybe do something over here. Yep. But an expert poker player would, you know, it's like... [23:55] We played a couple of hands. We lost a couple of hands. Maybe it was optimal strategy to play the hands. We got unlucky or whatever. Now that everyone can see the chip sack, like it's,

[24:03] Right. [24:04] It's visible. It's a... [24:07] And so I think everyone can see that it's rational. But when you play poker, you have to bluff sometimes. Like you can't just only – the optimal strategy is to still bluff. You can't only play perfectly rational because then – Yes, you still have to surprise your opponent. You still have to surprise your opponent. And so like the next few decades, I think we should still be doing some very surprising things. Let's talk about U.S. politics a little bit and sort of what sort of – I guess specifically – [24:35] the switch that tech has gone through. And I actually want to draw a parallel between, you know, from... [24:41] the tech side, like inside companies, you know, but Brian Armstrong at Coinbase did where he was like, I can't remember exactly what, you know, when he said this, but, you know, at one point he was like, it was roughly May 2020. Yeah. And he was like, I know, because I had been with him like a week or two before. Yeah. He's such a legend. And when he did that, which now, you know, be like, you know, our, we're focusing on our customers and our products. That is a very bland [25:11] So that transition happened where, you know, he went out on what seemed like a limb, but now it's like, of course, he was right. And like everybody remembers that at the same time. [25:19] Tech sort of politically went through this big transition where Trump in 2016 and 20 was like, you know, it was like a pretty heretical thing. Like people were, you know, very quiet about if they were supporting him and most weren't. I was not. [25:32] For the record. Yeah. And now it's, you know, it's it it is definitely not heretical. And it's also much more mixed in terms of the way people are falling politically. So I guess those two things seem to have something in common, of course. But like, what is that? What is the underlying thing that's happened so that tech, I think, has like swung as sharply as it has in this sort of like deeper way than either just like how it's voting or how it's behaving at companies? I just need to say one thing about

[25:57] my own personal journey that then I'll tie it into the broader tech journey. But [26:04] I came on, and in the beginning of this, I'm saying how I'm some misinformation expert and information warfare expert. And I honestly... [26:11] I think I am, and I have a pretty long track record of... [26:13] public and non-public stuff there. But I got fully manipulated in 2016. To be very candid, where it came from is I was too trusting of the U.S. intelligence community. And I had worked... [26:27] in and with [26:28] the USIC. And I think there's many unbelievably great people, some of the greatest heroes in America and the world are in that community. But I was raised, like a lot of people in my family were in the military, and I was just like extremely trusting of that community. [26:44] And my root vulnerability in 2016 and 2016 [26:47] 2016 to 2020 was I deeply trusted that community. And so things like, you know, [26:52] you know, [26:53] when people in the IC said that Trump has ties to Russia in 2016. Like that, for me, worst possible scenario would be we'd have a Manchurian candidate, president that is owned by Russia and somehow skates through. And then in 2020, I wrote about this a bunch, but I still didn't vote for Trump there, even though I had really... [27:15] like his policies in his first term, right two weeks before the election, 51 Intel officials, five former CIA directors came out and, you know, said that the Hunter Biden laptop was Russian disinformation. And that sowed enough doubt in my mind that, wow, maybe if Russia wants this guy to,

[27:34] to win. [27:35] then like I want whoever... [27:38] Russia... [27:39] doesn't want to win to win. And [27:43] That really worked on me. I think everyone had different vulnerabilities, different things that they, you know, data sources that they waited. But for me, two things coincided. [27:53] which were, one, I learned that I had just been deeply manipulated, and that was infuriating to me. And the second thing was Silicon Valley... [28:03] Starting around 2016 or so, [28:06] The Trump administration... [28:08] And I think that became such a so much animosity and people got more political and more vocal against Trump. [28:15] It coincided with the DEI movement, James Damore at Google, Me Too movement, cancel culture. [28:23] I think a lot of people in tech simultaneously got so frustrated by... [28:28] like cancel culture in the Valley. [28:31] at the same time where people were realizing that they had been manipulated, [28:34] for a long time. And those two things came together and it was just like, [28:40] Fuck it. I'm pissed. Like... [28:43] I'm going to start speaking my mind. And once some people started to speak up, it's just like every time one person says something, it gets easier for the next person. The next person is so balls. And that is something like from an information theory perspective, you can codify. [28:57] That was something that was... [28:59] I was actually going through my mind was like that you can trigger an information cascade with like small, small change in bits. Yeah, it goes back to sort of the prisoner's dilemma thing where once you see other behaviors like that, it just gets much, much safer. I also really like the root vulnerability concept. I think that is like, I haven't heard it quite like that. But I think that's something that like all positions have that and understanding your own weakness and where you can get trapped is probably really essential.

[29:29] from the left to the right. What's interesting is I think we like historically weren't that deeply embedded in the government. And now we actually are in what looks like a much more obvious way to me. And you see like a bunch of players inside of tech having much more influence with the government. The government seems to care about tech in a way that it never did. People get invited to the White House. They're like, you know, it seems like they've got a voice in all of these places. It's ironic that this happened with Republicans, not Democrats, for what it's right. It's crazy. And it seems like that was like a real fumble from the Democrats. [29:59] They had that at their disposal this whole time, but here we are. [30:05] One of the things I'm wondering about is will... [30:07] Will this drastically change the types of companies that become important? And so I'm thinking about the types of companies that, you know, maybe this got most, like, accurately, memetically captured by, like, Andreessen's, you know, American Dynamism Fund. You know, there's all these companies that require the government to do government-scale things. And, you know, you see this with Anduril and SpaceX and OpenAI and, like, many others of, you know, what are kind of, like, the biggest, most important companies at the moment – [30:35] you know, coincidentally have these big government relationships. And so I'm curious, will this, will this sort of tech government hybrid, will this change companies or will this just make it a little easier to navigate? Like what's the impact for tech here? There's a lot of directions to take this. One, [30:50] Initial quick thing. [30:52] In some sense, it's challenging and risky to have tech be closer to Washington. You know, like whenever you take a side, you're more at risk if something goes wrong. The conflicts of interest are more acute and...

[31:06] So, [31:07] I hope that, like, I think that [31:10] we should have a deeper relationship between tech and Washington, and Washington should be funding more ambitious, [31:16] tech projects. And I think a lot of times... [31:20] like the government investments into tech or loans, whatever, have been... [31:26] I think, treated unfairly. They're seen as a conflict of interest, whereas... [31:31] Oftentimes these programs were, you know, there's a broad agency announcement and a lot of people bid on it and the tech company happened to win it or the loan was given to 10 companies, nine were incumbents. One is a tech company, but the media focuses on the tech company. [31:46] So look, I think we should be doing a lot more and I hope we don't jeopardize that, but [31:52] The even bigger thing... [31:54] to me, and I haven't said this too much publicly, I wrote a manifesto for Sequoia a few years ago. I think it was three years ago. It's a hardware manifesto. [32:03] I've been obsessed with Harvard my entire life. You know, I have a, [32:06] teaching physics and... [32:08] literally was obsessed with semiconductors as a kid. I completely obsessed, obsessed with the chemicals industry. And I was also obsessed with early Silicon Valley. And if you look at Sequoia, [32:19] Sequoia made almost all of its money the first 25 years on hardware investments. [32:25] you know, [32:26] one of the first investors in Apple and, and, [32:29] Cisco and [32:32] you know, linear and many, many other companies.

[32:37] Nvidia, first investor in Nvidia. The last 25 years, Sequoia has made almost all of its money in software. [32:44] And kind of a question I asked myself was, like, is hardware dead, long live hardware? [32:51] Is this kind of a secular trend? [32:55] Secular, you know, in a finance perspective, especially like Wall Street oftentimes means like longer trends than just one economic cycle. And so I wrote this manifesto three years ago. [33:07] Now, it's obvious to everyone, but literally in my manifesto, I had two stocks. One was NVIDIA, one was Tesla. And I really strongly believe... [33:17] we were just in like a secular... [33:19] period where it was all about software and not about hardware. And some of my realizations, there's a bunch of things, but one... [33:27] that I think is not... [33:28] Fully appreciate it. [33:30] almost by definition to [33:33] Like every software revolution is preceded by a hardware revolution. You can't have some new software platform if you don't have the hardware to run it. And for example, you can't have... [33:46] Uber and DoorDash if you don't have the iPhone. [33:50] Well, you need to have the App Store and the App Store needs to have the iPhone and the iPhone required 20 years of [33:57] development with Qualcomm, Broadcom, [34:00] you know, [34:01] Fiber, CapEx, all sorts of things, Cisco. But there was 20 years of hardware before that.

[34:08] you know, deep learning, you need to have the GPU to have, [34:13] The cloud, you needed 20 plus years of cheap technology. [34:17] CPUs getting really cheap. Kind of the whole point of the cloud is that we got to a point where hardware got cheap enough that you can put a ton of [34:28] relatively commodity hardware in one place and operate and you get a like an economic advantage to to do it that way. Anyways, you can name any software evolution history and is preceded by a hardware evolution, something like VR today. Part of why we don't have great VR software is the hardware is not good enough and is getting better very quickly. [34:50] And the input to VR hardware... [34:52] Like a lot of it, it's byproduct of what's been happening with the iPhone. Like a lot of the [34:57] I think sensors in general are one of the most underrated technologies in the world. My estimate is about a trillion dollars of market cap of companies that are basically [35:06] sensor companies are selling sensors. A lot of the market cap of Honeywell or Raytheon are from sensors, but there's many other companies. Keyence is a $100 billion market cap company. Anyways, [35:18] is [35:19] Hardware dad or... [35:21] or [35:23] Are we missing something? And I really think we've been missing something. I think that thing is that... [35:27] Basically, the way Morsela happened... [35:30] You know, hardware matured in a bunch of different areas all around the same time, which led to a bunch of software ecosystems all catching on. And now as you go to and now a lot of those have run out, like the app store is very few new interesting apps getting created. VR is not.

[35:44] quite here because the hardware has to get better. In my opinion, AI is still hardware limited. You see it with the AI data center build-outs, but [35:52] it's not just going to be NVIDIA that benefits. There's going to be a bunch of other companies on the hardware level that benefit. We're seeing the beginning of autonomous vehicles, just on and on and on. I think we're at the beginning of – [36:04] a new hardware revolution that, [36:06] Silicon electronics are starting to become silicon photonics. We've just kind of gotten the first... [36:11] commercial applications of silicon photonics that are actually economically viable. And anyways, I think we're entering a new era where the next 25 years through robotics and AI and space and technology. [36:24] and [36:25] And then some of the later platforms like... [36:28] VR and then bring computer interfaces and medical technology, all these things. I think we're entering another golden era, and obviously defense technology to a massive extent. I think we're entering another era where that's going to be a very money-making opportunity. And that requires... [36:45] government partnership, or at least it's enhanced. And if we want to compete with China... Because it's such big dollars or something else? Because it's big dollars. Also because a lot... Look, the biggest driver... [36:56] Historically, as an early adopter of hardware, has been the defense [37:00] department. I don't think that's something that [37:03] should be ignored. And I think one of the other biggest trends in the world right now, it's literally one of the biggest macro trends is... [37:10] For 25 years, we were outsourcing manufacturing to China. Now we're bringing it back. And that re-onshoring is just an insane driver also of hardware development in America. So if you think about this as it applies to investors, Sequoia did all this investing in these crazy companies a long time ago. Over the last 20 years, software, not to do another Andreessen thing, but software's eating the world. That's obviously what dominates. Software, Andreessen, amazing at marketing, obviously. Very good. And that was right.

[37:40] And, you know, that's kind of what's played out. And I think, you know, outside of a couple names, which are incredibly important asterisks, you know, like Anduril and SpaceX and, you know, some other companies that are very hard to... [37:53] hard tech problems and companies. I think most investors who invested in, you know, hard tech over the last 15 years, maybe let's say, like just like mostly didn't go well. I don't know what the overall returns is. If you, you know, if you, if you include these like outlier companies, I don't know, like maybe actually with the long tail, it's actually very good still. But my question is, was it, was it a bad time and now it's a good time? Or is your mental model that hardware companies [38:23] It is very good to invest in this kind of stuff if you are capable of doing it. But if you're bad at it, you are like more likely to lose your money than average. You know, like another way to say this would be, I think, the sort of prevailing wisdom, which you might agree with, you might disagree with. But, you know, would be like a smart enough, reasonable in the mix San Francisco investor who runs around making investments in a lot of sort of good, you know, what we would think of as, you know, sort of classic AI software companies right now will probably get a hit. [38:53] probably do fine. But, you know, going and investing in hardware companies, like unless you hit a big one, you're going to like you might get zeros all around. So like, how do you think about is it about the time or is it about not everybody's equipped to do this type of investing? [39:05] I think all of the things you just said are true. I think it was a bad time to be investing in hardware.

[39:12] on average the last 25 years. [39:15] With the caveat that I think simultaneously some of the best investments [39:20] of the last 20 years were hardware companies, SpaceX, Tesla. I'm of the camp that Interl is going to be an absolutely giant company. Me too. But I think on a relative basis, I think the amount of market cap [39:33] that will have been created retroactively by hardware companies in the last 25 years will be very little compared to what it was for the 25 years before and what I think it will be the next 25 years. I believe it was a bad time. [39:44] I also think that most VCs were not equipped to invest in hardware, but that's okay. It's like the industry evolved for that because it was a bad, like just bad macro time. Are you saying that VCs were not in that moment? Or do you think even in a good moment for hardware, most VCs are not equipped to invest? I think a lot of the VCs that were good the last 25 years are not equipped to invest into a hardware era. It seems like a really different mindset because, you know, to your point, hardware comes first and then the software comes. [40:14] way differently. For sure. Versus when you see this thing's working, now I can just bet on a some application of it. Earlier today, I was reading the [40:22] story of how Kleiner incubated Genentech. And I mean, like, they literally incubated the company. I can't remember the name of the partner that really did the work. But basically, this person had been following some of the top biowork out of [40:38] you know, in the Bay Area, and there was... [40:40] faculty member at Stanford, faculty member at UCSF, where he'd been following their work. They basically approached them to

[40:46] try to commercialize the work and Kleiner convinced them to start a company. [40:50] Brooke Byers, my friend, Blake Byers' dad, was one of the people that was kind of on this team, but it was three people from Kleiner that basically approached them and convinced them to start the company. The idea of like you have a VC that's reading the scientific literature and finds these two professors... [41:08] and thinks they can create a new platform in biotech and then goes and starts to come in with them. And then, you know, you build a company for decades. That's just not how most VCs have been doing the job the last decade. [41:19] few decades. And granted, I think it's still how a lot of people in life science VC have been doing it. But actually, I don't think that's [41:27] where the biggest hardware opportunities are going to be the next 20 years. And so just to use that as a thought experiment, that's just not how... [41:34] VC has been done the last few decades. But I think there's also something very important you said. [41:38] which is, look, I do think the power law... [41:41] curve of hardware is steeper. One of my [41:45] insights around, I think, why. I think there's a fundamental difference between software companies and hardware companies, many, many, many. But one is, [41:53] With a software company, [41:54] If you have a successful first product, [41:57] The probability that you come up with an organic second successful product in your company is very low. It's like almost if you do the conditional probability, like it's. [42:07] is basically the same probability of being successful with the product from the outside. You don't get many advantages once you have a product. There are some distribution advantages, but you don't get much of an advantage for...

[42:19] creating a new [42:20] product, you can acquire things that already work and roll that in, such as how Facebook acquired Instagram or Google acquired YouTube. [42:29] But with hardware companies, [42:31] it's fundamentally different. Like, [42:33] Basically, every hardware company in history that's had a first successful product will have dozens of subsequent successful products. Like name any hardware company, go on their website. They'll probably have dozens, if not hundreds of products. Go look at, you know, Broadcom, Qualcomm. [42:51] Cisco, Danaher, [42:54] Tesla, NVIDIA, and for a lot of reasons. One is... [42:58] One's Apple, like any hardware company. If you're doing it right, I think sometimes you're not taking much market risk. It should be something where if you build it, people want it. But building it is very hard. And the hardest part about building it is oftentimes not designing it. It's like figuring out the supply chain. It's... [43:16] You know, it's it's figuring out the cost at scale, cost at scale, the capital cycles office. Yeah. Like take Apple. [43:23] Because I think it's a simple example. [43:25] They have so much leverage over their supply chain now that after the iPhone, if they want to make the iPad, they have so much leverage and they can go get production capacity from TSMC or whatever. Apple probably is the best sensor team in the world, especially for MEMS-based sensors, like micro-mechanical machines. [43:46] They reuse their sensors across all the different product lines. They use the infrared camera for face ID in the phone and in the iPad and now in the VR goggles. And, you know, and each individual component is a moat in and of itself.

[44:00] And so what happens is I think with these hardware companies, once you're successful, that success just compounds. I think at an even greater rate on average, it's harder to get success with the first product. But once you do, it leads to a lot more. It leads to a lot more. Whereas the software companies. [44:17] The first market you go after is much more important. Yeah, I mean, I think that's mostly right. I mean, what was going through my mind as you said that, as I was starting to think of examples like a Microsoft or a Workday that do become these, you know, we saw what happened with Teams and Slack. [44:33] So it's not to underestimate that distribution advantage. For sure. It's more that the hardware company also has that distribution advantage, and now they have this creation advantage. They also have a supply chain advantage and a creation advantage and just much bigger moats. [44:47] of the world [44:49] I think they're few and far between compared to – [44:53] compared to hardware companies. I challenge you to find a hardware company that doesn't have... [44:58] dozens if not hundreds of successful organically grown products. And it's just because once you've already got all that stuff, it's so much easier for you to make the next one. For sure. [45:08] For sure. It's actually amazing that Apple restrained itself to as few products as it did for as long as it did, given that they could have clearly just proliferated. But like everybody else did. And they're like, we're going to have five. For sure. You have to get to the point where you're really strong before you go start taking on all these other products. But once you're really strong, like with one product, it's going to compound like crazy. As you were talking about, you know, how. [45:30] A lot of investors operated decades ago and, you know, incubating and Genentech and all these things. There's a part of me that as I reflect on my own sort of.

[45:40] journey here where, you know, I've been building, you know, a company up until a year ago. And now I'm, you know, investing, which I really love doing. If there's an insecurity I have, sometimes I'm like, I was, I was doing the work. And now I'm supporting the people doing the work. I don't think that that's like a productive mindset, necessarily. But like that, you know, if I'm on a, if I'm on a day where I'm like, insecure about what I'm doing, that's what the insecurity is. And you also you'll grow out of it. Yeah, hopefully. Yeah, I'll just forget that I did [46:10] You know, you built a company and you were very successful and now you've been an investor for a while. Like, talk me through, you know, will I grow out of it? Like, did you have that? And how do you think about this? And are there ways to do it where you agree with that statement versus not? You'll definitely grow out of it. You know, that said, you may get the itch at some point to start another company. Like, I think they're different phases of life. [46:30] I'm going to use a few weird analogies. I have a lot of weird analogies. One is... [46:35] with mathematicians. So there's this saying, it's not 100% historically accurate, but that a lot of the best [46:43] Almost all of the best mathematical results are from young mathematicians, call it before the age of 30, roughly. And the Fields Medal, it's for work before the age of... You have to be 40 by the time you get it, so the work has to be a few years before that. There's this saying that mathematics is a young person's game. There are some counterexamples, like some incredible mathematical work that people did later in life. But I think for the most part, it's true. And...

[47:12] As a former kind of mathematical physicist, the reality of life is... [47:18] I think he [47:20] There are different phases of life. And I have two boys, little boys is the best thing in the world. But it would be very hard for me to do math the way I used to. [47:31] like with... [47:32] All the responsibilities I have now and having a family and being a parent I want to be. Yep. [47:37] And all this, like it's basically impossible. [47:39] When Gregory Perlman solved the Poincaré conjecture, basically spent years living in Russian forests, you know, just – [47:50] Literally just thinking about this one problem – [47:53] in Max's [47:55] autist way, which I respect a lot, they push humanity forward in the process, or when Andrew Wiles solved Fermat's last theorem, like he, I think it was seven years of work where he was just [48:09] fully, it's the only thing, waking up, breathing this thing. And I think it's hard to [48:15] For those types of problems, I think it's hard to do anything else. You have to put all the concepts into RAM in your head and just keep them there. 100%. And process. If you shut down the computer and purge the memory, the I.O. time is very long, and you'll just never make any progress. I don't want to interrupt, but I'm sure you've read the book, A Mathematician's Apology. The G.H. Hardy book. It's about this great mathematician who reflects back on his books. And Hardy's the one that basically found Ramanujan.

[48:45] And he's sort of like reflecting later on, you know, I did this work when I was at a certain age to do that work. And then, you know, how do you grapple with that over time? And I think different people grapple with that differently. I think for some people there's like a, I think the unhealthy, the grappling I don't want to have with, you know, to bring it to where I think you're going to go with the metaphor of building a company is the negative grappling would be to sort of have those be the glory days. [49:15] I imagine would be to be like, that was a beautiful chapter of life. And now how can I productively make that something I pass on and support people in their, you know, being a 25 year old mathematician days? [49:45] He still does incredible work, but he... [49:47] Then got really involved in coaching the Putnam team at MIT and basically trained a whole generation of young mathematicians. He basically broke the Putnam at MIT. And like that work of mentoring the next generation and then taking on a lot of students and teaching them how to kind of write their first paper and do research. [50:06] real research grade mathematics. That's just a different [50:10] skill, but I think that probably has even more of an impact, like, you know, enabling that whole generation of next people. And then,

[50:19] I think Yuffe is going to have another chapter in his life coming up. [50:23] which another analogy from science, my PhD advisor, John Preskill, [50:28] He's an amazing guy. I haven't seen her talk to John in a while, but he's I'm sure John's gotten shit for my political views recently, given the state of academia. But John, I still love you. You're an amazing guy. I've learned. [50:41] so much from you. But John had a saying where basically every 10 years you should kind of switch subfields of physics. He started his career doing physics. [50:52] particle physics, high energy physics, like... [50:55] had one of the proposals for a dark matter particle candidate. And then he moved into, in the early 90s, kind of the... [51:04] early... [51:06] like string theory days, like holography, how does all that work? [51:10] work. And then in the late 90s, it was one of the first [51:14] kind of [51:16] real theoretical physicist to go into quantum information theory and early quantum computing. And then about 10 years, he went... [51:24] And I mean, he built like the top quantum information group in the world. And then over the last 10 years, he kind of went back. [51:31] to thinking about what happens inside black holes. And going back early in his career, he had won a few bets against Stephen Hawking, and they were some of the early people that thought about, like the black hole information paradox. But he went back to it 30 years later with much deeper tools. I think there's another philosophy is like, I think that every 10 years to... [51:48] switch what you do a little bit,

[51:51] It's very powerful because you then cross-pollinate [51:54] ideas, you basically cross-pollinate [51:58] relationships, you have wisdom from this area. Also things just naturally go stale over some period of time. Yeah. And you force yourself back to some amount of beginner's mind. I think that's what I've appreciated as I got to a place where I was like, I kind of have my ideas about how I should do this as a CEO at this stage and whatever. And you unshackle when you do a new thing and you're like, I have no choice but to go back to thinking like a beginner, which feels healthy. Completely agree. I think... [52:24] Good articulation. I think that was, in a lot of ways, the key point he was making. So you articulated it better than I did. [52:45] more in line with what you think is... [52:49] The word honorable just came to mind. That's not the word I mean exactly. But like that are more fulfilling and meaningful versus types that are more just like, who cares? Even if I'm going to make money on this, it just doesn't care. What am I doing? I shouldn't answer this, but I'm going to. As you started in the beginning, I don't hold back. Look, I generally get way more fulfillment from hardware companies. [53:07] I think that at the end of the day, [53:09] hardware companies are the ones that really move the ball forward on humanity in general more than software. And that's not exactly, it's not always true. But as a general statement, I think that hardware companies are something that

[53:21] Like if one hardware company doesn't happen, the whole category might not happen. Like if Elon didn't push SpaceX. [53:29] Who knows where humanity would be with space today if he didn't push Tesla? Who knows where [53:34] EVs would be if Steve Jobs didn't push Apple. Like, who knows where personal computing would have been. [53:42] And it would have happened, but maybe five, ten years more slowly. [53:47] There's just so many examples of this where I think that an individual hardware company can really change the timeline in which... [53:55] whole industries happen more than software. I think software is more inevitable. Once the macro ingredients come into place, like I think the software companies are... [54:06] almost inevitable. And going back to science, there's this concept in science that ideas are in the air. Calculus was invented simultaneously by [54:14] Newton and Leibniz and other people like all the raw ingredients are there. I think with hardware, [54:20] It requires more of like a stroke of genius. Like great man theory type stuff? Yeah, great man theory. It's like Edison – [54:27] And Tesla were like psychos. And if they didn't exist, you know, like it's hard to know if the Wright brothers, like it's hard to know what would have happened with aviation. So I do believe a little bit more in the great man theory with Tesla. [54:40] this. And so therefore, [54:43] The places where I've gotten the most fulfillment have been... [54:46] like seeing something with hardware earlier than others and being willing to stick with it. Yeah, I mean, the other big difference is like science and math, you know, you might say eventually must get discovered. Like if it's true, eventually someone's going to find out that it's true. Versus with like nobody had to make Apple. Like Apple didn't have to be like Apple. It could have been anything. You know, somebody decided to make it that way. And because Steve did it.

[55:09] with such a specific... [55:12] intent like being like i just watched an amazing clip yesterday of him talking about microsoft and he basically says like [55:22] they had no spirit. And he respects them, they built a great business, but that they had no design sense, no spirit. And that he viewed Apple as a [55:34] cultural movement and that it's, you know, that, for example, [55:39] if he didn't bring in [55:41] type setting like it might not have ever happened and we might have had ugly text forever you know it just not only does it take someone to do it but the person that does it gets to really set [55:53] the agenda for how that field happens. I saw a clip of him the other day talking about how like, this is like a long time ago, decades ago, like 80s or 90s or something. He's talking about how, like, wouldn't it be amazing if you could ask Aristotle questions and how like, you know, and, you know, maybe that's kind of where AI is going now with stuff. And but like to have that kind of vision and narrative, and that's what rallies people to build towards that vision. And I think that's like a big part of what's required for these is somebody has to just like, [56:23] so long to so many people because it's so hard to make this stuff happen. Maybe we can end on this question with that backdrop. And you think about the way that VCs... [56:33] can be operating or should be operating or it's because, you know, what we're talking about right now, like, I'm like, this has me fired up. Like now I'm like, I feel good about being, you know, an investor. Like, this is great. But so much of it is so, it can be so, you know, mimetic and there's so much herd mentality and there's so much because you're, you know, because you're abstracted from the building, unless you're, you know, incubating and building companies yourself, it, you know, seems very easy for people to get caught up in the wrong stuff. And so I guess,

[57:03] what VC should be practiced like today. [57:06] Trey Stevens [57:08] and Marky, I think her last name is Wagner, they... [57:12] wrote this piece a couple years ago called Choose Good Quest. Yes, I loved that. It was amazing. [57:18] And I thought that was a great articulation. [57:21] Of it. [57:21] I'm not going to knock anyone that... [57:24] wants to create a widget, [57:25] If that's what gives you joy, go create a widget. But in my opinion... [57:30] Like, I get my joy from... [57:33] either choosing GoodQuest myself or trying to help other people on GoodQuest. There's probably 20-30% of the companies I've backed are not GoodQuest. [57:42] They're kind of more widget companies. And something just to call out, I also love people. And there are some humans where if I... [57:53] deeply love some founder and they are obsessed with some widget, then sometimes I get excited to give them money to go do that. Also, by the way, widgets sometimes surprise everybody. Exactly. This is like the YC thing that great things often start like a toy, and sometimes it looks like a widget now. It grows into something unbelievable. And so widgets aren't always widgets. Exactly. Exactly. Like the Collison brothers before Stripe did Octomatic stuff. [58:16] which I don't even know exactly what it did, but it was, I think it was like Substack before Substack, and it became Stripe. And potentially even on day one of Stripe, you wouldn't have necessarily had this type of sort of like, like starstruck conversation about what it was they were building, and it evolved into something that turned out to be incredibly important. And I'll just say on that, John and Patrick, they're such curious people. They're just, they're like maximally curious, maximum voracious readers. Yeah.

[58:44] And I understand why they like I can I can imagine sitting in an automatic pitch from them at 20 years old or whatever and just seeing their passion for it and be like, OK, these are brilliant young guys that are I don't care about this widget, but they're obsessed with it. So I'll give them money to go do it. And I think there's something very pure in that. And then it kind of not in that company, not that kept table, but those people and like the kernel of lessons from that. [59:12] company became Stripe, which [59:15] I think is going to revolutionize [59:18] financial technology, it's already well on the way towards that. And so I [59:22] I, to your point, like, you never know where these things are going to go. And that is something investing I do. But I also I really admire people that choose good question in the beginning. Sorry, I keep I have to keep asking. No, keep going. Do you think that outlier people are the limiting ingredient to us having more outlier companies? I don't think it's the only limiting factor. But I think it's [59:43] Definitely one of them. And if you look... [59:46] kind of in any given situation, [59:48] year in history like [59:50] at the number of [59:53] unicorns or decacorns that are formed like the [59:57] In terms of realized exits, I'm not talking... [1:00:01] VC fake marks in capital cycles, I'm talking like actual exited companies, it's relatively consistent for a long time. [1:00:11] And I think that says something around like at any given moment, there's not that many real big new market opportunities. So I think it says something about that. And I think it also says something. It either says only purely the number of market opportunities or it says only something about the number of outlier founders or companies.

[1:00:28] some combination of the two. But I think it's both. And I do think we're limited by outlier founders. But I also think we're limited by VCs that are willing to give money to, like that are able to recognize outlier founders that have a good idea and then give them capital. And there's other constraints. [1:00:46] as you've sort of, you know, working at a [1:00:49] place that's obviously, you know, kind of at the top of the game. Have you picked up things there specifically about – [1:00:57] how to identify what that looks like. Because one of the things I... [1:01:00] you know, don't have a good answer to is, I mean, I know, I know somebody who's outlier now, like I know, you know, when I see, you know, the Colisons talk, or you see, you know, these amazing people now, you're like, okay, I know good when I see it now. But that's a very different thing than, you know, they didn't look the same when they were 20. And all the rest of it, like, is that is that is that kind of taste in people? Is it teachable? Are there things that you learn by watching? Or there is there things you can learn by reading or thinking? Like, what is that? [1:01:30] On the first part of the question, I have to say I have learned an unbelievable amount being at Sequoia. Like, really, I have learned a lot being at Sequoia. [1:01:39] And that's on all dimensions. It's how to think about markets. It's also how to think about [1:01:46] founders, how to [1:01:48] find outlier founders, we're pretty obsessive around [1:01:52] the language we use for how to talk about founders. And I've just learned an incredible amount about company building, everything, the place,

[1:02:02] There's 50 years of institutionalized knowledge and an obsessive culture around documenting lessons. That language thing is so hard because you'll hear so frequently people be like, oh, this founder is really spiky. Yeah. Or you'll hear this language that to me comes off as. [1:02:18] It's horrible. Inaccurate. It's totally inaccurate. Yeah. And most people are completely uncalibrated. It's unhelpful. It doesn't help me be better at assessing next time. Genuinely, I have learned an unbelievable amount from Sequoia. [1:02:29] And... [1:02:30] There's people like [1:02:33] Mike Ritz, Ruloff, Jim Getz, [1:02:37] Doug, Alfred, such like these people are all unbelievable reads of people and I've learned a lot. But I also think I've contributed a few things. And one of the things that I contributed that I'll share here is one way that I think about outlier founders is I think. [1:02:52] think about the chess rating scale. [1:02:55] And I kind of wrote something internally on this. And here's a thought experiment for you. Have you ever played chess? Yeah, I didn't get that far up the ELO rating, but I've played chess, yeah. So in chess, the highest rated players, you know, 2,800 plus, like Magnus and others. And still like a pretty good... [1:03:13] players, even a thousand or whatever, and not... [1:03:16] But like a thousand rated player will be... They're not making stupid mistakes. Yeah, they're not making stupid mistakes and they'll beat someone that... [1:03:23] doesn't play any chess. If you, here's a thought experiment. [1:03:27] If you ask like a thousand rated chess player, you have them watch...

[1:03:33] They get to see a game and they don't know who's playing. They just see the moves of like 1400 rated players, 1800 rated players, 2200 rated players, 2600 rated players. They get to see just the moves of each of those games. [1:03:46] and you ask them to place on the rating scale, like from one game, what is the elo in each of these games, they can't do it. Like, they can't tell the difference... [1:03:58] of the 2600 versus 2200. As I was saying, the farther from you it is, the harder. The more unable you are to do it. Yeah, 100%. Yeah, within like one standard deviation or something, you usually can, but beyond that you can't. Whereas if you do it the other direction, if you ask the 2600 rated... [1:04:13] player to judge a 2200 rated game, an 1800 game, a 1400 rated game, a 1000 rated game, they can watch 10 moves and they can basically tell you the elo of those people. [1:04:24] And so one of the things that I obsess over is calibrating on – [1:04:31] Like if someone says this person is an outlier in AI, are they? Are you a 1000 saying the 1500 is an outlier? Exactly. Are you 1000 saying the 1500 is an outlier? Or like, what is the rating of the person making the statement? [1:04:44] Um, [1:04:45] You know, is this... [1:04:46] Noam Shazir or Ilya Suskevar or whoever, like, saying that this person's really good, or is it like a Stanford undergrad in CS? And a lot of people... [1:04:58] They struggle to understand the level of the person giving them the reference. That really fails with tail behavior and tail behavior.

[1:05:09] Okay, I love this analogy, just to push on it so I better understand it. [1:05:12] And I do watch chess on YouTube of 2600s playing. And they make a move. And I'm like, that was a dumb move. And I'm like, I know I just don't understand the move. And so I just, I know that I'm that. I do, I'm terrible at basketball. But I know what an NBA player looks like. And I could see an NBA player. So what parts of being a founder are which? I think the... [1:05:37] chess scale [1:05:39] only applies to purely intellectual things. I think that [1:05:44] A lot of other skills, like say sales ability, I think that is something that's more like being good at basketball. It's much easier to audit with your own eyes that almost anyone can tell compared to... [1:05:57] Outlier intellect and I think the outlier intellect it is [1:06:00] it doesn't always translate. So being really good at chess doesn't mean you're really good at math. I'll give you – I'm going to – I have to kind of go through this because I think this is important. I'm going to take off my jacket. Yeah, it's a little warm in here. No, no, it's great. See the lights. So I'm going to give you another crazy – [1:06:14] example of like just how many levels of outlier there are in mathematics in mathematics the [1:06:19] I mean, I think there are like... [1:06:21] 20 distinct levels that order of magnitude beyond just like really good at math in a high school class. At the highest end, you have like best mathematician in... [1:06:32] decades or century, you know, these are the like Terry Tao level mathematicians. One level below that, that's kind of a distinct

[1:06:40] level below is someone that's basically guaranteed to get a Fields Medal. It's like best in a decade. It's someone that does, you know, three or four things that would be worthy of a Fields Medal. And they're basically guaranteed to. And they're unmistakably different from each other. They're unmistakably different. The Fields Medal person would be like, I know that I talked to that person. I know I'm not that person. And the like. [1:07:02] Best in 100 years. [1:07:03] the Gauss or Euler or, you know, there's a small number of these people in history. They could... [1:07:11] get a field's medal in almost any sub-area of math. Like, they can just go and they know more about almost any sub-area than the experts in that field. But it's like, you have the once in a century of once in a decade, then you have the, like... [1:07:24] You know, the person that... [1:07:25] easily got a Fields Medal. And then you have the person that kind of got a little lucky to get the Fields Medal. Like there's maybe, there was like a 25% chance or 50% chance and this political star is aligned. But they picked the right subfield for their brain and for the moment in time and whatever is like that. Yeah. And like the political, they wrote the thing and like the political tides aligned in their favor. Their subfield was a little hot at the time or whatever. And then you have another level would be the people that, [1:07:50] Like, [1:07:51] Don't get a Fields Medal, but get tenure at a top [1:07:54] five math department at a really young age. At Harvard, there was this guy, Noam Elkes, who was one of the youngest, if not the youngest, tenured math professor there. Like early 20s kind of situation? Yeah, early 20s, mid-20s. [1:08:06] Um, you know, and it's like, this is like another level. Then you have the people that either easily get tenure at a top five department or get tenure, but called it in their early thirties or get tenure at like mid twenties at a.

[1:08:19] five to 20 university. Does that 23-year-old Harvard tenured professor [1:08:26] Do they yet know that they are not one of the higher echelons? Or is it still an open question at the time they're 23? It's a little bit of an open question, but not much. I think that the best mathematicians in the world, like Terry Tao... [1:08:39] can talk to almost any mathematician and basically understand where they are in the, like almost instantly. Age adjusted and everything. Age adjusted, you know, and not at the age of like 12. Yeah. [1:08:51] because so much of that is just like how passionate have you been in it, and what was your early exposure. But by the time people start doing research mathematics, I think that people can [1:09:01] tell very quickly. Another level below that is like, okay, you [1:09:06] Just you get tenure at a top five math department at any age, you know, and then it's like you get tenure at a top 20 math department. And then it's just you get a PhD from a top math department easily, you know, and then you get a PhD in math from like a top 10. [1:09:22] 20 department easily. And then as you just get a PhD in mathematics, and then it's [1:09:28] you studied math in college, and you basically were summa cum laude, and then it's like you were... There's still another five levels, and then it's even the kids... And all these levels are still impressive to most of us. Yeah, and they're all distinct, and everything I've mentioned, obviously this is 800 on the math SAT. Of course, so it's not a useful metric. It's not even a useful metric. There's like 15 distinct levels minimum above 800 on the math SAT that someone at the top can basically look down... Tell you about all of them. And tell you about all of them almost instantly.

[1:09:58] founders, you have to think about what is the outlier trait that I'm [1:10:03] that is important for this company. If it's a robotics company, being really good at robotics is important, but being good at chess is irrelevant. If it's a sales-oriented company, it doesn't matter how... [1:10:16] great you are. But kind of knowing what dimension matters and then being able to place where on the scale they are so you can actually have like a nuance. You can actually underwrite in your investment thesis. It's very important. Very few people do this. So if you're just you're merely the summa cum laude, [1:10:30] Or you're merely the, I got my PhD easily. [1:10:34] but you know that these hard companies are built by mega outliers. Like, should you play the game even? Like, if you're like, I'm not going to know... [1:10:42] oiler when I see him. What's the point of even trying that? There's many factors that go into building a company. My comment is really just around [1:10:52] underwriting fruit. For this type of company... [1:10:55] I think... [1:10:56] you know, being like, [1:10:57] For this AI company, [1:10:59] foundation model, just how good are the researchers? You know, like, I think that is an important part of underwriting, you know, like an AI, [1:11:08] research or in the like robotics world model companies, like just how good are these researchers? And you want to back the absolute best because they're going to attract the best, et cetera. But in, you know, a basic SaaS company, it's not that important. How does this language show up in like a Sequoia sort of analysis of a company? [1:11:25] Like, how do you use this metaphor, which I love? How do you apply this? Is it about...

[1:11:30] Is it about which partner we should be listening to and like who's got like, you know, like sort of Bridgewater style, like who is the trustworthy voice here? Is it about like, you know, what's it about? Like, how do you apply it? [1:11:41] Yeah, we just try to get very specific on what traits matter for this company. [1:11:48] Is this founder an outlier? [1:11:50] in those traits that matter? What level of outlier are they? When we're doing the references on them, qualifying, what is the rating equivalent? Is this an ELO 25 or 15? Yeah, like who is giving this assessment? Is this a thousand rated chess player rating [1:12:07] giving the reading or is it a [1:12:09] 2,400 rated chess player. And so it's, [1:12:13] But you can't just say this is a spiky founder. That will not fly. And we care a lot about the texture of how we describe. It's interesting because I think people who are investors who are sort of more market-driven or, you know, I think this often comes with later stage investors are, I think, much more precise in the way that they evaluate versus people who are more founder-driven or often more vibes-y. But I don't think it has to be that way. It feels like you could get nearly as... [1:12:42] intentional and specific about the way that you are founder driven even and that it's just most people just don't do that work and know even how to talk about it i'm naturally not very [1:12:53] good at language and was never interested in language. And it's one of the things that I've learned from Sequoia. Sequoia is obsessed with language, obsessed with words. Things like we don't invest in companies, we partner with founders. We don't do deals, we partner with companies. When we write memos, we don't say I, we say we. I think Don Valentine gets a lot of credit for that. He was obsessed with that. I think Mike Moritz was a lot of credit for that.

[1:13:20] took it to a different level. You do speak in metaphor a lot, though, which is something that I think is very necessary if you're going to try to be accurate about founders, because I think so much of it is having those devices to try to take this very amorphous thing and put it onto a map that you can do anything with. I think that way, but I didn't connect it to language before Sequoia. So this is one of the many things that I've learned there. Cool. All right. I'm gonna let you go. It's been a long time and it's warm in here. Thank you for doing this. This was really awesome. Thank you for having me. It's been a blast. Thank you, Jack. [1:13:50] Bye.