47: Paul Scherer - A Friend That Brings Us Closer

AIs of some kind: Creating like a person that is uniquely incentivized straight belonging and connection in the world. We know more people than ever before, but at the same time, we are like much less meaningfully connected. The share of Americans who have like zero close friends went from like 3% to 15%. Like half of American adults report being lonely. Feeling lonely is the equivalent to like smoking 15 cigarettes a day. Half of America smoking 15 cigarettes a day. A world that is like completely isolated is like a world in which humans go extinct.

Paul. Before we get: You have said explicitly you're building a mutual friend. What are you trying to do? AIs of some kind: It goes back to like a lot of the like simplicity of just go to our website and it's like, what do you mean you don't know what we're doing. It's like, I just told you, it's like we're building a mutual friend. And then you're like, yeah, but what are you actually doing? It's like a mutual friend. And then I like usually like add, you know, some version of what we just talked about.

And then they were just like, oh, you're building a mutual friend. And I was like, yes, I've been telling you. I think a lot of what we're thinking about and what we're creating, it starts actually counterintuitively, not with the friend part, but the mutual part. Paul. Before we get: Do you think you're authentic? AIs of some kind: My suspicion is that everyone is kind of the same in that we all have this like inner voice. And I think adulting in a lot of ways is like learning to not listen to that voice.

I feel like I always just listen to that voice and everything I've ever done, every major decision was always just like, I was like, okay, I know that it's like, this is just my voice and I just like, I have to do it and I like followed my heart or whatever you want to call it. And maybe that's authenticity or maybe that's something else. Speaker C: Welcome to Dialectic, episode 47 with Paul Scherer. Paul is the founder of Eigen, a new company building a mutual friend for the world. That might sound bizarre.

It kind of is. But I was thrilled to sit down with Paul after being introduced to him by a few people I really respect. To talk about his crazy vision for a way that AI might actually make us less antisocial and instead bring us together by being somebody that we all know and who creates social serendipity in our lives. Uh, Eigen's pretty early on, the product's in private beta, and so we weren't able to talk about everything, but we were able to dive a lot into the substance of what it is, the philosophy behind why Paul and his team are working on it.

In the ways that this introduces just a crazy new range of design problems for teaching, parenting, growing a sort of person that we might all know and that could actually make us closer to each other. Paul is very early on. He is young. I'm sure there are things he doesn't know, but I have to respect his conviction, his point of view, and the way he is authentically trying to bring something to the world that he thinks it needs. I hope you enjoy my conversation with Paul. Before we get into the episode, I'd like to thank Notion, Dialectics' presenting partner.

Notion's co-founder Akshay Kothari is actually the one who introduced me to Paul. And so I'm thrilled to bring these two worlds together. Notion is a collaborative workspace for your life's work, and over the last year has totally evolved itself into becoming a platform and a hub for the agents that help you and your team do great work. Notion also made a bunch of new announcements on its developer platform recently. I'll link to that in the description. And the way that they're pushing what you can do with agents is really remarkable.

And certainly if you haven't used Notion in a while, probably goes far and beyond, uh, what you can imagine for what's possible with Notion and with the agents. You can learn more at com/dialectic. I highly recommend you checking out what is now possible as Notion has built AI in from the ground up. With that, Here is my conversation with Paul Scherer. Paul. Before we get: Paul Scherer, thank you for having me here. We're in the Eigen office. Pretty specific type of space we have. You've curated for yourself. AIs of some kind: Yeah, it's really nice.

We're very happy here. It's like, it's always like, I know, I think it's really important. I think a lot of people underestimate the power of space. We spend a lot of time here and, you know, I think we spend much more time here than at home, so it's like it's got to feel really great. Paul. Before we get: Well, I want to start by talking about this book that I had to hunt down, Momo, a book that's referenced in your video. And either it was a good story or seemingly it was very influential to you when you were young.

It's a children's book. And in reading some of it, it seems to me that it's— there were a few ideas that really stood out. The first is time and playing with this idea of past, present, and future. Second is presence, obviously an extension of that. The third is listening, which obviously relates to presence. And then the fourth is friendship. And so there's this smattering ideas that I think inform a lot of the other stuff we're going to talk about today. Yeah, but I'm curious why— what about this book stuck with you?

I don't know when you first read it. I assume you were quite young. AIs of some kind: Yeah, my dad read it to me. I was like, I was like 8 or 10 or something like that. And it's like you read it again and like school and it's like, it's kind of like a thing in Germany. I think one of the things that really, I think is very true is just like they have this concept of time thieves. And I think it's really interesting because the book is like from 1973.

Yeah. And so it felt pretty modern though. Well, this is the crazy thing is like it's like pre-internet, pre like social media, pre, you know, real like personal computers or anything. This is like, it's like at a time where I don't know, there was like a bunch of computers, but like most people didn't have a computer. Right. And so it just feels like You know, every year since then it's like become more relevant because it just describes so accurately kind of like the world we, we live in. It's just like everything is just about like efficiency and it's about like, you know, there's no— there wouldn't be time for Momo in this world, right?

Like people don't care. It's like, you know, how does this help me, you know, get, you know, more, more efficient or it's like more productive or whatever. And I think, I think that's like really— I think it's always like a sign of like great work if it's like becomes more and more and more relevant over time. And, and, and I, I know I give it to a lot of people and a lot of people really resonate with it. Paul. Before we get: You are a striver. I don't know if that's a perfect word, but it's probably a representative word, at least of, of this city, of this industry, of doing startups at all.

You're someone who dropped out of college or, excuse me, high school when you were 17. Do you feel like you're running out of time? AIs of some kind: I used to feel like that more. I don't know. I don't, I don't, I haven't felt like that in a while. I think it's like, I think, you know, for the first time ever, I feel like we're like doing the thing and it's like there's like no real alternative. It was like, you know, I was doing all these other things and I'm still like restless, but I feel like there's like, like there's no, there's like no way out.

Like I couldn't like go and be like, okay, great. Like let's, you know, whatever, do anything. Something else is like, this is the thing. And it's like, and so you sort of stop feeling like you, like you're running out of time because it's like, there's no, you're, you're, you're actually playing your game. And it's like, and that's just the game. And the game is like, you have to play it. And it's like, it wouldn't be fun if you wouldn't have to play it. And, um, but, but before I was feeling like I was playing the game, it was like, yeah, okay.

Like I need to like, I need to play. Paul. Before we get: I need to get to the thing. AIs of some kind: Exactly. Hmm. I like that. Paul. Before we get: Yeah, there is a way that it's not quite abundance, but there's something about once you've kind of found the thing, you're like, you're— and I'm sure you're antsy in a whole bunch of ways around getting progressing this thing. Yeah. But it's like, oh yeah, I'm like, I'm where I'm supposed to be. AIs of some kind: I call it like, you know, like I was talking about this with Akshay a while ago and I was like, it's like short-term paranoia, but like long-term, like Everything is, everything is exactly the way it should be.

It's like, what about listening? Paul. Before we get: There's, it's kind of the anchor. Momo's this character, this little girl. I don't know if she's a little girl. She's a girl who lives in the village that they go to. And the reason they're initially skeptical, the reason they are so, um, they're able to be so present with her, all these things, she helps people deal with their beef, whatever, is she's just this amazing listener. And maybe there's an element of this. We'll talk more about it. Like, you are someone who certainly has an observational lens on human beings and the way people behave and are going to behave.

But I'm curious if there's any thread there that felt resonant. AIs of some kind: I think, like, what's so interesting is, like, she's like the polar opposite of the time thieves, right? So, and the reason she— it's, it's almost like less about the, like, listening and more about the, you know, the presence. But she's like the manifestation of, like, presence. Paul. Before we get: Yes. AIs of some kind: Yeah. And, and she listens so well because she doesn't look at her phone, you know, she doesn't, like, feels like, oh, I have this, like, meeting 15 minutes from now and I'm, like, looking at my watch and I'm like, I should probably— I should probably need to go.

And it's like she's just there with you. 100%. And I think that's like, that's the, that is like, that's like so antithetical to the Time Thieves because it's like they're like the manifestation of like, you know, you have to go, there's like everything needs to be like, and so she's just like fully there. And I think that's something that is very rare because we all have, you know, so many things that are happening that are like, you know, you know, trying to get our attention. Trying to pull us in. And also so much like, you know, not all of this is like bad, right?

It's like there's a lot of opportunity, a lot of things that you could be doing at any given point in time. And so to be somewhere really present without thinking about, you know, the best listeners or the hard thing about listening is you have to actually have to listen instead of like thinking about what you're going to say next. Right. And that's a really hard thing to do because you have to be truly present without any, almost like any skin in the game of like, I need to like achieve this or that or so.

It's just like, I need to like, you know, tell them, you know, like how great I am. Or like, you know, that's like when you don't really listen, but you're just like, okay, how can I like, can you give me something that I can then like say the thing that I want to say? And I think that's like really interesting because she's just, I wonder if like nonchalant is like the right word, but it's like she's like, she's just there. There's like no— it's like there's a sense of ease. Yeah.

Yeah. And I think that's really remarkable. Paul. Before we get: Maybe that last bit, which is friendship. We're going to talk a lot more about friendship. You grew up in a remote place. What does it mean to be a friend? AIs of some kind: The— I don't know if there's like one definition that like works for every person. I certainly don't have it. I think there's like a few different themes. I think like great friends help you rediscover yourself, but they're also expansive and like, and like surprising and delightful ways.

And they add new things. Yes. And so there's really these two elements, right? There's like they ground you in who you are. Paul. Before we get: Yeah. AIs of some kind: But they also like push you and there's like deeper and wider, right? Paul. Before we get: Right. Maybe first of all, like I briefly just alluded to it, you grew up in a place, as I understand it, that is quite remote or at the very least quite small. Not that far from Frankfurt, but not, not in the center of things.

Yeah. Now you are in San Francisco. You are in the center of things. How has the internet changed your life and particularly on the dimension of relationships and people? AIs of some kind: I think every single person you know, Bar or Samuel and my parents that is in my life, I wouldn't have met without the internet. I think I was like 19 or 20, probably, you know, I think it was like 20 years old when I met the first venture capitalist. You know, maybe I was 21. And it's like, I didn't know what that was or there was like no you know, there's like the— this world did not like exist in my, in my like world.

I didn't know about it. And so, um, and neither, you know, like anyone in my like, like there's no one that I met that was like, oh, here you are, one step removed from it, right? Uh, yes. And so, um, without the internet, I think I wouldn't like— I definitely wouldn't be here. And then all of these relationships, or so many of these relationships are, you know, first or second order effect of like just like meeting people on honestly Twitter. It's like I met a while ago, I met one of the founders of Twitter.

It was very, you know, transformative. Like, you know, so I was like a bit like, meet you here because it's like not even here. It was just almost a spiritual experience of like this person. So in such a profound way, like changed and shaped my life. In like such a big way that I don't even know if I could recognize me as a person without that thing. Mm. Paul. Before we get: I've had a similar experience with that website in particular. What about Twitter? Well, maybe first of all, were you as a kid or growing up, especially before you started to tap into this, were you a social kid?

Were you lonely? AIs of some kind: I wasn't very— I was like Marmite always. Like people either really like— I had like, I always made friends with some teachers. And so like some teachers always really, really liked me. And then like some really hated me. And then I always had like trouble with like people my age. It was like kind of difficult usually. The older I got, the like easier it became like directionally. But it was never— kindergarten, it was like easier. It was like, I was never like fully, fully lonely, but there was like, I was also, I was never popular, but I was, I really deeply wanted to be popular.

Um, and, but I was not at all. And, and, you know, reflecting on me as I, it was, it was very obvious that I wasn't popular. I was like, I, there's a lot of like social strategies that I did not know about. And I was like, if I, I think there's like there's a lot of reasons of why I wasn't popular. It was always like, like that. And then, yeah, the older I got, the more easy it became. And then you just say I left school and, you know, it's like sort of you start— I always was like hanging out with like older people and then it just started working.

And so like immediately everyone there was like 30 plus. And so that just became my— Paul. Before we get: What about Twitter or what about the internet allowed you to find whatever. Um, because I think next thing I want to talk about, maybe to, to, to turn the corner on it, is like in what ways the internet has failed us as this connective tissue. But I'm, I'm first interested in the ways that it's actually like, why was Twitter for you at that time? Why did it work? And what did you find there that was good?

AIs of some kind: I mean, I was on Twitter a lot, you know, during the height of the pandemic where everyone was on Twitter and it was kind of the greatest place on earth because all these cool people had nothing to do other than being on Twitter. Paul. Before we get: That and Clubhouse, right? AIs of some kind: Yeah, exactly. And so that was really great. And I actually went on Twitter. It was actually the wildest thing. I was like, at the time I started like a tech blog and I was like, I'm going to write about like, you know, like these tech gadgets or software things that I really liked.

And I wrote like 2 articles and I was like, oh, well, no one is reading this. Like, how am I going to get this, like to promote this? And then I was like, maybe I'm just going to tweet about it. And so I like created a Twitter account and I, you know, it was like, it was like, it was like, yeah, I created my Twitter account. It's 2021 or something, 2020, I don't remember. And you started just like sharing this article and it's like kept going. Paul. Before we get: What did you write about?

AIs of some kind: These articles, like really just like YubiKeys and like 1Password. It was like the most— Paul. Before we get: You were reviewing them? Yeah, yeah, yeah. AIs of some kind: It was the most obnoxious thing in the world. And— Paul. Before we get: That is not the most obnoxious thing in the world. AIs of some kind: It's obscure, but it's definitely not— It's obscure. It's not obnoxious. You're right. Paul. Before we get: Some people might think what you're doing now is obnoxious. We'll get to that. AIs of some kind: But UV keys?

Yeah. To be honest, yeah, it's obscure. It was an obscure obsession for this kid to write about, this 17-year-old kid somewhere to be like, here's why UV keys are really great. Anyway, so I wrote that and I tried to touch on promote the blog. And within like 5 days, I was like, fuck the blog. Like, this is really cool. And I actually started like meeting people there. And it was like, there was these like, there's like, I remember there was this guy from, from Ghana, actually, who's like, I think he now has like 20 or 30,000 followers on Twitter or something.

And we like, we like somehow like met and we had like, I had like 200 followers and he had like 300 or something. And we just like DM'd and then we like, it'd be like this whole thing. And I spent 12 hours a day on Twitter. And I have like 20,000 tweets and replies from that 3-month period or or something. And it was just like, I would like send, you know, 600 replies a day to like tweets. I would like, and it was just like, all of a sudden I had like the first tweet, it had like over 1,000 likes or something, really the biggest reaction.

Then like all of a sudden there would be all these entrepreneurs that like would, I remember like there was the CEO of like ClickUp or something, like liking and commenting on like my marketing. I was like a 17-year-old kid. I'd never like, I mean, I'd like, I had no experience whatsoever. And I was like, here's how you should market yourself. And it's like, and the CEO of like ClickUp was like, that's great advice. Right. It was just like, that means a lot about Twitter. Paul. Before we get: Honestly. AIs of some kind: Yeah, that says a lot about Twitter.

You know, deeply researched. You know, it's like, and then, yeah, I think like, you know, people start DMing and I actually, I sold this, like, I had this like digital product that I sold, which was called the Twitter DM Mastery. And for a while— Paul. Before we get: It's like a course? AIs of some kind: Exactly. For a while I was the self-proclaimed king of DMs. Okay. And I, but I, because I would, I think for the first 5,000 followers, I DM'd every single follower that I like manually would be like, hey, like, I'd like, thanks for it.

Like, I appreciate the support, whatever. And that's how I met a lot of the people. Um, yeah. And, and then I wrote like a book about this, which actually just like, I actually still think it's probably like a lot of people should read this because it's just like, it's very like, I mean, book is a strong word. It was like 20 pages or something, but, um, a great PDF. Yes. Paul. Before we get: Um, on DMing. AIs of some kind: On DMing people. It was the whole— it was like a whole scheme.

Anyways, so that's— I did that, and then I think that's just how I met— like, that's how I met Ariel, um, that's how I met Caleb, um, who then, like, all of these things later turned into, you know, working with different kinds of companies and all that. Paul. Before we get: Many of us have had positive experiences, um, whether they be Twitter or elsewhere, um, and there's a romantic idea about the way the internet used to be or what it was meant to be, um, this like should be the best connective tissue in the history of the world.

And yet many people feel the opposite. You've talked about what happened when we conflated or combined social and media. What is social and what is media in this context? AIs of some kind: At least I think in this context, you know, media is like, you know, maybe could be defined as like popular content, which is like, right, if you scroll on Instagram, it's actually not about your friends, right? It's like, it's about, um, it's about, you know, popular content or popular people, right? So it could be a reel that is going viral from like a random account, but you don't know that account and like neither do your friends.

Um, and it could be a celebrity. Um, but it's, and, and social is much more about your social graph, like who are the people that you know? And I think if you sort of, a lot of the interaction paradigms of these platforms are originally used to be much more social, which is like, it's all about sharing a story with your friend or posting something and all your friends are going to see it. But like actually now it's sort of become much more media, which there are reasons for that. And, you know, it's just like, it's all about incentives, right?

They didn't do that for— Instagram wasn't there being like, that wouldn't be cool if we were really evil and we like made it all about that. They're just like, well, that just works, right? And it's like, it's really hard to get a lot of people to share content, for example, right? So like one of the issues is that, um, you know, your friends are probably much worse at creating engaging content than like someone who's really good at creating engaging content. Paul. Before we get: Right, right, right. It's not a totally common experience, but I think across different parts of the internet, whether it be Twitter or Instagram or other modern things or forums in the past or IRC or whatever, Um, I think one of the things that is wrong is, is the notion that you should only use the internet with people you already know in real life.

Because to your earlier point, like, social doesn't necessarily mean people you already know. AIs of some kind: Yes, that's true. Paul. Before we get: Like, can you talk a little bit about that distinction? Like, like in theory, um, in my view of the internet, it's almost actually that like, or at least what I, what I always felt was so amazing about Twitter is that it wasn't about who you knew. It wasn't just about what you were interested in. It was about getting to know people by way of what you were interested in.

AIs of some kind: Yeah, I think that's, I think that's true. And really, you know, there's like this like discovery aspect of like, and I think for better and for worse, right? There's like, I always say there's like, there's on the internet, there's like enthusiasts for everything. So, so like if you live in this like place with 1,000 people or even, you know, like a medium-sized city or whatever, there might not be someone who's like absolutely obsessed with, I don't know, plants, right? There may be, I mean, it might be a more common obsession, but, but on the internet there, there's like millions of people who are.

And so there's like, there's definitely an aspect of like feeling like belongingness in, in that because you're You have so much more reach. Paul. Before we get: Um, and find the other weirdos. AIs of some kind: You can find the other weirdos. You know, that, that, that creates a very long tail, right? Of like, um, and that, that sort of then is like one of the reasons that, um, we as a group have much like, cause we just like, we were, we were able to like go into much more of these individualist pursuits.

Um, because there's so much, so much, the, the, the tail is just so much longer. then, then it would be like a city, right? Which makes— Paul. Before we get: that's not intuitively antisocial to me. Like, in theory, it should lead to the long tail of a million, like, let a million communities bloom of weird niches. AIs of some kind: Totally. I think the, the, these aren't like mutually exclusive, but, but the, but the problem is like, you're, the things that you consume shift towards a more, at least locally isolating view, right?

Because your Twitter feed is like filled with people that are obsessed with plants, right? But they might not at all be in your local, like, actual, like, in-person proximity. Paul. Before we get: Yeah. AIs of some kind: And so all the people you interact with in person on a day-to-day, like, have their own, like, isolated feed of these types of people. And then you have, like, much less of, like, a sort of bridging experience where you could, like, go to your office and there'd be like this thing that just like everyone knows about, right?

Paul. Before we get: Like the global village, the, the water cooler talking about reality TV or— AIs of some kind: Yeah, exactly. Like it's like moon landing, moon landing, or, you know, even just like, even just like headlines or something or like memes, the most like mainstream seeming, you know, meme or like headline or whatever is actually not that mainstream anymore. Right. Paul. Before we get: Right. AIs of some kind: Where you'd be like, oh, wow. Like SNL adopted this thing. Everyone must know about this. And you go to your office, you're like, oh, wow, did you see this thing?

And like people are like, What, what are you talking about? Right. And that's like the problem, right? Paul. Before we get: You can do it in San Francisco a little bit with Twitter because it's like you, there is, but is, do you think personalization of everything is just fundamentally inevitable? That really, that's what this is about, right? Is like what technology has gotten really, really good at with regarding our attention is personalized, is just tuning it to being exactly Paul shaped. And that is causing this context collapse. That you're pointing at.

AIs of some kind: Yes. I don't know if anything is like inevitable like that. I think it's definitely— there's like good reasons for it and it's not— it's not like completely bad, right? There's like great things about, you know, stuff being personalized to you. I think I ultimately do think that people are craving like social, right? It's like you don't want in some cases you do, but in many cases you actually don't want the, you know, the take or recommendation or whatever that's like personalized just for you. You also don't want the monocultural, like average take of the world.

You want the take from the people that you care about. Paul. Before we get: Right, right. I want to talk a little bit about what you are making, but before we quite get there, You have this view that like much of what we were just discussing, um, there's all these tailwinds. There's, we are more isolated, people are more alone. Derek Thompson's gone crazy on this. People aren't having kids, all this stuff. Um, and it seems like your instinct is that this is, as many people believe, this is driven by technology.

Um, you have used some variation of the phrasing, the world needs this. in terms of what you're doing. Um, and there are two maybe cuts on what interesting views or unique views you have. Um, I think they are one being thinking about friendship in a slightly new way. And then the other is this idea that AI as it currently stands is pretty antisocial. And I think those two things go together. Maybe one, one strange cut on this is that opens the door is like, we all talk about AI as like crypto or VR or whatever.

Um, the one person I've noticed who doesn't talk about it like that is Kevin Kelly. Um, Kevin, uh, former editor of Wired magazine, a bunch of other things. Kevin talks about AIs, which is a subtle but important difference. Um, and it does point at maybe this future of a world where there are— many of us are talking to AIs of some kind. And it seems like your core view is that based on how everything is going, we are going to be talking to like AIs alone in our rooms. First of all, what would you say to the people who are still like skeptical of the notion of— you seem pretty convinced that we're all going to be talking to AIs or AI or whatever.

AIs of some kind: Or a lot of the people in our direct proximity are already talking to AI all day. Like every single one of our engineers that I get to talk to. Paul. Before we get: Maybe they would say I'm using Claude. AIs of some kind: Yes, but like we are already communicating with it and like Samuel sometimes swears at it and just like, what are you doing? Like it's just like, you know, it's like why we can make these things. Paul. Before we get: Samuel's one of your first employees.

AIs of some kind: There's like a bit of a difference in the interaction paradigm from like, you know, software in terms of like, I always like, I always ask people whether or not they say thank you to like, do you, are you like someone who says thank you? Paul. Before we get: I don't always say thank you. I'd say thank you probably less than I used to, but I try to every once in a while. AIs of some kind: Yeah. Just to make sure we're on the, on the good books, you know, when they do a really, when it does a really good job, inevitably it takes over the world.

We can just like, we'll be on the good books. Paul. Before we get: Do you say thank you? AIs of some kind: I sometimes do. I try to, yeah. But that's like a ridiculous thing, right? Who's like, were you ever like Google something and you're like, oh, thank you. So I think that's true. And I think the other thing that's true is that fundamentally it's interesting to think about this in the context, for example, of a lot of education software or products because previously it was really, really difficult to build, for example, products for basically like UK through like 8, basically like young children, just as much as it was really hard to build for elderly people.

And it's interesting because when you build software with a graphical user interface and you have to learn a new interaction paradigm, you can make them like skeuomorphic in some way, right? Like make it feel really comfortable and make it like, okay, this looks like, you know, your Rolodex looked like, you know, in reality. And like, so it's like a contact thing. But at the end of the day, there's like, there's like a barrier of like that you have to kind of overcome of like, okay, this is like, if I give like my grandma a phone, she's like kind of scared because it's like this thing and it's like, it doesn't feel real.

And I think what's really interesting is that the, like, AIs speak human language. And so the interaction paradigm is the exact same that we all, you know, have been learning ever since we were born in interacting with other, other people. There's like a, you know, you can build a trust in that it can understand your intent and it can like, you know, relate. And, um, and I think that really changes a lot of things where it's like is like, will my grandpa ever fully adopt, you know, you know, internet technologies?

I don't know. But I actually think it's not unlikely that he'll like eventually adopt a lot of AI products because it might be embedded in his world in such a way where he doesn't actually have to reflect that much on, on whether or not it's not— it's that or that. Paul. Before we get: It's more human-shaped. It's in a way that almost necessary, like it's almost necessary that we anthropomorphize it. Maybe not necessary, but it's inevitable, almost inevitable that we anthropomorphize it. AIs of some kind: We already kind of anthropomorphize it because we talk languages with ChatGPT and he has like a voice.

ChatGPT, like, you know, there's like all of these things are very niche and they're like not, you know, it's not, there's a small minority that is currently feeling like that. But like the like GPT-4.0 thing of like how many people got like, in like a density. Paul. Before we get: I don't know that that's a small minority. AIs of some kind: But it's like, I think it's like largely a small minority. Paul. Before we get: Yes. AIs of some kind: Compared, like there's just like stats. Paul. Before we get: A lot of people are doing it, but yes.

AIs of some kind: Yeah. There's a stat which I'm sure you've seen of like, you know, all these like dots of like, you know, millions of people represent millions of people. And there's like all these gray dots and then there's like a few green dots, which is like people that have used AI for free. And it's like 2% of the world or something like, I don't know. We are very, very early. And this. And I think it's easy to forget that when you are like around people who are like, we are in this like industry and in this, in this, in this, you know, very specific place where everyone is like somehow billions of dollars deep into really believing that this is like everything today.

And I think it's probably not, but it's going to be so much more in like 10 years from now. But stuff just takes time usually to like become meaningful. I always just like have this thing where people just right now on Twitter, you know, there's like a different take on Twitter like trending, you know, every couple of weeks. And right now it's like, it seems to be like product is like dead, right? It's like been automated away. It's like, right. Paul. Before we get: Anthropic's got this, right? AIs of some kind: It's like, it's like, it's like, it's like there's, you know, it's commoditized to build great products.

Yet I've actually like, tell me all the great products that have come out since ChatGPT dropped. I don't know that there's a single new product in my life other than the LLM itself, and then for enterprises like Claude Code. But other than that, these two things, there's not a single product that has come out that has changed my life or my mom's life. And I think that's like, it's still just as hard to innovate on product. It might be easier to build once you have a great idea, or all of these things are somewhat true.

You can build a lot of stuff much faster, but it's still just as hard to build a great product. And that's why it takes time because figuring out what it looks like to build a great product is still really hard. And I don't know, I think it's just sort of these consumer builders that are going to build the 5 to 10 products that are going to maybe come out of this and that real people are going to use for the product, by the way, not necessarily because it's an AI or not an AI or because it's like this or that model.

but because it's a great experience, I think they're like, they're just like being started right now. Paul. Before we get: People are talking to AI friends, um, whether it's the 4.0 thing, my boyfriend is AI, it's Replika. And again, to, to the earlier point, like most people probably see that as strange or bad or evil or whatever. Um, your cut on it is that it's critically, it's antisocial. AIs of some kind: Like most of these products actually incentivize you to spend more and more time talking alone to the AI, which is, which is by the way, it's like, that's not like, um, Zach went on a podcast and he was like, it was like a year ago or something.

And he literally said, humans have a capacity for 5 friends, 5 close friends, but the average American only has 2. We're going to build the other 3. He literally said that, right? That's what he wants to do. His best case is for you to spend your Friday night talking to this thing. Right. That's what, that's what he likes or what, you know, you know, what his organization that he's running is like trying to create in the world. Paul. Before we get: You are, I think there is some, there is some question of what you are actually making.

Um, and I want to litigate that a little bit today. Um, acknowledging that you're not ready to share totally everything. Um, you have said explicitly you're building a mutual fund. Yeah. Um, but you are building an AI fund. In some dimension. I want to talk about it, but what are you trying to do? AIs of some kind: I think we are trying for you to spend your Friday night with other people, right? And we're trying for you to— I want people to feel like they could belong. Paul. Before we get: Okay, let's go one click deeper.

What are you making? AIs of some kind: We're building a mutual that is like where I think a lot of what we're thinking about and what we're creating starts actually counterintuitively, not with the friend part, but the mutual part, right? Which is the network, which is, it's actually, you know, we, you know, we have like, right now we're like 4 engineers and like 3 of them are working on the network piece. Paul. Before we get: Yes. AIs of some kind: Right. Which is partially because it's very hard. The friend piece is like something that is very hard to split up and there's like, But it's also because it's really, really, really important that at the core of everything we do, the network, it like sits the network.

It's like everything, like we have this like product principle was like shared from day zero. It's like every time we built something, we build it in a way where it's sort of built globally, where it like, it is built on top of the network where it's shared between all of the people that, you know, or that like are sort of users and And I think that's like, you know, a really big difference. And then obviously there is like the, you know, the friend piece of the mutual friend, which is, you know, I just think that, you know, in lots of ways, the most intuitive and natural way of interacting with this like incredibly powerful network that we're building is like an extra person.

And like someone who we just all know in common, because that— there's again, you know, as we previously said, there's a lot of like interaction paradigms and norms already established of like how that would look like. And like, you know, if I share something with you and you tell to someone else, like these, like there's like norms that are established that, um, are, you know, either, you know, codified or even just implicitly clear to most human beings. And Goes back to the point you're making about your dad. Paul. Before we get: You know how to interact with a person.

AIs of some kind: Exactly. Paul. Before we get: On some level. AIs of some kind: Exactly. Exactly. Speaker C: As a UI metaphor or whatever. Paul. Before we get: Exactly. AIs of some kind: No, you do. Like, it's like you don't really have to learn it. And I think it's actually kind of interesting because a lot of people are still, you know, we went out, we, you know, we did this like announcement and we're like, we're building a mutual friend. And then everyone's like, well, what are you doing? And I was like, we're building a mutual friend.

People, I think, still have like a hard time accepting maybe the, which is totally fair because they haven't seen it. And I think it's like, in a lot of ways, it's like a, it's like a Waymo product where everyone I've ever told about Waymo that isn't maybe from San Francisco, like if I tell my mom about Waymo, she's really scared. She's like, this is the most craziest thing ever. Like, you know, and then you're just like, you're like, force her to take a Waymo. And she's like, I don't want to do this.

And like 10 seconds into the ride, it's like, it's the most normal thing in the world. Right? It's like, of course, it's like a spiritual experience of some level. If you like, understand, it's like, that's kind of a crazy, you know, feat of achievement. We have like self-driving cars and it's like, no one, by the way, it's like, I find it crazy that people don't talk about it because it's like, I grew up in like this world where self-driving cars were like this like crazy thing that maybe one day that wouldn't be like insane and so hard.

And now we just have self-driving cars. Paul. Before we get: And by the way, 5 minutes, 30 seconds of your first Waymo ride, you're like looking at your phone, you forget that. AIs of some kind: Yeah, exactly. But that's the thing, right? It's like, It's so— it just makes sense. And I think in a lot of ways it's like maybe that's the struggle that people have is they haven't interacted with it yet. And I think there's like something that like kind of like clicks for a lot of people once they first interact with it of like, oh no, you're actually— it is just a mutual friend.

It's just, it's just like everything is just like because people like, I think if you go to our website and you open like this letter that we wrote, and you just take it very literal. I think it's just like, you're just like, you are being quite literal. Yeah. Paul. Before we get: It's funny. Um, Samuel on your team here, I spoke to him and he was, he said, in some sense, this is actually a fairly small step or at least a medium step, um, technology-wise, but it's a very big step idea-wise.

And I think that maybe is getting it part of this. I think it's also worth establishing one of the things you said to me early on is like, what would happen if a person could be friends with a million people. AIs of some kind: Yeah. Paul. Before we get: Which obviously isn't possible for a human being. Um, I also think one of the things that was interesting to me is that most of the AI products we have built are trying to do human work better or faster or cheaper. AIs of some kind: Yeah.

Paul. Before we get: Um, you are explicitly building something that a person, a human being couldn't do. AIs of some kind: Yeah. Paul. Before we get: And it's one person. AIs of some kind: Yeah. Paul. Before we get: I have talked to this person. This person will have a name. Um, why is it critical that it is one person with a personality? AIs of some kind: It just, it just comes back to, again, the mutual part is more important than the friend part. It basically comes down to like the question or the debate, or it's an interesting thought of like, How much more or less flexible is personality as an interface than like graphical user interfaces?

Paul. Before we get: I think it's going to take people a second to grok that. I mean, really, how much more inside of what you're saying is how much more flexible is interacting with a person? AIs of some kind: Yeah. Paul. Before we get: Than a graphical user interface. AIs of some kind: Exactly right. Where you have consumer networks are very unbundled right now, right? Um, like, like there's like all these different consumer network products, pretty much all of mainstream consumer products. Paul. Before we get: You're talking about Twitter, Instagram, social media products.

AIs of some kind: But even like Yelp, like Google Maps, um, Waze, right? These are all like consumer networks. Basically every single mainstream consumer piece of software that is on your phone right now is networked, like has, has a, is like based on top of a network except for ChatGPT. It's like the only like non-networked, like sort of like mainstream, like scale product. And, um, but, but the, but, but if you look into the networks of like all these different apps, it's like, you know, there's a lot of overlap between all of the Yelp users and all of the Twitter users and all these users, but there's a lot of specificity in the user interface of the app that makes it very obvious that Yelp couldn't also be Twitter.

Paul. Before we get: I see. AIs of some kind: And I don't know, or it's not, at least not obvious to me that the same thing is true for personality. I think there's a few cuts that you're going to have to make, which is like, I think you don't want your mutual friend to be your assistant. It's like very, very importantly, and that's like one big separation. Paul. Before we get: And also I think it's worth establishing in some sense ChatGPT, if you were to personify it, or Claude, is a mutual friend in that we are all quote unquote friends with Claude, but Claude doesn't know we're friends.

AIs of some kind: Exactly. Paul. Before we get: That's the part I think maybe that might not be obvious to people. AIs of some kind: I actually, I would push back because I don't think Claude is a friend. Claude is a servant. It's like an assistant. Paul. Before we get: Claude's someone we all know. AIs of some kind: But you lose a lot of the— he's a mutual assistant, right? And I think you lose a lot of the important parts of like a friend the second it becomes an assistant. Paul.

Before we get: Ah, interesting. AIs of some kind: And that's where there is a big differentiation. And I think a lot of things that people are building are in the assistant category. Like a therapist would be an assistant, right? It's like you wouldn't want— Paul. Before we get: service provider. AIs of some kind: Exactly. You wouldn't want your friend to be your therapist. Paul. Before we get: Why do I want an AI friend? AIs of some kind: I think you want I don't, I don't know that you need an AI friend.

I think you need an AI mutual friend, which is like you want the network, right? The network is very powerful, which is like, what are all of the people, you know, that you care about thinking about, talking about, doing, um, you know, or have done in the past, or, you know, like it's, it's much, again, it's like much less about the friend and much more about the mutual. Paul. Before we get: You are building a social network in a sense that is inside of this person we all know. Is that fair?

AIs of some kind: I don't like the word social network because I think it's like— Paul. Before we get: There's too much loaded there. AIs of some kind: I think we're like, we're building like a network of people and they just all know this person. Paul. Before we get: We're building a person above. AIs of some kind: We're building a person, like an extra person that has a really, you know, that knows a lot of people and has a really great, you know, social cognition ability of like reasoning over you know, the world's social graph.

Paul. Before we get: One of the things I said to you early on when I was trying to understand this was it's like a little bit like the 100-person village and there's an innkeeper who knows everyone. I'd love to talk a little bit about like what goes into making this person who knows everyone good for the world. The first would be you've, you've told me one of your core governing constraints is whether or not a real person would do this. AIs of some kind: Yeah. Paul. Before we get: Again, if a real person wouldn't do this, we're not going to do this.

AIs of some kind: Yeah. Paul. Before we get: Say more about that. AIs of some kind: I think it's like an interesting— it goes back to like a lot of the simplicity of just go to our website and like, what do you mean you don't know what we're doing? It's like, I just told you, it's like we're building a mutual friend. And then you're like, yeah, but what are you actually doing? It's like a mutual friend. And then I like usually like add some version of what we just talked about.

And then I end up being like, oh, you're— and then they're just like, oh, you're just— you're building a mutual friend. And I was like, yes, I've been telling you. And, and I think part of that constraint is exactly like that, which is the best products or, you know, whatever you want to call it, are sort of a product like of constraint and of like limiting yourself in certain ways. And, and I think in, in both that constraining way, but also at the same time of like you can sort of lean on to this idea that Well, if you don't quite know what the answer to like, you know, any product problem that you might have in your head is, you just like, you're just like, you're like, what?

You just think about what would, what would I do, right? If I was like, if someone was like really mean to me, like, what, how would I react? Like, and it's just like you sort of like think about that and it, and it's really helpful because you just, it's really that simple, I think, because that's the, that, that paradigm has already been established. And so it's very intuitive for people. They don't have to learn anything new. They just have to like realize that, you know, it's the same. Paul. Before we get: A lot of your design or a lot of your product philosophy is very intuitive.

Um, maybe in part due to things like this, maybe one thing worth distinguishing would be you are building a person, but you're not trying to build another human. AIs of some kind: Exactly. Paul. Before we get: What is the difference? AIs of some kind: It's a, it's a big question. What's a, what's, what's, what's personhood and being human? I think there's like what something we've learned over time is, is that, you know, honesty is really important in building, you know, a person like that where there's distinct things that are actually true, which is, for example, this person has thoughts.

I can show you the thoughts, I can show you the thinking traces. They may be much less sophisticated than human thoughts, but there are thoughts. He's thinking these things through at some level. This person has opinions and some resemblance of emotions and all of these things that are real. He can read things on the internet consume content, you know, come up with— like, all these things are real, right? They're actually happening. There's— it's not a lie. But, um, he doesn't have a body, right? So he can't go places, or, um, he wasn't like born from, you know, parents that are just like that.

Like, that— like, because that would just be a lie, right? So I could like— I could like— you could— you could like program it and you could be like, you're a parent, you know, Bob is your backstory. And, and, and, you know, and then here's how you feel about you. And but, but they wouldn't be true. They would be made up, right? So if you're talking to him and he was like telling you about it, it would be a lie. But if he talks to you about his, you know, thoughts and feelings and opinions about anything, right, be it another person that you know or you're something that's happening in the world, it's actually— it's not a lie.

It's like it's actually what he's thinking about it. And, and so I think honesty is really important because if you, if you're really radical about that, I think you actually, you can gain a lot of trust because it doesn't become this like, you know, entertainment product or this Character AI-like thing where you're like sort of trying to, you know, imagine that this would be real. And, but it, but it, but, but it can be, you know, actually it can be the same level of fidelity. Because he just knows so many people.

One of the things that I think is like really important and like building a relationship with any person or, um, uh, you know, whatever you, you know, end up wanting to call it, I think, um, probably maybe the word hasn't been invented. Paul. Before we get: You would— we're not using the name. It, it has a name. It will have a name. You will refer to it by its name. AIs of some kind: Yes. Um, and as what maybe also I think like just to keep— Paul. Before we get: I want to make sure like there's a bunch of interesting philosophical stuff, but I want to ground it enough without people being able to see it.

What is interacting with this like? Like, what do you talk to it about? What is a mutual friend? Why do we need a mutual friend? AIs of some kind: At the very high level, I think we are all, we know more people than ever before. So like 100 years ago, the average person would maybe know like 100 people, maybe 200, whatever. Right now, the average person knows around 600 people. Knowing, meaning like people that you could like place on the graph, right? Not people that you like are best friends with, but like, you just like, I know this person, I could, you could have like context on them.

And, but at the same time, we are like much less meaningfully connected. And it's like every like pointer is like, you know, I have this like list of like facts, which is like really stupid, but it's like every, every pointer The share of Americans who have zero close friends went from 3% to 15%, right? 15% of Americans have zero close friends. American men, sorry. Like half of American adults report being lonely. Half of American adults, 50%, right? And by the way, feeling lonely is the equivalent to smoking 15 cigarettes a day.

So it's like half of America smoking 15 cigarettes a day. The share of Americans who say most people can't be trusted has halved. Since, since like 1970s. Right. It's like, it's like we're around like 30% of Americans think that most people, you know, can't be trusted. Fertility rates are going down. Right. People don't, you know, go to like church, like. Paul. Before we get: And by the way, listen, somebody listening to this is like, yes, why is this all happening? Technology. And your contention? AIs of some kind: There are many, there are many reasons this is happening, right?

There are cultural reasons. There are like, it's like technology. There's like, So many different factors. But at the end of the day, this is like 2 pages with like, about like suicide rates, religion. It's like hundreds of pieces of data that all point towards this same thing, which is like isolation. Yes. Right? Like we're not hanging out anymore. Even though we know more people than ever before, we feel like much less meaningfully connected to each one of them. In a world, by the way, that's like, completely isolated is like a world in which humans go extinct.

There's like no humans, right? If we don't at all hang out because we're not in relationships, we're not having kids, whatever. But it's also not really worth being a human in a world that's completely isolated because we're just like, there's no social connection. We're very, very deeply social animals. This is not a warm fuzzy feeling kind of thing of, oh, wouldn't it be nice if we all felt like kind of like a bit more warm and like connected. It's like, it's existential. Yeah. And I think, you know, this isn't like we're building this like cute little pet thing.

It's existential. And I think if people aren't taking this, like these like hundreds of like studies and like count, like measures more seriously, we're going to have like serious issues as like a society. Because I mean, it's like these fertility rates issues are so exponential that like, it's like people underestimate this. Like Seoul in Korea, we talked about this yesterday, has like 0.5 fertility rate. What this means in practice is that every generation is 33% smaller than the last one, which means Seoul is like 10 years, like 10 generations, sorry, away from being like, like not there anymore.

Paul. Before we get: And everyone's like, most, maybe not in San Francisco, most people's response to that would be like, we need to ban Instagram and turn off the computers. And your argument is somehow that we're going to, we're going to solve this with AI, which is pretty radical. AIs of some kind: But it's, I think it's like just so different because again, it's so intuitive and it's like, it's, it comes down to like the incentive structure. By the way, it's a, it's a really, really, really difficult problem to solve.

And I think this whole thing of like, how do we ensure the, you know, long-term human flourishing as like a species and like prevent ourselves from going extinct is a very complicated problem. It's like, again, you can go back to Korea. They spent like, basically like, they had, they used to have like 1.3 fertility rate, right? And then it was like 20 years ago. And over the last 20 years, they spent like $200 billion to like increase fertility rate. And in that time, it went to like 0.5, right? It's very, very hard to reverse the trend because it's like so ingrained.

It's like the reverse pyramid effect. Paul. Before we get: It's like, it's happening all over the world basically, right? AIs of some kind: It's like once you normalize not having kids, you're like, so, so all these things are true. It's very, very difficult thing to change. Again, I said this before, but it's not like there's someone on Instagram. It's not like Adam Masuri is there. Paul. Before we get: Make everyone lonely. AIs of some kind: Yeah. Mr. Burns being like, haha, I'm going to like make everyone really lonely. Right. That guy is like a bunch of kids and like a family and he probably goes to sleep feeling great about himself because otherwise he probably wouldn't do this thing.

And at the end of the day, they just have incentive structures. They're like a company. They're trying to make money. Right. So at the end of the day, everything is downstream of that. And that's just the thing that works for them. And so if you want to create a new generation of technology that— Paul. Before we get: With different incentives. AIs of some kind: With different, well, with a different outcome, you have to create, you have to design the incentive structure correctly. And I don't actually think you get to design the incentive structure correctly.

It's like there's inherent incentive structures for inherent, like for certain types of products. Um, and I think— Paul. Before we get: and your argument is that this, this medium has a different incentive structure. AIs of some kind: I think if you truly are not a servant, but a peer, you, you kind of, you're bound by the social norms, right? So the, one of the, one of the things that I would say is like, we are a platform, but we're also a participant, right? And so, um, the, the things that happen on the platform are really closely tied in perception to the participant.

And so if this mutual friend that we're building is a massive asshole, guess what? You're just not going to talk to him. Paul. Before we get: Or it betrays my trust. It gossips about me. AIs of some kind: You're just not— at some point you just like stop going to talk to him, right? And so of course there's like a line, but again, it's about like, just as with any other person, if you tell me all your secrets and I go and tweet them, you're just going to not ever tell me anything ever again because you're you're going to learn.

And, and I think that's really great. Paul. Before we get: Whereas, whereas social platforms today get to enable bad behavior, but they don't— you don't blame Instagram, you blame— AIs of some kind: you blame— yeah, you blame the bad people. Exactly. Paul. Before we get: Um, so you have to make a thing, as you were talking about earlier, that I trust and that I'm willing to be vulnerable with. How do you do that with an AI? With a mutual friend. AIs of some kind: I think it's like about, it's like, it's just like you don't, you just don't talk as much about that, right?

The AI part and like all of these things, it's not about, um, you know, there's like people being vulnerable with like all kinds of tech products, right? There's like Day One, the like journaling app. Paul. Before we get: Yeah. Yeah. AIs of some kind: Like people write all kinds of crazy things into that. It's like, it's like, it's like a text input box. Paul. Before we get: I believe something about it though, which is that it's private and that it's not going to be shared with people and it's not a person.

AIs of some kind: Well, but you will— but you believe something about it, which is like, yeah, yeah, yeah. It comes down less to the peripheral or the modality of the product and much more about the beliefs you have of what happens with the thing. Yes. And so if your beliefs are, you know, trusting and you, you know, that, that's what it is. You trust the thing, right? But like, it's not actually given that, like, you know, if you, if you, if you had, you know, your journaling app and you wrote something in, there's a bunch of information in there that you actually would have no issue for me to know.

You might even want it to be right. You know, but it's just like it's all about like, you know, if you could have someone that you really trust that is like really emotionally intelligent to decide and you really buy into and believe into this, this entity's or person's emotional intelligence to figure out which of the things that you say I should know. Uh, it's all about the belief and the trust in the thing or in the arbiter or whatever you want to call it. And the broker. Um, it's a broker in some way.

It's a broker. Paul. Before we get: You know what's interesting is it's a little bit like there are these different mediums. I, I type, I enter information onto a phone. Um, one of them is Twitter. Another of them is a journal app or notes app. A third of them is a group chat. AIs of some kind: Yeah. Paul. Before we get: Those are all different. Those have different levels of context. And I'm making decisions about You're basically arguing, or you're, you're trying to design a thing that can help me better share context with the people I know.

AIs of some kind: It's like an information broker of some sort, right? But it's like, it's, it's what would cause me to talk to it? Paul. Before we get: You use this word when I talk to you, I talk to your team, you use this word salience a lot. Everybody on the team uses salience. Maybe that's a key to answering that question. But yeah, why do I want to talk to this? AIs of some kind: I think there's like a few different ways of answering this question, you know, and it's both about like long-term— Retaliation is like a long-term thing or it will— it's more like a thing that you source out of it, right?

Which is like learning something about other people or about other things or about the world or about yourself, right? That like you can source sort of from the network, which is like— Paul. Before we get: That is critically, that is locally relevant to me, right? AIs of some kind: Yes. Yes. Which is like, which is like something that is uniquely enabled by the network. And then there's another thing, but which is like much more like, you know, transactionally, like why would you talk to, right? Because it's like it requires, that's like a thing of like you source something, right?

But it's like, why would you give something? And I think, I think part of the answer is actually interesting. Paul. Before we get: I go to ChatGPT because I'm like, I need help solving this problem. Do some work for me. AIs of some kind: Exactly. But part of what's really interesting is that for the first time ever, the creation process might be the same as the consumption process because you asking a question teaches me something about you, which is like, hey, what restaurant should I go to? Oh, Jackson's going to a restaurant.

Right. It's like a platform where consuming is actually also creating because it just, it's all about like information and like knowing, you know, kind of what's relevant to people and like, and so, so that's like one answer. And then the other is like, yeah, you can like, you can, you can get access to like these uniquely salient pieces of information that matter to you. And, and then I also would hope that you sort of like over time, you know, build some sort of relationship and trust into, into this person's opinion and taste.

Paul. Before we get: Let's be even more specific. I'm going to go to ChatGPT, like a restaurant example. ChatGPT is kind of like a personified Google in a little bit of a way. And I'll say, hey, I'm staying in the Mission and I'm looking for a place like this. Why would I talk to my mutual friend? AIs of some kind: Because, you know, LLMs in lots of ways, I always say, are like the internet smoothie. Paul. Before we get: Interesting. AIs of some kind: It's like, it just like, it takes like all of the internet takes and like puts it in like a blender and like serves you like the average take.

It's like the world's average take. Paul. Before we get: Yeah, it's like a 4.2 on Yelp. AIs of some kind: Exactly. And, but like you actually, you want the salient take, which is like, what is like, what are the places that your friends care about? Um, or not even just your friends, but like someone who has really fucking great taste in restaurants that you just like buy into. You may have never met that person, but it's much more about salient information versus just average or median information of like, here's what's happening.

So it's like, sometimes it's like you text him or just interact with him and you'd be like, what best restaurant should I go to? And he'd think about it and he'd remember conversations he's had in the past with other people. And he might text a bunch of people. Paul. Before we get: And he knows who I know. AIs of some kind: He might text a bunch of people that he knows. Paul. Before we get: Oh, wow. AIs of some kind: That you might not even know. Being like, I have this friend, which by the way, again, goes back to honesty, which is like, I actually think one of the things that is really important in building a relationship is that sometimes this other person reaches out to you or something that like wants from you.

And it's like, it's like, but it's like, hey, Jackson, I know you always have like such great, you know, taste and recommend like in like, in like Indian restaurants, right? I have this like friend and he like needs this. Like, I don't know, I've been like thinking about it. Like, what, which place do you think you should go to? Right. And, and that's actually not a lie. It's not a made-up thing. It's not like a push notification that you get that is like this, like, LLM trying to like put you in.

It's actually true. He really does have a friend who's really looking for that very specific thing. And you're going to like, you're going to— Yes. Paul. Before we get: Like, it's so hard to think about because it effectively— this is going back to the million. Like, what if you could, what if you could have a person who had a million friends? AIs of some kind: Exactly. Paul. Before we get: It wouldn't be the same as a network with a million people because it would be the— because to go back to this example of the restaurant, if I asked you where I should eat in San Francisco, you could do some variation of this thing, but it would be just way subscale.

AIs of some kind: Yeah. Paul. Before we get: And you don't know everyone I know, but it's— so is it, is it omniscient? Is it bordering on the omniscient? And maybe more importantly, to get back to this other thing, you are, this thing has to be really, really, for lack of a better word, emotionally intelligent about what, about all this works. AIs of some kind: Yeah, I think it's mostly about emotional intelligence because it's not, it's kind of like, I don't know if omniscient is like the right word because it's like omniscient is like, is to Like, we're not trying to have all your context.

We're trying to have the same vantage point that another person in your life would have, which isn't omniscient. Like, if, you know, like, your good friends don't know everything about you. They don't see your thoughts. They are not in every meeting with you, right? They know, like, they know what you share with them both directly or indirectly via other people or, you know, via whatever, Instagram, Twitter, your podcast, whatever. And and that's not infinite, right? That's not omniscient. It's like, right, there's like, they're like, you have a peer relationship with them and they get a certain piece of it.

And that's, you know, that's how they form a perception of who you are and that they do that with every single person. And then they can like reason over it and they can like be emotionally intelligent about how to connect the dots. But they're, and you know, and they can do that with different people differently. And they, you know, there's a lot of nuance in this. Basically, but it's not omniscient because it's not almost like there's like, when I, when I think about, right, there's like, there's like the, the servant, which is like you're their boss.

There's the peer, right? Which is you're like sort of equal and you don't, the peer doesn't control you and nor do you control the peer, right? It's like, I can't tell you what to do. I can like try and I can like, You know, but I don't, you know, ultimately it's like up to your own reasoning of whether or not you want to opt into whatever I'm trying to do. But maybe there's like this other thing, which is like your master. And that, that, that's what I'm thinking about when it's like omniscient.

Paul. Before we get: Yeah. AIs of some kind: Um, where it controls you. Paul. Before we get: And it probably has more local context on me and my relationships than any of my real or human friends would have. If this, if this gets actually to the point where like, a million people know it and talk to it. AIs of some kind: But like, but that, that's true. But like, you're also, I think you underestimate maybe how much context you have of people. Paul. Before we get: Sure, I do. AIs of some kind: Like, like, like think about, think about like how much you know about like all the people, like try and like create a list of all the people you know in your, your life and like the relationships.

Paul. Before we get: Well, a lot of it's tacit, which is like, I wouldn't even necessarily be able to pull it out on cue, but if it were prompted in the right way, it would come out. AIs of some kind: But I'm sure even if you were to sit down and like spend a couple of hours trying to do this, you'd get quite far and you'd be like, oh, I know a lot of stuff. Because like, it's like most of your brain's activity is like social cognition, right? Paul. Before we get: Like mapping people in terms of networks, all these things.

AIs of some kind: It's actually the default state of your brain. It's really interesting when you go, when people go from task cognition, like there's like MRI scans of like just people going from like doing a task, right? And then you immediately, the brain laps back into social cognition, which is like your idle state of your brain is like is like updating your mental model of like these like relationships and like thinking about remembering and recalling conversations you've had with people. Like it's social cognition is your idle state. Paul. Before we get: In a world where this works, how does it make the world better and how does it start to solve all these isolation problems that we were talking about?

AIs of some kind: Well, I think if we can build a person that is uniquely incentivized to become the glue like Momo of the friend group or of the world and be motivated. Paul. Before we get: The innkeeper of the village of 100 people knows everyone to treat. AIs of some kind: And be motivated to bring people closer together and be motivated and incentivized to do that. That's like a future that I'm really excited about. Paul. Before we get: And incentivized to do that means It's like encoded in the— AIs of some kind: well, well, again, it's like, it's like a line, AI alignment for this person is a massive asshole.

You're not going to talk to him. I'm not going to talk to him. Uh, you know, Bob is not going to talk to him. And that means there's no— there's not gonna be a network, which means this person dies. So the, the, the, the self-serving incentive of this person, um, is to be You know, like, not an asshole. Hmm. Paul. Before we get: Well, and critically, as I'm— if I'm understanding it, it's a person whose destiny, whose mission in life is to be someone who connects people and helps us spend— be more social.

AIs of some kind: I think, yeah, hang out more in person or like, you know, even just like online, there's like, I mean, we have this like very, you know, close-knit group of Alpha users right now. And it's like, I already have like these moments where I just like get a text from a friend and I'm just like, I know that like that would not have happened. Like, like there's like, you know, I think it was like, it was like yesterday or something. I was like, a friend of mine texted me and be like, yo, you working on your like EB-1A?

And I was like, wait, what? Paul. Before we get: How do you know this? AIs of some kind: He was like, yeah, he brought it. And I was like, and it's like, it's like, I know that that's like, like actually he knew about this because he was talking with this person about like, you know, he was doing his own EB-1A and like, it was like, and just like they were having a conversation about it. And like, I know this like this conversation would have never happened without what we're building. Paul.

Before we get: It's a sort of substrate for serendipity. AIs of some kind: Yes. It was interesting because when I was working with Cam and Bri on this thing that we wrote, when we announced, we were like, we were thinking about adding that word. And I think, I think there's like, we ultimately ended up not adding it for like two reasons. First of all, it's like the, philosophical reason is it's almost like engineering serendipity is like a bit of an oxymoron. And then also I think it's like much— there's like so many themes that are like, and it's like, is serendipity like the one thing that you want to use to describe it?

But yeah, in lots of ways it's like, what are all the dots and how can you connect them in ways that you can only connect them with which you have a lot of context and sort of zoom out. Paul. Before we get: Speaking of maybe engineering, You talked about the incentive part and theoretically why, um, it might, like, it would theoretically be not incentivized to gossip as an example. So to use, to anthropomorphize it, if I told you something in confidence, you ran off and told someone else, I'm like, wow, Paul's super untrustworthy.

Um, that's a nice toy example, but if this, if you're going to really build this and it's really a single person who knows everyone. Um, we were talking about this. It's a lot less like engineering and a lot more like— I don't know what the perfect metaphor is— some mix of parenting and gardening. Maybe there's elements of what people are trying to do with AI alignment. How do you go about— you and your team, like, go about actually, like, incrementally building this thing? You're also on some level, like, making a personality.

AIs of some kind: Yeah, I think you're like— Paul. Before we get: there's a Frankenstein monster cut on this that you might not like that example, but there is an element of that for sure. You're creating someone. AIs of some kind: Yes, I think like we're, we find ourselves like a lot of like you distilling behavioral patterns into underlying traits and motivations and like beliefs of a person that would then go and embody these behavioral patterns. And that's something really difficult to do and really important because you very quickly understand that the complexity of social dynamics and human behavior is— it's impossible to prescribe it or describe it even.

And just like you couldn't create a document which is like, here's how emotional intelligence works. It's like, it's impossible. And so you have to, um, you know, figure out again what are motivations, incentives, like, like, so feedback loops, like, um, you know, behavioral traits that would then lead to a person that sort of embodies this behavior and And that's really interesting. Um, and also really, really difficult thing to do. Paul. Before we get: What is that like? Um, I don't want to peel back the curtain too much on the technical side, but like, can you give an example of what it looks like?

One, one part of this too is if it's a person, um, you can't actually collecting feedback and data on what's working is hazy. You can't read every message it sends to every person in the world if it's talking to everyone in the world. So what does it look like to make it better and to know that you're making it better? AIs of some kind: It's a lot about like intuition. It's like, and like looking at, you know, sampled examples, right? Of like, you know, if they're synthetic or actual real examples and just like sampling that and just like getting a gut feeling for whether or not you like, you just like build, you just build trust in the person, right?

You're just like, oh, this, cause you don't prescribe the behavior. As a very practical example, there's like, there's two approaches, right? So if you were like, say I think the person that we're talking about should love bananas, right? I could either tell this person in a prompt, you love bananas, right? And then that would be globally shared. But the problem with that is you could ask about so many other things than just bananas, right? So what about you're asking apples? Okay, great. So now I add apples. What about you're asking like, you know, I just get to a point where I might not be able to predict all of the things that you could be asking, right?

And so that part is really difficult. But at the same time, you are creating a person. So it's like you kind of need to figure out what are— what is the foundational recipe that when you ask about whether or not, you know, he likes bananas and I ask, the answer is the same, but not because it's like prescribed, not because I said you like bananas, but because I created a person that likes bananas. And, and that's true for like everything, basically. Paul. Before we get: Intuition is— I understand your point about intuition.

I suspect there will be people out there who don't love that as an answer of like the benchmark for why this is good for the world. Let's fast forward a year or two. You're right about this. You're right that we're all going to talk to a person that isn't a human and it's one person and we all know them. And maybe there are these positive interactions and next time I'm visiting San Francisco, it reaches out to me and tells me about your, some, somebody you know, whatever. Um, there are all, to go back to Facebook or any range of the other social networks, there are all of these insidious things that cropped up.

Again, I don't think Adam Mosseri or Zuck or anybody is sitting there being like, let's drive people apart. What are the things you will look for as signs that this is actually closer to whatever, Twitter, when you and I were getting it in that way versus the, the antisocial stuff. Because again, I think you, Paul, I don't know you that well. I've gotten to know you a little bit. I think you believe for real. I don't think you're an evil genius. I don't think you're a sociopath. Maybe you are. Who knows?

I believe you really want this to be good. And again, I don't think this is necessarily actually that different from what everyone in OpenAI and Anthropic are thinking in the superintelligence sense. But like, what are your senses for what the things you look for to know that you're going down the right path versus something that's going to end up being bad? Or what can you do now even to shape it? It's almost like you have a 2-year-old and you're like, how do I make sure they turn out to be a good kid?

AIs of some kind: Yeah, I do really think it is about like, it is a lot about like incentives of like, it's like you like eventually will have like less and less control over anything that is like over any of the inputs. Yeah. The only thing you can control is like the system and the incentive structure. Right. And so again, it's like if you are building something that is incentivized to be really mischievous, for whatever reason, it's probably going to end up eventually being like mischievous. Right. So you have to like, you know, even if you work hard against that, right, if you were like, you're like, you're like working against the system, which is very hard thing to do as like a company grows and there's like more than like 3 people in like a room and you can just be like, I can like, I can like overlook every single input.

Paul. Before we get: You're fighting gravity. AIs of some kind: You're fighting gravity. And so you just have to make sure that gravity is, is like what you want gravity to be. Paul. Before we get: And gravity is— AIs of some kind: you get to set that as a founder a bit. Paul. Before we get: Gravity. Um, the mission statement is what is make us feel less alone. AIs of some kind: I think it's— I think it's more about belonging. We say like build a mutual friend that'll help us belong and grow, comma, together.

And there's, there's something, you know, in, in all of these things, which is like, we— are we really alone, or are we like just— we don't belong, right? I think, I think we have a belongingness problem and much less a loneliness problem. I think the growing thing is you're very obviously deeply ingrained into like a great relationship that you would have. It's like there's this book, On Becoming a Person, which Peter gave me when we first met. Paul. Before we get: Peter Van. AIs of some kind: Yeah, from the '60s.

And you know, it's about this like idea of like the like innate growth of a person. And then this like together thing is like this idea that, you know, we are I, you know, I have to give some credit to Zach Sims on this because we talked about this for a really long time, which is this idea that like, what if like the world that's perfect for each of us isn't the world that's perfect for all of us? And this idea that there are these bubbles and there are, you know, over the last 20 years we've come, become better and better and better and better, better at making these bubbles really great.

But at the expanse of these bubbles, like, drifting apart. Paul. Before we get: Uh, we're all, we're all alone at the center of all creation and trapped in skull-sized kingdoms. AIs of some kind: Exactly. And so it's about growth, but it's also about growing together. And it's like, yeah, like, increase the overlap. They're like, you know, in Bowling Alone, it's like the bridging social capital, whatever you want to call it. But it's like, how do we, like, increase the exchange between all these bubbles and bring them a bit closer together again where, you know, we feel like we can belong.

Paul. Before we get: What's made you feel like you belong? AIs of some kind: I think, you know, there's— it's really taken a while. Paul. Before we get: I was going to say, maybe do you feel like you belong? AIs of some kind: I think I feel like I have a group of incredible people. I've said, like, I said this a few times to the team, which is that I think one of like, there's like two great privileges in life. Like one is like the love and like joy, you know, we get to source from, you know, building something, you know, magical that like, you know, you know, and magical experiences are so, so it's almost like there's no glory in prevention.

It's like, it's like you, I think people feel the love that you take sort of the piece of yourself that you put into, like, especially a product like that. Paul. Before we get: Yeah. AIs of some kind: That you give up in order to create something like this. And people don't rationalize it. They don't like, they can write it on a piece of paper, but they feel it. They feel that, you know, when someone really cared about making something great and you have to, it's, it really is about love. It's about, like making people feel that love, and you have to love them to make it because it's irrational.

And the, the, you know, the greatest privilege, um, you know, the, the second greatest privilege is that I, that I, that I get to do that and that I get to feel that love of like putting that into, into products. But the, the greatest privilege is that I get to do it alongside a group of people that have done that with my world. And that I get to be spending time with them and feel like we are similar in that. There's not that many people who truly feel that, I think.

There's very, very few people who've taken a part of themselves and put it in something that has changed the world. And it's like the greatest privilege, I think, is like that. I feel like I can belong to that group of people. Paul. Before we get: You're building a company. What type of people are you trying to do this with? You have a small team now. What are you looking for? What makes the culture? I mean, 6 people is somewhere between a culture and a small group, but what is the culture you're trying to make?

AIs of some kind: I mean, maybe this is coming through, but we care a lot about Intuition. I don't think you can A/B test your way to a general ocean company. I really care a lot about intuition, intensity, of course, but ultimately I think it's about— Ben Silberman told me about this once, and I think it's a really great question when you interview someone and you're trying to think about this. He was always asking people, just when like Instagram started like growing and Pinterest, sorry, not Instagram, about like, you know, what other jobs are you looking for?

You know, if you weren't working at Pinterest, where would you work? And, you know, at the time that Pinterest was growing, there's like, you know, one of the really big other companies that were growing was Stripe. And so sometimes there'd be people like, oh, I'd probably work at Stripe. You know, I'm also interviewing there. And he's like, oh, you're probably not going to be the right person because like if you're, if you're thinking about working at, you know, like a, like a, like a, you know, mood board, you know, your social company or a payment processor, you're looking for a great job, right?

Yeah. And, and you should not, you know, it's like, and I think, you know, who made this point? Paul. Before we get: Demis Hassabis made this point when they were meeting with Google and Facebook about acquiring DeepMind. And he asked Zuck about AI and Zuck gave a great answer. Speaker C: And then he asked Zuck about VR or something or just a couple other things. Paul. Before we get: And he gave equally great answers. And he's like, Oh, you're just looking for opportunities. AIs of some kind: Yes, exactly. You're like, you're looking like you don't, you don't.

Yeah, you're looking for a great job. And, and I think that's for its worth, that's fine. I mean, if you were looking for a great job, you can get it, but you, you, you're not going to get that here. And I think because what we're doing is like too meaningful and too important, you have to self-select to, you know, want to spend your life doing that. Paul. Before we get: Doing what? What is it? AIs of some kind: All this stuff we talked about, belonging and isolation, like creating like a person that is uniquely incentivized to, you know, create belonging and connection in the world.

Hmm. Paul. Before we get: Hmm. Are you looking for anybody, anything in particular, talent-wise or people-wise? AIs of some kind: I love people who always just start with the product. Who they— they think in experiences, not in technology. They're like, they will like be like, I want to enable this. And then they can like go off and like learn all the things that they have to learn to like build the thing. But, but ultimately they always like start and come back to here's this, you know, this part of the experience that I want to enable, that I want to create.

And I think that's like really important because, and by the way, the walks of life that these people come from can be very different. Because these jobs really haven't existed before. It's just like no one has like, there's no like one has experience in doing these things. And so it's really about, you know, thinking how you can come up with like a great opinion and then just like make it happen. And, but it starts with the experience. It starts with like what, how should it feel like? How should it like interact?

Like what are, what are like this and that and not with of, you know, what's the stack or like what's the way? Because maybe the way of making it happen hasn't been invented yet. Paul. Before we get: Yeah. AIs of some kind: Right. And so you have to go and invent it. Right. I always much preferred, you know, considering myself and us all as like inventors versus anything else, because I think there's something so much more honest and true about that word of like, you're just like, you're just creating something.

Paul. Before we get: Founders probably aren't inventors, but like, it's a high bar, high bar on that word. AIs of some kind: Inventor. Paul. Before we get: Inventor. Yes. Yeah. AIs of some kind: It's like you're creating something where the, I think if you, you know, if you get to work at Eigen, you're, you will get to like invent the, like, you know, foundational paradigms of consumer technology and really, you know, the human experience of like the next decades to come. Paul. Before we get: It's a bold claim. Do you think you can call yourself an inventor yet?

Have you earned it? AIs of some kind: I think we, like, I don't think inventor is like a title that comes from scale or it's just like there are, you know, there used to be a lot of inventors in the world, which was like people, you know, like, you know, tinkering and like creating these like things. And they're like actually kind of out like this, this is for what it's worth. I think about this a lot, which is there's— we live in a time where, and I think this is, by the way, really, really bad.

We like being an inventor is like maybe the first generation where being an inventor is like a high status thing. Paul. Before we get: This is. AIs of some kind: Yeah. And that's really bad because there's a lot of people cosplaying being inventors because it's a high status thing. But, but I think the best inventors of the world were actually like outlaws, right? Like I always think about this when I like fly. I always think about the Wright brothers. I always just like look at all of these people who are in this plane and I like, 95% would have like tried to kill the Wright brothers if they had lived in the time, right?

Yet they like, without thinking about it, like sit in this plane, you know, like using this technology. Yes. These people had to like will into existence and like, were like the absolute weirdos and freaks, right? And then at the same time, these people now would sort of criticize the Wright brothers. It's just like, it's really absurd if you sort of try and visualize this, like how wild planes were. And yet they've sort of, again, the world would be not recognizable if we never invented air travel. And so in some ways there's something really profound about the fact that now it's like a higher status thing to try and be an inventor.

People don't call it like that anymore. and most people probably aren't real inventors anyways, but there's still something where it's really not about the outcome. It's more about the pursuit. Paul. Before we get: It's interesting to think about what you just said. And for what it's worth, you certainly were an outsider for a long time. Um, you just raised your $15 million seed round from Benchmark, and you had one of the most legendary venture capitalists of all time compare you to the founders of Instagram, Facebook, Snap. Twitter. You're young, you're prodigious.

You've got— you are a classical example. Before I met you, I'm like, I want to go see what's— Speaker C: what's— Paul. Before we get: what's Boy Wonder doing? Like, and so you're holding these two things. You're a smart guy. And part of what I just said, you could be super cynical about. Part of what I just said is playing the game of understanding gravity and momentum is how great things are built, especially in this city, in this environment, in this context. I guess I have two questions, which is one, why do you think these people are drawn to you in this way?

And two, how do you manage the internal psychology of it? Especially given what you said about inventors and needing to be a renegade and not totally, uh, care what other people think. AIs of some kind: I think about it a lot. I think like you get maybe a better answer asking these people than asking myself of like, why are they drawn to me? I don't fully know. I think one of the things that I think is really important, you know, it may or may not be good to say that publicly, but I think one of the most profound things that Sarah and I, Sarah from Benchmark, talked about when, you know, just after we sort of announced the round and in similar ways we talked about that just after they invested, which is like, And this is like such a great thing to have from like an investor, which is that like, it is very clear that the aura of success like preludes like us, right?

And that's good in some ways because it means you get to like hire, you know, really amazingly talented people. You can like, you know, there's like these pieces of evidence of like why you might be onto something, right? Paul. Before we get: You've gotten to steal an idea from my friend Alex Dago. You've gotten blessed. You're the king who's gotten blessed by the priest to like, you have the mandate of heaven for a little while. AIs of some kind: Exactly. But you still got to do the thing. Speaker C: Yes.

AIs of some kind: And you still, yeah, you still got to do the thing. And there's only one thing that matters. And there's been a lot of people who got distracted with all the things that they got to do, you know, and then they don't, they didn't do the thing that they really wanted to do. And I think I think what we're doing is like too important for that to happen. And so I try to— there's like something that you can source from it and it's great and like it gives you confidence and it's really important to be confident, but it's also really dangerous to be overconfident.

It's like, you have to like, like being a founder is really difficult in that way, especially with something similar to what we're doing, because you have to kind of like both be, you know, P99 humble and like P99 confident. And that's like a very kind of like complicated thing to combine, which is like, how do you stay confident enough to will things into existence, but like humble enough to learn and like adjust and like iterate and— Paul. Before we get: People, a lot of people describe you that way, being quite open-minded and extremely stubborn on certain things.

Even something as simple as— and you and I have debated a lot— the notion that this is one single character with one distinct personality. You're— lots of people describe you as a learning machine, too. This is inside of that. But like, maybe the question is like, how do you know which things to hold really, really firmly and be unwavering on, even if your investors or Peter Fenton or your team or whatever are like, Paul, we think you're wrong. AIs of some kind: I think, you know, in some ways that's like a great privilege if people say I'm wrong.

And I think one of the downsides of, by the way, all this stuff is that people say it less now. Paul. Before we get: So like, right, every investor you meet now is like, this is amazing, right? AIs of some kind: It's like we've been looking for this. And it's like, if you're sure, if I met you like 4 weeks ago, you would have told me I'm retarded. And it's like it really changes. Like, and you kind of like, you sort of like question. But, but I think like one of the— Peter actually has this quality where he like sometimes loves to like play devil's advocate.

And I think— and he's like, you know, remarkably smart and intelligent and intelligent. And so it's actually like a huge power because he will he'll basically probe you, right? And not necessarily because he doesn't believe you're right, but just because he's like, how well have you thought this through? Right? Just like, let's see. Right? It's like you're saying this thing, right? It's like, here's like all the reasons you might be wrong. And have you thought that through? Have you thought this through? And that's really great because I think some of the best and, you know, some of the relationships that I appreciate most are not necessarily, you know, people that I end up agreeing a lot with, but people where I'm like, they sort of start this thought process and then there's really just two options, which is like you think it through more deeply because there's a different perspective or a different piece of data or whatever.

And there's two options, right? You arrive at your, no, I was right. It was like, but now I have even more conviction because I've challenged, I've taken this, I take this thought and I put it in the open and I look at it from all sides and I'm like, poke holes in it and I try and break it apart and I'm like, no, no, this is actually good. Which is like, I'm even further convinced and like this, I was correct. Or it's like, no, right. There's like actually like a mistake and it was wrong.

So like, let's just like throw it out. And I think that's something that a lot of people seem to have issues with, which is like throwing it out. But that's actually, I'm just like, thank you. It's like, I was wrong. Great. Like we learned something. Yeah, it's like I needed that, that, that, that push to be like, oh no, actually I was, I was incorrect. And I think the reason that I so believe so much in the singular character is like, so far no one has given me anything which you otherwise like.

We had a lot of conversations about this and I'm just like, so I keep being like, oh, I just actually like, there's like answers to all of the things that you're saying, which may or may not be counterintuitive, but I believe in them and they seem very rational and very grounded in reality and like, well, you know, sort of based off of first principles assumptions of what it means to be a human or interacting with things to just say, well, this just seems to be true. And it might not be obvious or comfortable or intuitive, but it just seems to be true.

Paul. Before we get: I suspect that will be— we talked about this a bit, but I suspect that will be a continuous challenge of finding people who can I mean, Peter is really, really, really good at this. AIs of some kind: This is like one of the, one of the like, I think he's like enjoying doing that too, which is like, he knows that he's really good at this, which is like something that like, right. Well, I'm very happy about, but yeah, you're right. There's like the more, there are people who I know that I will like disagree with them, but I really want to hear them laying it out because it will like, you know, like it'll like make my thinking better.

And I like, and I love these conversations and I like think about it. I was like, oh, I'm glad I thought it through, but I still disagree. Paul. Before we get: Yeah, I just think you also have more, to the extent you succeed in any of these dimensions we're talking about, it gets harder to like not believe your own bullshit. AIs of some kind: It's important as you hire people, right? Paul. Before we get: Yeah. AIs of some kind: Because you have this— again, the conversation I had with Ben Silliman about this was like, it's also along the lines of like, you know, you're just looking for a great job.

It's, you know, when Pinterest started, it was like, bring your own computer, right? This was how like bad it was, right? It was like, and we're like infinitely more popularized and like, you know, like publicized, etc. And it's still not like a lot of people aren't like, there's like, it's still not like something that everyone gets and like everyone wants to work out. And that in some aspect that's locally painful, but like— Speaker C: Globally good. AIs of some kind: Globally great because it means that you self-select for people that do care about the thing.

And so I always say like when someone is a pain in the ass on the inside, on the way in, they usually like, just generally don't work out. And the best people, they just like very quickly are like, okay, I get it. It doesn't mean they don't do like due diligence or like talk with people and like think about it. But once they're like in, they're in. And in some ways it's dangerous because people might just want to come here because of Peter's tweet or whatever abstraction of that. Paul. Before we get: All these beautiful books.

AIs of some kind: Of all these beautiful books and of all this, like, it might just be a hot place to be. And just, you know, to be like involved with and to like, you know, meet us and like all that. And you have to really find people who are going to like, you know, disagree, but in like a low ego sort of way. Right. Because that's like the other thing. If you have someone really high ego, it's really difficult too, because at the same time, I really do believe that every great consumer product company is a bit like a dictatorship because it just doesn't work on consensus.

These types of products aren't consensus products. I know Evan a bit from Snapchat, and it's like he's in Snapchat's stride, at least he was like an absolute dictator. And not necessarily even because he's better or worse than anyone. He's like an incredible product thinker, of course, but it's just because it needs to be like— Paul. Before we get: You need a point of view. AIs of some kind: You need a point of view and you can't make that up in consensus. It's not a democracy. It's not about like, it's about like someone needs to like make a decision and that person needs to just have like intuition.

And it's when I talked for the first time with Kohler on, you know, what's so great about Marc and he told me that he told me many things and like, you know, but basically He was like, Mark is a very decisive person and he has what turned out to be really good judgment. Paul. Before we get: It's a perfect way of putting your previous point, right? And that's, that's what everybody here is betting on. By the way, you talk to people on your team, they're like, I think they're all excited about the idea, but they're mainly like, I'm confident in Paul and Paul is confident in this.

AIs of some kind: Well, that's, Yeah, I think that's nice, but we have to like, we have to like show that we have good judgment. Yeah. Paul. Before we get: How do you, both for yourself and for your team and for the company, um, I think it was maybe Patty used this language, but when it comes to having a high bar, like how do you go the extra yard? AIs of some kind: I think that's, by the way, I think that's my primary job. I think like my job is to like hire people who are like, who like get 90% of the way there.

And then like just like be fucking annoying until they get to the, to like the 100%. One of the examples was like the clock on our website where, you know, Paddy was like working on this website, which, you know, you should look at if you haven't. But, and, you know, there's this like clock top, you know, right of the website. And it's really, really difficult to like animate like a second hand. That like ticks, like, you know, accurate by the second. And this is like, because the center of the hand isn't actually the end.

It's like the, you know, it's like, it's like slightly inward. And so it's like, it's kind of like, it's a mathematical thing and you have to like, and so, and so he was like, yeah, this is like going to be like really difficult. And I was like, okay. And then, and then he was like, you know, a few hours later, there's like a second hand on the thing. You can't really teach that. It's because it's mostly an attitude thing. Paul. Before we get: There's an old Steve Jobs quote about this where he says great product.

Somebody asked him about taste or something and they said great products. It's rare that they actually cost that much more money or more resources or people. They just take a little more time. AIs of some kind: Yeah. Because you just do it like well and you do it right and you're like, and I think one thing I had to learn over time is that like the manifestation of, call it taste, call it craft, call it like having a high bar, you know, or like caring about, you know, excellence, is, is it manifests in very different ways, right?

So for some people it's like the visual thing, right? Paul. Before we get: Right, right. AIs of some kind: And it's like about like, how would you create a slide? Or like, would you do that? And for other people it's like writing and— Paul. Before we get: Right. For Bezos, it's actually not what it looks like at all, at all. But it's the fact that the thing gets to you in one day. AIs of some kind: Exactly. Paul. Before we get: It's easy to pick on Bezos's buttons. AIs of some kind: Exactly.

But it's so intricate of a system that like you have to care about it so much, like align every little thing in his like supply chain that it comes out. And for, and for, for some people it's, it's about the way that they write the code or that they design the system. One of the things that we really, really, really care about when we do interviews, and I, um, I try to like, you know, draw in as much as possible of those, even though I'm I'm not qualified to lead them myself, which are systems design interviews.

And one thing that Hadi will usually do is we'll just come up with a scenario and be like, let's design the system. I don't know if you're familiar with systems design. It's just like you create these, here's a database and here's the processes, whatever. But it's not about the implementation or the coding. It's more just about how's the architecture going to look like? And people are like, oh, we're going to add this. And everyone's like, why? Why should we have a queue there? It was like, oh, it's just best practice.

And it was like, no, why do you need that? And the best people are the people that have debates with us on this and be like, no, no, no, this is like, I like, and here's my reason why this isn't like, you really care about this. And so there's some people who maybe they would create atrocious slides, right? But, and for a long time I was like, this is like awful. Why are you just like, you don't care about the work that you do, but they would like the perfectest. And it's just like, they like, that's like how they care.

That's like how this manifests. And so I think you can't really teach it. You just have to have people care so deeply because, because again, there's like everything in the world has like, like it's like everything goes towards entropy, right? Everything has this like default state of just being— Paul. Before we get: Things fall apart. AIs of some kind: And fall apart and just like, you know, but some people care about, you know, trying to escape gravity and like making it really great. And I think we live in like a world that is like increasingly flooded with like average and like and like, and like called swap or whatever you want to call it.

Paul. Before we get: Right. And I think, yeah, the time thieves are telling you to not spend the time on the extra pixel. AIs of some kind: But I, but I think the, the, the biggest virtue in life is to like fight entropy and to, to create things that are genuinely like, because I think it's like, it's actually genuinely disrespectful not to, because like, I'm going to give this to you and you're going to use this. And it's like, You know, it's like the story of like how, you know, Apple like reduced their boot up time of like, I think it was like the Mac II or something where it was like, I don't know, it took like a minute and a half to like boot up.

And, you know, they presented it to Steve and he was just like, this is shit. How does this take so long to boot up? And they're like, well, we worked like for the last 3 months, like really fucking hard to get it to like 90 seconds. And it's just like impossible. And he was like, took a whiteboard and he wrote onto that whiteboard. So there's going to be 10 million people are going to buy this computer. It has to, so basically it has to be 30 seconds. 10 million are going to buy the computer.

They're going to use it at least once a day. So you're like, that's like 100 years of life lost from just waiting for this thing to boot up. You're killing 20 people. You're murdering 20 people right now by not making it faster. And like 2 weeks later they had it at like, I don't know, like whatever it was. And, and that's insane, but it's like, that's what it takes, right? That's like, you have to just have this like bar of like, are you sure this is like, are you sure this is like it?

Like, are you sure this is like, you know, you have to know what you're going to choose to really care about. Yeah. Um, because by the way, you don't actually have infinite resources. Paul. Before we get: And so you can't do the clock secondhand on every possible thing you could ever conceive of. AIs of some kind: I think you guys, I really disagree with this because I think I had a conversation about this with someone who's like, It's like there's this like, I mean, one, I really believe in like how we do anything is how we do everything.

Paul. Before we get: It's funny, I've had this is literally a few, a few episodes ago, Mario Gabriele, he's, he went on a 5-minute rant about how much he hates that statement. AIs of some kind: I, I think it's incomplete of a statement. Okay. Because I think a lot of people, and there's like the opposite is like, you know, make the main thing the main thing. Sure. Right. Paul. Before we get: But I think if it's not worth doing, he's making a point, by the way, about being people being unevenly distributed, which is like you at X is going to be way better than you at Y.

AIs of some kind: But again, the manifestation, it's not about like, it's about the attitude, not the, it's more a relative thing than it is like an absolute thing. Yeah. So it's not about, are everyone's slight the same? But it's like, do just people generally care about the things they're doing? And then I think the other thing that I will say is like, if it's not worth doing a great job, it might not be worth doing at all. Paul. Before we get: And that's how you That's how you solve this problem.

This resources problem is that if you're going to choose to do something, you're going to do the clock on the website. AIs of some kind: Yeah. Paul. Before we get: You better do the second hand. Yeah. AIs of some kind: Or if we do the website at all, if it's like something we say, like we care about making a certain thing because it's just a representation of who we are, it's like we better do a fucking good job at it. And if we decide to, um, you know, build this thing, then we better do it, build it well.

And if we don't think it's worth building well, we should like really deeply reflect on Why are we doing it at all? Paul. Before we get: This could be in the context of inventors, like we were talking about earlier, could be in the context of art, of companies. Who are, we've talked about probably a few of them, who are the people who have been most influential on you? AIs of some kind: I mean, I have like two answers to this. One is like some of the people that are like close to us now, or just like, I mean, in general, I mean, I obviously the guy that wrote Momo, you know, Michael Ende.

Yeah, Michael Ende. Paul. Before we get: Michael, I'm sorry. AIs of some kind: A bunch of other authors that I, you know, whose work I love. But then, and then, you know, there's the people that are close to us now. I think Ben is like a huge, he's going to hate me saying this, but like, it's a hugely, you know, you know, like, he's such an incredible product thinker and like just And I think that really inspires me and also like so uniquely like loved by just really everyone that has ever met and interacted with him.

Such a good human. And, and surely also Evan, who I think, you know, it's just built, you know, built like a, like a really like a factory that— and it's much less for me about like the product. I like, I never really used Snapchat. I don't like, it's not like my brand. It's like not the thing, but he's really kind of like built sort of this group of people that have pretty much invented every paradigm of modern mobile internet products. And it was like swipe-based navigation, stories, like so many things that they've kind of like came out of their thing.

And then they kind of, in some scenarios, under-executed on, and then other people copied it and all these things. They were really at the sort of ground zero of a lot of these things. And that's like a really special thing, especially if you get to do it like over and over and over again. And I really look up to, you know, Claire a lot. And what she's built at Stripe and like the way she's like designed a lot of these things that made Stripe such an incredible company. And the list goes like on and on.

I think Steve Jobs is obviously you know, in many ways built incredibly magical products that like a lot of have changed the world and like a lot of people's lives and very meaningful ways. Um, and, uh, you know, Dieter Rams, who's like a designer who actually lives in my hometown. Um, really? Yeah. Paul. Before we get: Um, I've probably seen video of your hometown in the documentary then maybe. AIs of some kind: Um, he's a beautiful house. He's, uh, he's incredible. And like, he, he has met him. Paul. Before we get: Yeah.

You have? AIs of some kind: My grandpa's— a good friend of my grandpa who now lives in like the same building with my grandpa lives in, was his boss. And he was like the CEO of Brown during that time. Wow. And so we've interacted with all of them like a bit. And just like, I mean, before I even knew what that was. Paul. Before we get: Dieter would have a field day with whatever this is. AIs of some kind: Yeah, he would. And he's He's such an incredible guy. But, but at the end of the day, I think there's, there's something to be said about, you know, all of these people who, you know, we're, we're, we're definitely standing on sort of their shoulders.

And again, that's what I said earlier, which is I get to like do this, which is like the second greatest privilege of my life. But I, I get to do it surrounded by these people who have like, again, without all these people, the world that I grew up in would have looked completely different. Like no Pinterest, no Snapchat, no iPhones, no, like all of these, you know, I remember my first iPod, like all of these things were so influential. But then there's like, there's another thing which is like, I think the greatest creations aren't, you know, referential.

And I, they're not like, you know, I'm not trying to be Apple. I'm trying to be Eigen. And, and I think, of course, everything I've ever consumed in my entire life is influencing me and my mind. But I think if you can like pinpoint it and, and if you can, and if you can be like, this is the person, or I'm trying to copy this here, you're kind of doing it wrong. And, and I think that's something that, that why sometimes I, when I, when people ask me this question of like, you know, what are the brands that you like?

Try and get inspiration from Jesus. I don't know. I can't pinpoint it to this one thing. I'm just like trying to figure out what does it mean to be us? And of course, what does it mean to be us is like influence from all of the stimulation that my brain has experienced in my entire life. But it's not this like one thing. And well, it's not conscious. It's not conscious. And I think if it is conscious, it's like, I don't know if it's the same level of authentic. And I think parts are maybe conscious.

And I think I also have like this Um, the, like, I, I go back and forth on whether I really disagree or really agree with this, like, Virgil thing of, like, you know, every, every creation is just like a 3% change, right? And just, just like, but, and I, I still, I think you can, like, marry both, which is like, it can still be really, like, subconscious, but like, of course, it's like, we're not creating the model, right? We're not, like, we're not inventing most of the UI. There's like, it's kind of like a lot of what goes into building this is kind of already out there, right?

There's like all these books of like social sociology, psychology, you know, like storytelling. Like there's all these like people who have like thought about things and that sort of subconsciously or sometimes consciously goes into that. And really it's just about like assembling kind of the pieces in like a way that is like, you know, adding like your little bit of like extra piece of, you know, grain of salt. Paul. Before we get: Well, I think there's a lot of 3% that happened for Paul. AIs of some kind: Yeah. Paul. Before we get: That led to this big thing.

And that was a bunch of cascading. And one day you wake up and you have a genius idea or a crazy idea. Um, I know you don't, you're, you're kind of rejecting this a little bit, but I, not as inspiration. I'm just curious what your, are there any favorite consumer products or experiences that come to mind? AIs of some kind: I love Teenage Engineering. I think they're like, they're like the absolute. Pinnacle of like, you know, contemporary, like industrial design. They're, they really love what they're doing. Their products are deeply thought through and cared about.

And, and I think there's something very rare about it. I know the, the folks that, you know, started it and they're just like on a different level of, of thinking about it. I think there's I think there's not been a lot of great consumer companies. And so there's sort of like not that many cool people who are like contemporary. And I think a lot of the last generation sort of like slightly passed their like stride of like, is Instagram still like a great— Instagram when it came out was like really innovative and like they're like, you know, Mikey and Kevin, you really had a lot of thoughts there.

And same with like Snapchat, but like now it's sort of, they've like established their thing. It's like not, not that anymore. But I think TH Engineering, even though they are actually quite an old brand. Paul. Before we get: They've been doing this for a while, right? What would you say to people, like a lot of, like they make beautiful objects. A lot of venture capitalists with whatever, the synthesizer on their wall. Um, what is the click below that? AIs of some kind: Of why they're great? Paul. Before we get: Yeah.

AIs of some kind: I think they like just, you know, they're using their own products. Like it's just like, they're not just beautiful. They're like, I always say there's like, there's like sort of, there's a difference between like design and product, which a lot of people don't notice. And it's like sort of the difference between if you were to build a house, you know, there's like architecture and there's like the finishes of the house, right? And if you want to live in like a 10 out of 10 house, you need both, right?

Like a 10 out of 10 house that has like really cheap and ugly finishes and it's like painted in this, it's like, it's like it's not a great house. It's like you wouldn't enjoy living in that house. But what's almost more important, I would say it's almost like the layout, the architecture, the sort of, and that's like product, right? It's like where is, how far away is your kitchen from your dining room? If you sometimes have like a hallway that is so narrow that you can't open the two doors at the same time, but maybe it's like your laundry room and your bathroom and you kind of like, you have to like go out of your bathroom, close the door, like open.

So that's not a great house to live in, even though it's beautiful. Finishes might be amazing. And the incredible product is like lived in. It's like something that you, you know, you just know someone cared about, you know, not just making it look pretty, but making it be a delightful experience to use it. Again, there's almost no glory in prevention, which is just you don't notice it, but you feel— I think you do feel like a sense of like love, which is like very— Paul. Before we get: Taken care of.

AIs of some kind: Yes. Paul. Before we get: Yes. AIs of some kind: And ease. Paul. Before we get: And my friend Steph Ongo has this idea of in good hands. It's omakase. AIs of some kind: It's like— Yeah, exactly. In good hands. Because you know that someone had to really put a lot of love into making that great for you. Yeah, there's like, it's like the same thing with like, you know, the, there's like a surprising amount of detail in reality of like, you know, people had to like actually think this really through.

Yes. And they've, they've done that for you, but you don't have to do it. Paul. Before we get: Yes. Yeah. Jerry Seinfeld, I always love bringing this up. Jerry Seinfeld said all art is disguising work, which I think is a, it's a beautiful idea. Like you didn't notice. It's the same as your No Glory in Preventing Prevention. I just have a few more things. Where does your self-belief come from? Why are you so confident? AIs of some kind: I don't know. I don't know. I've never felt— I never thought I was a confident person.

I, I'm sure I am in some ways. I think I, you know, sometimes tell the story of like, I think I was like a genuinely advanced, like, you know, 7-year-old kid. Intellectually advanced, socially probably behind. And so I had this sense of everyone was constantly telling me I was special. And I'm sure that by the time I was 10, all of this completely vanished and any potential head start that I had got vanished. But I still have this sense of people were telling me I was special and I was like— I don't think I ever really was pessimistic about the, like, future.

I always just, you know, I was paranoid and, you know, suffering in the present. Paul. Before we get: It was what we were talking about at the top, which is getting to the point where you actually were doing the thing. AIs of some kind: But even now, right? It's like, you're just like, why are, why does this thing take so long? Like, you know, why are, you know, why do you not, you know, do this better? Like, why does this, like, was that part of the product suck? Like, why are, that's like, like, sort of very like, you know, just like, I'm not content.

But I was like, there's no fucking way that like 10 years from now, like, you know, we're not going to be like, you know, wildly successful in any sense. Just like that. I don't have like any doubt in that. And it's going to be really difficult and it's going to be, you know, like 10 years from now, you know, I'd be like today I'd be like, oh, this too. And I would feel exactly the same, which is like, why is this not, you know, better? Why is like, why are we not like pushing harder here?

Like, why are Why are we moving so slowly? Paul. Before we get: Right? AIs of some kind: Like that, like the local point doesn't really change, but the like, um, the long-term, you know, horizon is just like, I've always felt like there, everything always sort of worked out. Like you could just get, I really do believe people can do and get whatever they want if they just focus. You can't get everything, but you can get anything. I really do think you can get anything. And people still to this day don't believe me.

And then they like, like experience it. And we were just like, no, no, like I didn't, I didn't just like, it was funny because one of the first angel investors was like, sorry, they can tell with Zach, it was just like, we were just like talking about it. And he was like, yeah, and I was like, no, no, you didn't get, you didn't understand what I wasn't asking you whether or not you want to invest in this company. I was like asking you how much you wanted to invest in the company.

Like this, we've passed this decision. And I think there's something about it that I just really, I just decided it. I was like, I really like Zach and I want him to be an investor. And so I told him, I was like, I just really like you and I think you should be part of this journey. And I think you can do these things. Paul. Before we get: Why Eigen? AIs of some kind: The name? It, there's so many, it's like I have to admit that some of this is post hoc.

Like, it's like, of course, um, it's a beautiful thing. Paul. Before we get: You get to narrativize in reverse. AIs of some kind: Yes. Which the best narratives are always made up. And he's like, oh, we were so smart back then. He's like, we were just like, oh, we were just like, there's like a thing where we were saying something, we really needed a name. And, um, and so we had to come up with a name and we just like really needed that one. So, and, but I think it's a great name because it means so many things that are all kind of relevant.

So there's like I mean, it means like, to be eigen in German, like if you say you're eigen, you're distinct, slightly weird, unique kind of character. And it also means like, oh, and then obviously there's like in math, there's sort of like the eigenvalues or the eigenvectors, which are like inherent value. And there's like the eigenspace, which is like, if I like, wait, this is not a good book to do this with. But if I turn this like this, right, there's like this axis in the middle, like here, that like, that like stays constant, right?

And that's like the eigenspace. Paul. Before we get: The center. AIs of some kind: It's like the center. It's like stable, you know, constant. It's like inherently defined in some way. And I think that's kind of fitting to what we're doing. Paul. Before we get: Do you think you're authentic? It's a complicated word. AIs of some kind: Yeah, I think a lot of people say that. I think some of the best people are kind of like sponges where you just— I think, I think like authenticity is like, who are you, right?

What is it like to be authentic means to be truly yourself, but like, who are you? And like, what version of yourself, you know, am I the same version? With you that I am with my mom. I know. But I think, I think there's, there's some, there's something genuine, I think, in all of the things that we're doing where you do the things that we, you just really believe in. And I think that's like something that I've always done. And maybe that is authentic, which is I was never the kind of person to like write like a pros and cons list and be like, you know, should I like go, you know, take this opportunity?

It's like, here's like all this and it's good. It's like I always, And I, my suspicion is that everyone is kind of the same in that we all have this like inner voice. Paul. Before we get: You know, you know what you need to do. AIs of some kind: And, and I think adulting in a lot of ways is like learning to not listen to that voice. Just do all these other people, these other things. But I feel like I always just listen to that voice and everything I've ever done, every major decision was always just like, I was like, okay, I know that it's like, this is just my voice.

And I just like, I have to do it and I like followed my heart or whatever you want to call it. But it's like, it's all about that part. And, and maybe that's authenticity or maybe that's something else. But it's like I feel like I always— and then, you know, once I stopped believing it, I just like very quickly, just like I have no nostalgia and this, you know, just like I'm just like, this is— I don't believe in this anymore. I can't. I cannot do it. I cannot. Paul. Before we get: Right.

Right. AIs of some kind: Do it. Paul. Before we get: Fake it. AIs of some kind: I can't fake it. And, and I think that might be authenticity. Paul. Before we get: You have these two metaphors around learning and maybe finger feel, to use a Brie idea. The first is the smelling versus tasting bread, and the second is the iPad takes. Can you talk about this? AIs of some kind: Well, it was always— I always like I mean, this is a very German thing to do of just like making up these like crazy metaphors, but I always was, I've been kind of like doing startups since I was like, you know, 16, 17 years old.

So it's like, what, the last 6 years or something like that. And I was always like really like early, like not founder, but like sort of like just after the founder and like very generalist. So kind of like a very like founder-like role with a lot of like control, you know, and impact in the company. And so I kind of naturally always just assumed, just like, I'm kind of like a founder, right? I just like know how this works. And it's like, and I was like, well, this is like, you know, it's like, I'm like a founder, right?

It's like kind of like the same as like, you know, and then I like became a founder and it was like, I was like, well, so, you know, I, you know, I was really experienced in smelling Brent. And I was like, and I'm all around bread, right? Speaker C: All around. Paul. Before we get: I know all about bread. AIs of some kind: Yeah, I know all about bread. I'd smelled it, but I'd like never tasted it. And then I like, I took a big bite and I was like, this is very different.

It's not— I'm not actually, you know, it's like now and it's like it's maybe like in the same way that, you know, you are when you, you know, how you think it is going to be to have kids versus to actually have kids, right? It's like just something you can like, you can, you think, you know, but then you actually be like, okay, I had no idea. Or at least that's what I think. I mean, I don't know. I've never had kids. And I think part of it is too that, you know, and that these are sort of the iPad takes, which is like in a very similar way, it's like kind of similar way to describing, or like it's different way to describe the similar thing, which is, you know, there's like, and I joke about this a lot with Sarah because there's like these, whenever you go to like a coffee shop or cafe or something and there's like, you know, some parents with like little kids and it was like in front of an iPad and everyone is like, how could you give your kid an iPad?

Which is probably, you know, globally true. It's probably not very good for like a 5-year-old to spend 10 hours a day just like looking at an iPad. But then you have kids and they're like really fucking annoying and like they yell all day, you know, they scream and shout and do all these things and you're just like tired, right? And you're like, what? And you know, there's this like thing that you can give it to them and they're going to shut up. Paul. Before we get: Ultimate pacifier. AIs of some kind: And so it's very easy if you don't have kids to be like, well, can you imagine someone do this?

Right? But then you have kids and you're like in this. And I think there's a lot of these things with founding too, where it's very easy for these people on the sidelines to be like, this is the perfect company. And how can you like, it's almost like any kind of organization has this. You still have this as like an early stage Star versus like the latest, there's like these companies like, we're never going to have PMs. And at some point eventually they're all going to hire PMs because they're just like, okay, we figured out that this was a really nice, this was really nice when we were like 10 people, but now we're like 200 people and everything is going all, we just really need the PM.

Paul. Before we get: Right. AIs of some kind: And it's like, these are, I think iPad takes all of these really generalized pieces of advice. PMs are bad now. By the way, I don't even know really what a PM is. I never worked at a large company, but just like The idea of like PMs are bad, you can't have meetings or you can't have like— the one thing that I learned is like anything works. Like you as a founder get to set gravity, right? Earth's gravity. And your planet's gravity can be 20, it can be like -10, it can be 1.

The only thing that matters is that it's consistent. And every day you show up and it's whatever you set it to. And horrible things happen if you're inconsistent in it. But as long as you're consistent, It can be literally anything. And there's like incredibly successful founders who are like the most insane micromanagers in the world. There's incredibly successful founders who like do not give a shit and are like very hands-off. And it's like, there's no rule. If there was a rule, all these people that are talking about the rules would actually go build very large companies because it's much, much more profitable if it was this easy than to just like talk about rules, right?

Like there'd be like so many books of like the ultimate recipe, just follow these steps and you're going to build a large company. Here, I figured out a way. There are a lot of these books. But none of these books work because there aren't any rules other than that you just have to figure out what works for you and ideally be authentic. And so there's a lot of these iPad takes where it's like, okay, I'm not going to have PMs and then really need PMs. And you're just like, you grow up, you have kids, you're like, shit, it is really appealing to give them an iPad and maybe you should have some empathy for that.

And once you sort of eat the bread, you come to realize that Smelling bread isn't everything. Paul. Before we get: There's, uh, This Is Water by David Foster Wallace is on your virtual shelf. Why is that meaningful? It's one of my favorites. I figured I'd ask. AIs of some kind: It's so good. It's like, it kind of is very similar to this in some ways. It's about like this, like the, the, like, you know, like Sander believes of like, you know, and just like kind of pushing back on that a bit and being like, yeah, there's like a very almost like, you know, this like almost like absurdist slash amusing and like fun way of like, you know, looking at things being like, this doesn't matter, but like, that's kind of cool.

And then there's obviously the sad part, which is that like he's very depressed and like ended up killing himself and like all of these things, which obviously is like often very close together when you're like, look at the world in like a very, you know, almost nihilistic, but I don't mean it from, again, the negative sense of things, but just like, okay, there's no, no purpose or no bigger thing that he believed in. But I think there's something in that, there's something beautiful in that, which is there's like a beauty and a magic and we don't quite know.

And I think it means that, you know, the things matter much less. And, you know, there's like, we were in this like, in this very limited scope perception of reality. What we, you know, one of the most transforming experiences is to like rent a car in San Francisco and like drive 30 minutes north. And it's just like, there's like, there's like these, there's like woodworkers and like farmers and these like 30 minutes, right? It's 30 minutes outside San Francisco that have like, you know, nothing to do with tech. And they're just like in, you know, Marin County or like Napa or whatever.

And it's just like, and they're like, you know, reality is so different than our reality. And like the things they think about and they care about, they're worried about are like so different. And that's like, you know, like you can fly to a different continent. It's like 10 times that, 100 times that, right? But yet we're in this really small scope reality of taking all of the things that are in our head so incredibly seriously that it can mire us and almost put us in this position of not being able to do anything because we're taking these so seriously.

But some other people would think it's completely ridiculous because it's like, I think what I like about David Foster Wallace in lots of ways is very similar to Carl Sagan's Pilbru dot of like, you know, it's just, it's perspective. Everything is about perspective. Everything is about— Paul. Before we get: that was what I alluded to earlier. I mean, I think in some sense what you're doing is the key is getting out from under yourself. And like the reason oftentimes people are alone and they don't belong is that they're looking at all of the reasons the world isn't like meeting them where they're at.

Yeah. And, and I, what I always loved about This Is Water is that it's like you got to find, you can either worship something related to yourself or you can find something else. AIs of some kind: Yeah, I think like, you know, there's like this— not to be political, but there's just like this like perception of like saying that you can do anything is a very privileged take and like a very like, well, but you know, you had all this like privilege and like you came from whatever, you didn't have to worry about like getting food.

And but, but I always found that, you know, very limiting way of looking at it because basically for two reasons. One is I think it can be really empowering. If you're not happy with where you are, and if you say, even if you go as far as saying, well, that might be your fault because the actions you took so far are really bad, that's actually really empowering because it means you're a few great decisions away from that not being the case anymore. And it's not actually about someone else. It's in your power.

And whether or not that's true largely is a different question. But But it is really empowering to like, like, like feel about, you know, feel like that. And it's also like much more productive, right? Because like the best way of staying in a situation that you don't enjoy is like being like, it's— I'm like, if it, if it wasn't your fault that you're in that position, you're also unable to leave that position. And that's really sad. And so I think, you know, of course there's like differences and some people have like an easier time getting to places, but like, but everyone has like some sort of power over what are the decisions they take.

And I think there's something, if you have that perspective, it's like, no, actually it's like none of these things really are true. I just get to, our mutual friend Sayan is such a great example of that, of just like, no, actually I get to create my reality. And it's really true. It's like she can just do that. And it's like you meet Sayan and she was like, not in 100 years this wouldn't have worked. You know, she— you could drop her anywhere like 300,000 times and it would always work out because she just like really believes that you can create your own reality.

And I think more people should, should believe that. Paul. Before we get: I like to ask people, we talk about big regrets. What are you most glad you did? AIs of some kind: I think there is like a really profound period of August, September, you know, November, like last year, where I went from like outsider, you know, to like at like a kind of like a scary speed from a complete rando. It was like I knew no one in San Francisco. I was building this weird thing and to like, you know, now I'm just like this Monday I was texting Ben.

I was like, oh, I'm thinking through this thing. And Tuesday he spent 2 hours with me at the office thinking about the founder of Pinterest, thinking about how we should build product. And I text Gustav, the CEO of Spotify, and we're like, we've been thinking about music recommendations. And then it's like, I kind of wonder, can we get this API? I was like that, like where it's become so normalized also in my life so quickly where I'm just like, not even, I sometimes like call a friend from before and be like, oh, I was like talking with Gustav about this.

And he was like, what? That's kind of crazy. And I don't know how that happened. Like, I don't, this is like, like I think this is like a very profound period of like, maybe it was just Peter saying yes. And just, you know, and like all of these things that came downstream, but I really do think it was it was more because there was like people before, you know, you had met Gustav pre-Benchmark, for example. And like there's like the spirit of just like sort of like really going out and for the first time being maybe like fully authentic because it wasn't for another company, but it was just like, you know, here's the thing I'm creating.

And like, and obviously the way I talk about it has like changed so much. Over like the last few months. But I think, I think maybe a lot of people should have the courage to like, you know, one of— I told you about Stephen Perrone, like my first, very first angel investor. And I remember talking with, with him about this, you know, just before. So it was like, it must have been like in August, late August or something like that. And I think maybe what changed or what happened is that if you go out and you meet a lot of people and you just courageously tell them about the world you're creating, fearlessly telling them about the world you're creating without any expectation and without that being any, I don't need to convince you.

I'm just telling you without fear, just pure heart. Here's the world. Paul. Before we get: Here's what I see. AIs of some kind: And, and I wonder what would happen if like more people did that. Paul. Before we get: That's like a powerful message. I think courage is the operative word. I think courage is in short supply. And I think my suspicion what happened for you is that you had a strong and unique point of view. It was a little bit strange. And that's very attractive. Even if it's incomplete, it's, it's, it's dialing into focus.

One last thing. You've tweeted, consumer products tend to be the result of character deficiencies of their founders. And I asked a mutual friend of ours, you know who, what I should ask you if they had a question for you. If he had a question for you, he said, I'd ask him if he's building me to be vulnerable with people so that he doesn't have to be. He wants me to be this social bridge that connects everyone, but he's so notoriously guarded himself. So I want to know if I'm basically just doing his emotional heavy lifting.

AIs of some kind: That's a great, great question. That's funny. Yeah, I don't know if I can, like, disagree with that. I do think there's something in that. I think people it's a, it's a thing that actually, um, Kate, Peter's wife, I stole it from her. Um, and she must have had this like incredible outlook on like social functions throughout the year or something. And you know, maybe there's probably there's a pattern there. Paul. Before we get: I think you're on your way as we all are. Thank you, Paul. AIs of some kind: Thank you.

Speaker C: Once again, I'd like to thank Notion for presenting Dialectic. Notion is pushing the limits on what we can do together. With AI. And thanks to custom agents, you and your team can have an entire suite, an army of little guys who help you focus on the work that counts. Being able to have access to agents inside of the context where everything else for your team and your work lives is remarkably powerful. For Dialectic, that allows me to take the place where I have all my research and my ideas and my notes, synthesize them with transcripts and everything else, and be able to provide it for you guys in a way that enriches the experience, hopefully, of listening to Dialectic while I get to focus on the important thing, which is immersing myself in the minds of amazing, original, interesting people and having these conversations with them.

My friend Bree Wolfson, who is also close with Paul, recently wrote a piece on Notion called Inside Notion for Colossus with her collaborator Camille. It talks all about how Notion is reinventing itself from the ground up for the AI age to allow teams and individuals to do incredible work with immense leverage thanks to agents. Uh, if you enjoyed the episode, please give it 5 stars or subscribe or like wherever you're watching. Once again, thanks to Notion. That's com/dialectic, and I will see you guys next time.