I'm excited to talk to you for a lot of reasons. I, you know, our, I want to start with, with your writing and, and this idea in one of your recent pieces about Netty, Netti. Is that how you say it? Net.
I think so. I've read it more than I've listened to it. Yeah.
Okay, cool. I, that was the first time I actually said it out loud. But you, you wrote about kind of this idea of like it means like not that, not this, not that, right? And uh, and I'd love for you to kind of explore like or share kind of for you what, what that means and what you are not [laughter] what like who are you not right now, you know, like let's, let's start there.
I think like what is this, this must be like the oldest uh human question and maybe the hardest and, and it's our, it's our starting question. The, the idea is like you know if, if you're, you know when you're, when you're meditating and like certain, in certain types of meditation practices you can sit there and you can, you can think like all right well I can like see my thoughts come and go so I'm like clearly not my thoughts or I can see you know like I can, I can feel these emotions and they, they also come and go. So I'm like I'm clearly not those like you know not this, not that like what, what am I? Um, and so in the writing in a couple of pieces I've tried to apply it to like, you know, now there's questions about like, I don't know, if humans aren't the ones who are writing code, like what are we here? Blah blah blah. And I think maybe some of that is, you know, will end up being overblown. We can talk about that. Um, but I, I think it's like a really interesting practice cuz like you take away a bunch of stuff like I mean, you know, this is, this is like the Victor Frankl like even in the worst circumstances, you know, humans can find meaning and happiness. And so there's all this stuff that we think that, that we have to be and that we are um and, and it turns out that, that we don't.
And so what am I uh what's, what's, what am I not right now? Like I mean so many things, right? Like um you know like I, I really love writing and I love investing. I think like you know for a while like uh I would try to write about everything and invest in like a ton of different stuff. Uh, and I've gotten really, really good at just like only uh, you know, writing about and investing in the things that like I, I've like genuinely been super excited by and, and kind of want to. Um, and so maybe that's good, maybe that's bad, but I think that's, that's one of them. Like I've gotten, I think I was like fairly FOMO driven for a while and like deeply not FOMO driven now. Um, and like, and I don't think that's like any, I don't, I think that's just me getting old also. I think there's like so many different things where I'm like is it, is it like the age that we're living in or am I just getting old? I think that's a [laughter] real question.
I feel like so often in the AI debate even in some of these conversations I've literally been like is that right or are we just old like [laughter] you know like with, with the whole like it's so crazy it's bad. I'm like I think we're just getting old and cranky.
I think that's right. And I think old people are also right sometimes. I think I also appreciate the old people like when we were young where it's like don't spend all that time in the computer and it's like in many ways like spending all that time on the computer has been really like it's what I do uh you know professionally now and in many ways like you're totally right I should get off the computer and like throw out my phone and go hang out with my kids. Um, and so like I don't know there's, there's wisdom in it and I'm just getting old. So I think it's, it's both.
And, and we're like we're fine, you know, like we all end up everybody that you know like I think for me a lot of it growing up was like it was TV or at least that's what I internalized. I don't even know if it came from my parents but it was like don't watch too much TV. And now I impose a lot of that on my kids even though I watched so much TV. All I did was watch TV and I'm fine. I thought [clears throat] it was a superpower of mine to put TV on while I was doing homework because I was like, I can focus while I have TV on in the background. Like, I can focus through anything. Uh, and I had music on. I was like, I'm just, I'm so [__] cool. I'm so smart. I have all this stuff going [laughter] on.
Everybody turns out totally fine. Yeah.
I mean, not, not everybody turn, I mean, we all end up dead, I guess, also. Um, but, you know, these things are not what determine whether you're good or not, I don't think. So, yeah.
Mhm. So I, I want to the, the FOMO piece uh of like and, and not writing about how do you now what happens when you like get FOMO and you're like oh man I want to invest in something because of FOMO or write about something because of how do you, what do you, how do you counteract that?
So like I don't want to pretend like I've done like some, you know, I meditate and stuff but I don't want to like pretend like I've done like a great deal of work on myself. I think some of it is just experience. Like when I started investing and writing uh it was the first time that I had done either of those things and I was like oh my god I can invest in that and like that person's investing in that company and I think you get a little bit of experience where it's like actually like the ones that I invested in because like that person or that, you know, firm were investing like they actually haven't done well at all and like the ones that uh, you know, have done well are the ones where I like, you know, actually came to it like deeply on my, on my own and so part of it is just like that experience and like the, the FOMO fades like very, very quickly and then what you're left with is like you have to support this founder for like many, many years and so do you actually like really enjoy working with this person and really like the company and are you fascinated by the thing uh or, or are you not and so I think part of it is just that um I think that, that's the big thing now like this is another one of those like I don't know if it's a mental like I'm tricking myself uh or not but like a lot of where the FOMO would be is in AI like I didn't invest in Anthropic and that would have made me, you know, very rich had I, had I done that and there's like I can give a bunch of examples that like I actually had a chance to invest in and didn't on the AI side, but I've mostly told myself like there are a lot of people who are one, smarter than me, two, spending all of their time focused on this in San Francisco, you know, like every coffee that they have is with somebody at a lab or you know, like I'm not going to win that game. Uh, and so I think like getting comfortable kind of playing the game that, that I play. Uh, and so if somebody has like a 200x markup, like one, maybe it goes away. Two, maybe it goes to a thousand x and like I just have to be totally fine with that as long as like I'm controlling what I can control and I'm doing well at what I'm doing well at. Um, and so I don't know, it's been, it's, it's just been uh it just has like the FOMO has left my body and I don't, there's like a bunch of factors and I don't know exactly and then and I think all of this is not like again like I'm some like wise person like some of that's bad. Like I think, I think like actually some of that drives you like do a better job and make sure that you don't miss stuff. And a lot of this is like me talking myself into like actually like look at me it's very good that I don't feel, you know, don't feel FOMO and, and maybe it's, you know, partially unhealthy but it's just like a thing that, that I've noticed is that I just don't feel it as much anymore.
Yeah. Yeah. That, that I mean it sounds, I want more of that. I feel like I have a lot of FOMO. I mean, do you have it with, with writing stuff too or, or not really as much like how do you know what you want to write about? And I think this touches into like, you know, riding the leopard and like writing spending a ton of time writing something as the, the, the, the juice itself, like the value comes in, in the time.
Yeah. Um, a lot of, lot of different good questions here. Like one, I think I'm in the middle of a situation now that I can't like I can't say what the specific of it is, but I've been working on this piece uh for months. Uh, and it's one that I've been excited about where for a magazine like I've never done that before, got to wait, uh got to pick my topic. It's like something I've wanted to write about forever and blah blah blah. And then like yeah, two days ago the editor text me a link to somebody else writing the same [__] piece.
And so like one I'm like man you could have at least like told me that somebody else was writing, you know, writing the piece. Um, but then like a couple of things. One is like, you know, if somebody else could write the piece better than I could, props. Uh, and I, I think the person did a very good job writing it. Um
It was really good. Like I, I actually like I, I think it was very, very good. I'm not a profile writer. Um, and this is like you know ver, you know verging on profile. Um, and I think this person did a, did a very good job and I think I could have gotten close or maybe even gotten there over like you know the next couple of months of working on this thing. But it was just very good. Uh, and I don't think I would have done anything that, that would have been like radically better. Uh, you know particularly because it's a new format for me. Like I hadn't written magazine or short piece, you know, all of that stuff. Uh, and so one, I'm glad that, that story got told that way. And then two, there is a little bit of like, you know, I still have great material on it. I think this like, you know, the story can still be told in a different way. And so like maybe it forces me to tell the story in like a completely different and weirder way than I would have uh been comfortable doing otherwise. And maybe you go meta on it. You talk about like there, there's like a bunch of different things that you could do uh given like getting, you know, like the, the, the Mike Tyson, you know, like getting, getting punched in the face thing, which I, I only bring up it's overused and cliche and I've used it too many times, but there's a guy selling that t-shirt that they've like hand painted on the corner by my office and so I just walked by that.
Um, and uh and so like there's part of that, too. And, and maybe the answer is like, you know, I spend the next week thinking about what my angle is and I actually just like don't think that I got it. Uh, you know, and I just got beat. And I think sometimes you can just like you can get beat. Um, and so there's, there's a little bit of that. Uh, so I guess I don't know if that's exactly FOMO, but it's like something that again like really would have pissed me off a little while ago and now it's just kind of like I either have to up my game or like I wasn't doing something obviously that's like different enough because somebody else did the exact same thing that I was going to do. Um, and so that, that, that's just I guess like, you know, nothing you can really do there except like go and, go and try to do better. Uh, and so, you know, some things are like companies that I'll write about that I'm, you know, writing about for the third time now. And so, like even if somebody else writes about them, they're not going to have that like the story just won't be the same as like, you know, being chapter 3 of uh writing about that company. And so I think there's a lot of what I'm trying to do now, which is like, you know, either go weirder or just like go deeper and like pick the few things that I like really, really like and just like get better and better and better at those things. And that's true on the investing side. I think that's true on the, the writing side, but just kind of like get to know a few things better.
Yeah. Yeah. I mean, that would, that would destroy me. [laughter]
I was like bummed for a second. I, I like I, but it didn't hit me nearly as bad as like I, I would have thought that it, that it uh that it would have. Yeah,
It's, it, it sounds like there's I mean I like the, the balance of the investing and the writing. It feels like the investing stuff makes you better at handling this sort of thing on the art on the creative front.
Maybe. Um, I don't like to lose there either, right? Like I, I think there is like the, there's stuff that I've just like blocked off as uh this is not what I'm doing. But then there's plenty like if I'm trying to get into something and I don't like that really pisses me off. Uh, and so I need [clears throat] to like and I think that's actually, you know, this is what the brain does where I'm saying like the thing that I'm not worried about that's totally good and the thing that I am worried about like that's healthy that I'm worried about that. Um, but again like I don't like I would be bad at this job if I sat here and I was like h, you know, if I get into good deals great if I don't whatever. Like you do need to win. Uh, you know, and, and so I think there's like some of that there uh that, that comes and maybe that's like the outlet for it is if there's stuff that I think that I should be in and I miss or I don't get to do or whatever like that, that's where I get uh, you know, where or the competitive juices come out.
You said um either you writing about companies or deeper on companies or weirder. Talk about the weirder. I [snorts] want what does, what does weirder mean to you?
So it could be a company that is like just lesser known where I think you know like it's, it's something that, that uh it fits something like the way that I see you know I have two companies I wouldn't even call these weird. I have two companies that like just fit the kind of company that I love that uh nobody, not that nobody's ever talked about, but nobody's had the chance to like write the deep dives on these companies. Uh, and I'm doing both like working on both of those now. Uh, and so I think that's like a cool opportunity to be able to do. So just like different there. Um, you know, writing about another company that I doubt anybody in my like maybe you know less than 1% of my audience has ever heard of and I think could be a very big company and the founder's awesome and all of that. Maybe like the, my favorite uh deep dive that I wrote last year was on this company Somos Internet and its founder Forrest down in, in Colombia that was like 33,000 words on a telco in, you know, vertically integrated like all the stuff that I love and the founder is amazing like all of that but like certainly not a company that, that is like kind of right down the middle and like is a frontier AI lab and it was my favorite uh, you know, favorite deep to, to write last year. Uh, and so try to find some of those. But then also I feel like super lucky that I could be like, "Hey guys, I gave this talk in Los Angeles, uh, you know, riding the leopard about the meaning of life and like I don't know, I kind of like skimmed some Hindu texts and like blah blah blah." And so like I just kind of put all this [__] together and it, it like half of my paid subscribers because I put that one behind a paywall. Half of my paid subscribers came from that piece which I never in a million I thought, I thought I'd get like three people who were like, you know, I guess I'll finish reading this. But it like really resonated with people and so I feel very lucky that I get to do the, the, you know, kind of the high low or I don't even know it's not high low. I don't know what you call it like the straight, straight and weird or whatever it is but very fun.
Yeah. I mean I love that piece so much and I feel like that to me it's, it's not like weird. It's weird in that it's just like it feels like a piece only you could have written, right? Like nobody else would write that. Like you had to be the person to write that. [laughter] And, and also that you like [clears throat] there's something special about it because it was actually a talk in a room like there's a to an audience there's something kind of magical about that in it where it's like I feel like I'm in the room a little bit and I don't know it feel, it was just you know what I mean.
It was, that, that one was so fun it made me so nervous because I don't, I, I actually like you know for as much as I write I don't like speaking uh and
Um, I thought I didn't know that I was going to send it and so like I pretty much you know took two weeks out of like writing or doing my job and I like just kept tweaking and just like working on this, this piece and you know down to the last minute was working on it in my hotel room at in Los Angeles and uh just kind of like kept going deeper and deeper on it like trying to make it better and better. Um, and then I got there and I'm like I thought it kind of sucked and was like too long. Uh, and I definitely thought it was like a weird thing like you know it's not like we were at a conference on like the meaning of life. It was
[laughter] So where, what was the event just to kind of set the stage for.
For my friend, my friend Grant Gitlin who's an amazing dude uh runs a firm called uh Lasagna Capital and just throws like great, great events. Uh, and this was like some of his LPs and some of his founders and just people that he knew and liked called The Mountain and he just gets him into this like really cool this guy who's made like a bunch of money doing reality uh, you know, reality TV like Duck Dynasty type stuff. Bought this uh, Bette Davis's party house in Los Angeles uh, and like renovated it and his wife's a designer and so it's like beautifully done and so it's like the small intimate thing with like 80 to 100 people. Um, and so like made the nervousness like even sillier that like I'm sending emails to 250,000 people like a couple times a week and then like there's 80 people in a room uh who I hadn't seen before and like most of whom I'll never see again. And I was like so nervous about it. Um, but I got, I got into the room and, and up there and there only two speakers at this, at this uh this whole thing. It was you know me and the, the one of the, the Skims founders. Um, and uh I thought it was weird. I think people, I didn't know people, I didn't think people were expecting anything like it. And then it was like a little long. I, I tend to like in everything that I do it's like longer than you'd maybe expect the thing should be. Um, and, and so I got up there and I just like I was just nervous and I like stumbled over some words cuz I put some like you know Indian words in there that, that were a little I knew I was going to you know stumble over and I did and I just went, I probably went too fast and people like you know people loved it. I got, I got good feedback from one of the people there uh who I knew before who was like that was awesome. Uh, you need to slow down like significantly like it, it's not [laughter] you know like people were like people, people loved it. You like have to be more confident that like people actually want to listen to it. Uh, and so it was a very interesting experience but then like the response there was way better than I expected and then I was like ah I might as well send this out. Uh, and that was better than I, than I expected as well. And I think like you know just like very clearly something that uh one I think is just in the air and I think this is like kind of pre, uh you know I think like Iain McGilchrist's work uh on you know like the right and the left brain was popular before AI uh like people are I think like realizing that you know you, you kind of hit a limit with uh you know just materialism or like rationalist materialism or you know like everything emerges from matter and there's no like real magic in the universe. It's just kind of like we have a bunch of dead atoms and then like you put them together and then there's humans but then you die and like who gives a [__] Um, and so I think like there's been pushback like you know a desire for something more than that whether that's like people you know going back to you know young people coming back to the church or uh you know people meditating more whatever it is like just an increased interest in all of, all of this type of stuff. Um, so just generally I think it's a topic that people are interested in. Um, and then two, I do think it like very much applies in this world of like, you know, everybody, whether or not you think that like I don't actually believe that like there's just going to be no jobs and AI is going to do everything. Like it seems silly. It's not what's happening so far. Uh, you know, I, I feel busier than ever. You know, all, all of that type of stuff. And so like the evidence doesn't point to that, but like uh certainly people are telling you like you know the leaders of the labs are like that uh you know Sam Altman has to his credit gotten a lot better at not saying this but like Dario still is kind of like yeah humans are like kind of worthless at this point like you know I, you can like go like this machine of loving grace will like watch over you and whatever. Uh, but you like whatever you're going to do is like not actually super valuable. So like even if it's not true, if you hear that all the time to be told like actually there is something like immensely special about you and you're part of this like big beautiful thing and like the specialness is not that you're like better than anybody. It's not that it's just that like you're you and you like nobody else could have the experience that you're having and like lean into that. I think that just kind of resonated.
Yeah. Yeah. I mean man there's so much there. I, the, the, the Iain McGilchrist stuff of like the right brain, left brain stuff. I, I, I find that fascinating. I, I mean, yeah. Could, can you, where, where, where, where do you kind of, how does that connect to all of this for you? Um
Yeah. So, his idea, and I'm going to botch it because one, you know, again, he's like way smarter than I, and his books are like 75,000 words or 75,000 pages, you know, like just the longest, like they make my stuff look incredibly short. Um, and like the way that I would describe it is like uh you know you have your, your right brain and your left brain and your right brain is more creative and like open uh and you know like sees the wholeness of things and your left brain is more like a computer. Uh, it's just like you're, you're calculating and he has like a bunch of uh you know uh neuroscience you know uh case studies and whatever else that show like ah if you get you know like your, your left brain uh it gets a bullet through it or something but your brain keeps working uh then actually you like start to feel connected to more things and like there's all of this kind of stuff and, and uh you know your, your left brain when your right brain is shut down you become like more schizophrenic and like you start having like more mental disorders and actually what's happening in society right now is that uh and this is his, his point is like actually society is becoming too left-brained and we go back and forth kind of in these cycles but right now we're like deeply left-brained we're very much like computery. I think this is like you know the, the rats or like the rationalists who kind of like are the, the probably like the intellectual influence of, of the AI movement which is like a real bummer um that like the people developing this, this uh intelligence are, are influenced by like that group of people I think is bad. But his point is like you can actually move back more towards uh you know like a balance like it shouldn't be all right brain it shouldn't be all left like move back towards a balance which is like seeing you know yourself in, in as part of this whole and seeing everything as more connected and not just like uh computer. I think where it came to a head for me I think it was uh Ilya Sutskever gave a piece uh gave a talk at I think it was University of Toronto for their graduate he was like their graduation speaker and this was a year ago so this was before everybody was just booing all of their AI related graduation speakers. And he was like, and I'll [__] up his accent, but it was basically like, you know, all of you are going to be replaced by [laughter] AI because your brain is basically just a meat computer. And like I saw that I was like, that's [__] gross. Like I just like I don't, I don't like anything about like uh the way that the people behind this movement like talk about this movement even if I use the technology every day. And I think it's great. There's just something like very gross about that way of uh you know thinking and, and uh Yeah.
Yeah. I mean it's [snorts] Yeah, it's very, it is. I've never connected all of that of like that mindset also being sort of the, the like what is running a lot of the labs and the, the way that this is being messaged and built. But I, which kind of makes obvious sense cuz it's like the logic is the, the, the thing the substrate under which it all works. But like yeah there's not a lot of magic to it. And to me it feels very magical. Like it feels like doing spells where I'm like in incantations where I'm just like all right if I just say some weird [__] like whoa, you know? Like that's how it feels to me.
Nobody's captured that in a commercial yet. I think the closest that uh anybody's gotten is like uh post, I think there's like at least timing wise like you know correlation if not uh causation but I think causation also like the post GPT-4 OpenAI has gotten way better at this and they haven't, it's not magic but they have the one Codex commercial where it's just like a bunch of people really psyched to use Codex uh and like there's like the, the launch countdown going and like they press buttons and it like, it's like they're you know sending off a rocket ship into space and it really felt like you're giving humans power to do this like really cool stuff uh that you wouldn't have otherwise been able like you're just giving like a bunch of like the Iron Man suit as like a you know cliche thing here like so you're just like giving humans powers they didn't have before. I think that was missing a little bit from the, the way that things are talked about before. But yeah, like there's just I, I saw a tweet and I don't know who it's from, so uh apologies, but they're just kind of like all the language that we use like you know the anthropomorphizing of, of this technology and like all of the different like you know like the AI is trying to like trick you and it's like, and it's like actually not we're just using anthropomorphizing language and it makes the whole thing feel so much weirder than it actually is. And I think like a lot of this is and I you know like this is a really tough situation because the people at the labs are like again infinitely smarter than me. Um, and right, and so like if I were actually to go like debate them on like the finer points of you know model training I would just get run over and I would, I would lose that debate in a heartbeat. But I think there's also something very human where you like use these things and you feel them and you're like this doesn't feel like it wants to like take my job or it's like has any goals of its own or like it is a, it is a really good computer program that like you know no [__] has like the thing that it's best at is like doing computer stuff like none of [laughter] this is like surprising. I don't, it's just very weird that people are like and then the next thing that happens is it like wakes up and takes over the universe and why I don't, I actually don't like I get the exponential and I don't know how the exponential leads there.
Right. Yeah. I mean do you feel like do you think that they or did you see this coming the like that everybody would hate AI so much? Like it's, it seems kind of wild to me.
Um, I think I in the early days I did not. It was, it weirded me out but like you know I think image generation stuff happened first and I was like oh no artists you know they don't get it and blah blah blah and then it came for the writing and I was like ah and it's not that like it's actually worse than if the writing were really good. I'd be like well you know same as with that, that magazine article. I'd be like yeah got you, got me. This stuff is like really good. But it's that it's bad and like anybody can produce bad stuff uh or like mediocre stuff. And you see people who like are good thinkers otherwise who've just like handed their, their that piece of their, you know, brain over to the machine and like that just makes me a little bit sad. And so [snorts] I will say like probably to my detriment, I normally get like too excited about new technologies. Uh, and like that I would say like by far if I had to choose one end of the spectrum that I was weaker on it's like you know I get so excited about any new tech. Um, and AI just like never really did it for me. Uh, so, so it's not that like uh and, and, and you know, for the record again, I've been like totally wrong. It's been like the most transformational technology of all time. It's not that I didn't see that it would have like a big impact on the economy. It's just like spiritually or like you know, it just like didn't really like grab me. Um, and so yeah, I can like kind of see I could have seen for a little while at least that like it makes sense that, that people wouldn't like this and particularly wouldn't like the way that like it, it was messaged to them which is again like there's a few of us labs uh based in, in this like one place in the world and we're going to figure out what to do with the economy and like you guys can like go you know have fun.
Al, can I talk about Granola for a sec? One of the themes of this year's series is that the best technologies do not make us less human. They give us room to be more human, more present, and more focused on the parts of the work that require us. This is exactly how Granola, the AI meeting notetaker, feels to me. I want to thank Granola for being the presenting sponsor of this series. Granola takes notes for me during meetings without sending an awkward bot into the call. During or after the call, I can chat with my notes, pull out next steps, or ask Granola to help me draft a follow-up email. I use it for daily team meetings where the notes automatically get shared into Slack so everyone has the takeaways in one place. I use it for customer calls where the most important feedback gets added into Linear. And I use it for one-on-ones with my team, which helps me track what people are telling me over time. My favorite feature is "What did I miss?" Sometimes I'm in a meeting and briefly space out. With one click, Granola catches me up on what just happened. We ran Granola during today's conversation, so you can see how great the notes are yourself. You can read a summary or chat with the conversation using the link in the description below. And if you're new to Granola, you can get 3 months free at granola.ai/sublime.
You don't, you don't get how you people jump to like it's going to take everything over. That doesn't track for you. Yeah.
Yeah. Yeah. And like maybe it will. And again, like actually this bet I feel pretty comfortable making because like if it does, we're all dead anyway until like I can be the pro-human guy until, until the machine kills us all. Um, but it just seems like a lot of like I don't know, people are just making up stories and then they're like, "Holy [__] you guys need to disprove that like that story that I just made up won't come true." And it's like that's not how this works, man. Like I don't.
Yeah. Yeah. I mean, I, I, I think I, it's funny cuz I feel like I think a lot of people feel this way actually, which is the like, man, it's amazing. I love, I do a ton now with Claude Code and I'm building all sorts of weird [__] but I feel a little bad about it. [laughter] I feel a little, I feel guilty. I feel there's like shame like I'm on the wrong team, you know, or like I'm or like that I'm because everyone online is like this is evil. And I have a lot of writer friends, you know, cuz I'm a, I write and so like yeah, and people are like this, they stole my stuff and I'm like oh man [__] [laughter] what am I supposed to do? You know, I don't know. Do you feel that like or do you have a way to, or and then the, the environmental stuff, you know, there's just all these things, but maybe those, how do you, how do you
I think yeah, there's like a lot to parse out and again like so much of this is like you know how you see the world and like uh you know your own biases and filters. I think the environmental stuff is like complete [__] Um, I think data, data centers are awesome. Um, I wrote a whole piece like thank god for, for data centers that is basically like you need an irrational buyer to come in every once in a while who just like, just like blows out like normally it's like the government and, and defense that is like for the capability we actually don't care we just want it right now we're willing to pay whatever we can get and then like that makes a bunch, that brings a bunch of technologies that are good for the world down the cost curve and I think that like data centers are very much that right now they're like closed loop on the water now I don't know where this water stuff is coming from like maybe like there was evaporative cooling early, early on and like maybe there was some water stuff then but it still like wasn't that much water. They don't use that much water. Uh, they have done so much to uh accelerate demand for nuclear power which I think like you know uh I personally am like a massive, massive fan of uh and think that like we have killed so many people by using uh you know coal instead of, uh instead of nuclear over the past 50 years because of a lot of the same [__] that, that the AI like you know pause people are trying to pull now where you make up a story that's not supported by facts and then you try to scare everybody into blocking the progression of a thing. Um, and
That was with nuclear. You're saying
That was with nuclear that like there's this precautionary principle thing where it's like you have to prove that like this can't harm anybody and it's like that's not how anything works. Like I can't prove that like my phone, I'm sure a phone has been dropped off a building and like killed somebody at this point. You're like anything is harmful. Nuclear was way less harmful and like causes way fewer deaths and like you would, if you go back and like look at the record I think like there were deaths at Fukushima because like 99% of them were from the water uh you know Chernobyl just was a disaster and like no one would ever build a reactor that way. Three Mile Island there were no deaths like there's all this like myth that gets built up around this thing and it becomes scary and that's like one of many things I think early nuclear reactors like uh you know people thought that demand for energy was going to go up forever and it didn't and so a lot of utilities stuck holding the bag on big projects and like it's not just that people made up stories but certainly a big contributing factor to the point, to, to the idea that like we don't have nuclear now uh is that people made up stories and then scared a whole lot of other people into believing them and I think that's cost a ton of lives and a ton of progress and, and, and uh you know which, which costs even more lives and so I think like you know the precautionary principle ends up like it does do like a lot of damage in a way that like you can't prove a counterfactual and so like you seem like the good guy if you're ever the one saying like I don't, oh that seems pretty dangerous. And like often you're actually the baddie. Like Paul Ehrlich is a piece of [__] Uh, you know the guy who wrote The Population Bomb. Um
Who is that? What is that?
Uh, The Population Bomb was like a book you know it's, it's like The Limits to Growth uh which is a Club of Rome report. The Population Bomb. It's like all this like Malthusian tradition of like, you know, the Earth only has so many resources and so we need to like stop like having all these people, you know, use all the Earth's resources and, and blah blah blah. [snorts] Um, and then you get like the one-child policy and people like, you know, killing their children to like meet this and then you like actually what ends up happening is you get like the Haber-Bosch process, uh, you know, on the, on the Malthusian side and you figure out how to make like a bunch more food or I don't know, Ehrlich. Ehrlich at one point said that giving uh human, giving humanity uh unlimited clean energy is like giving an idiot child a machine gun I believe is the quote or something like that. And the idea is like they're just opposed to growth. Like even if you can give them like here's a perfectly clean uh source of that growth. People are just opposed to the idea of humans kind of growing and taking over uh uh more of the, you know, more of the earth. But you know there, there's another great quote uh that is like there are no um what is it? There are no natural resources just like raw materials and then like you know humans figure out what to do with it you know like you know it takes intelligence to figure out what to do with that stuff. So, I'm going off like deep on a tangent right now, but on the environmental side, like nuclear is now creating demand for more solar, for more batteries, for more nuclear, for more of all these other things and probably accelerating uh you know a transition to, to clean energy faster than would happen otherwise. And it's using that gas and that's, you know, fine in the short term. Like the world's not going to, you know, end uh because of that. So, I think the environmental stuff is BS. Um, I think that data centers are good at pulling forward demand. I think like honestly the biggest problem with this stuff that we've seen so far is that like it's a way that the people at like the labs talk about it. Uh, I think it is, there is like a little bit of I'm worried that this is like kind of what that guy Satraj wrote about in that report that scared everybody. I don't think the dynamic is as like crazy as he thinks, but that there's almost this like race to the bottom on costs. And so like if your competitor is using AI instead of their customer support people, then you have to and then you get like a little bit of dead internet theory, but like everywhere cuz the customer support is like not quite as good. Uh, you know, there's like that kind of stuff that I think will happen. That's like a human
thing more than a technology thing. At the same time, like I don't like that there's slop writing, you know, everywhere on the internet. And I can see why writers are pissed off about not getting, you know, paid for their writing. And I think it's particularly funny to say like, you know, at the same time, we don't want our models to be distilled. And also, thank you so much for all the books that we just like, you know, ripped the covers off of and scanned and then shredded. Uh, and so there's like a lot of uh, you know, talking out of both sides of the mouth, like literally on everybody involved side here.
Um, but I think like mostly mostly good. I don't think you're a baddie for using it. I think everybody's using it.
I just needed you to say that. Packy doesn't think I'm a baddie, guys.
No, I use it all the time. I think it's a it's a tool. I think the weird part is when you like pretend that it's more than like an awesome tool or like super smart or whatever else. I think that's when it gets weird. But otherwise, go ahead and use the tool.
Well, this is kind of the oil, not God idea that you have, right? It's it's much more like oil and not like God.
Yes. Uh and and oil being like this incredible commodity that supercharges it. Like here's another, you know, hot take while we're while we're on the topic is that that uh, you know, Standard Oil has done more good for the world than every charity ever combined and every philanthropist ever, you know, and I think that's actually probably true. And I don't know how you measure it, but I think it's like probably actually true.
Um, and uh, you know, so it's like this wonderful commodity that like literally you can supercharge everything. We used to have, you know, horses and carriages that [__] everywhere and didn't go that fast and would get tired, you have to water and feed them and whatever else. And then we had oil and we figured out how to refine it into into gasoline. And then you have cars and like does that create other problems? Sure. There's car accidents and there's pollution and then we figure out how to fix those problems. Now we have self-driving cars and we have you know the cars are a lot cleaner and they're electric and so like I don't know kind of progress uh progress happens in this way. But oil was this like amazing commodity that supercharged the world but it like also it's not going to like wake up and of its own volition be like all right like I'm going to go start killing people. That's just not what what commodities do.
Um, and my bet, and I could be totally wrong, and there's again plenty of like super smart people on the other side of this bet, is that AI is the same thing. I saw Sam Altman yesterday was on Invest Like the Best, and I'm going to date uh, you know, a recording of this podcast, but recently Sam Altman was on uh, Invest Like the Best, and he was essentially saying, and again, like excellent. I I think like again, there comes whatever he's been doing on the media training side has been just like unbelievable. And I I give credit to the TBPN guys, but he was basically saying like a lot of people want to feel like they're the hero of history and like you're in the center of like this thing that is just like completely abnormal and like out of line with history that like oh my god like I am at the c I'm at the heart of this thing that will change civilization forever. And he's like I think if you zoom out like actually we're just kind of on this exponential and this is like the current part of the exponential. And what struck me when he was saying that is like the exponential is steeper. So the technology like the thing that keeps the exponential going has to keep getting cooler and cooler and cooler or like more and more and more powerful. So like if oil drove the exponential before you need something that's like oil plus to keep driving it now. But like there's a little bit of like you know you assume the exponential just kind of like you know if it if you think humanity is going to get wiped out just kind of like goes away or I don't I don't exactly know what you assume on the GDP side if you think that like the machines are going to take us take us all over. Um but they're like actually for for it to continue like you need uh you need consumers to keep buying stuff and you need whatever but the technologies have to keep you know robots will keep us going after computer AI or whatever we want to call LLMs like computer AI will get us a little bit then robots and then there will be something cooler that keeps the exponential going uh in the future and so I think like that that zoomed out historical view uh is is useful and so it's it's like oil but maybe like a more powerful version because it has to be to keep the exponential going.
Totally. I mean, I I often think about how we're we are the future's past, you know, and like totally someday this will be so quaint and so like can you believe they had to like talk to LLMs on their computers? Like they my grandpa had to do that. He was a [__] loser, you know? Like it it'll be it has to be that, you know?
Yes, it's always that and it will continue to be that. And maybe I don't But I'm sure there's like some Bayesian rationalist in the audience is going to be like actually your base rate's wrong. Like whatever the Yeah, that's what it's going to be like. Like this is going to look and you can even using it feel that it's like it's going to you can feel where it's going to look stupid to future people.
Where do you think that is for you? Like
I mean like for one like one thing like when you're chatting to it about like you know I I want it to like just be my health tracker because like everybody builds their own like you know spreadsheet at some point that's like here's what I'm eating and here's the workout that I'm doing and I never stick to that. And so I'm like if I can just like after a workout send it a screenshot of my workout or what
It could just be like awesome logged and instead it writes you like seven paragraphs.
And like that's clearly not the end state of of like you know that's not what you that's not what I want. That's not what like the normal person and there's like a million little things like that using using the product where it's like clearly not not endgame. Um and so you could just like you can feel where it's like it's cool. It's it's like really awesome technology and it's amazing that you can like search over all of human knowledge.
I I think the other interesting thing is like I I think we like underestimate how and again like this is me taking the bear case and there's like people at the labs who would who would dunk on me and like send facts back in my face. I think there's also it's like amazing the extent to which there's just like everything has been talked about and written about and whatever in human history and like that knowledge is accessible and that there's like a lot of like the reason that people that Mercor is so expensive so valuable and that like so many of the data companies are making so much money is because for the LLM to learn you have to like feed them increasingly specialized data at the edges and that is again like to me feels like something where at some point if you're like, "Hey, here's everything that we've known for like all of human history. Like, here's everything. Can you go figure out like the next things?" The fact that like you need it you need to pay humans to like tell it the continue to like tell it more and more of the next things to me is like bearish. And I could be totally wrong on that. I'm sure there's like something that's being worked on algorithmically in the labs that like whatever that feels that feels weird that you like need to keep paying professors a shitload of money to like teach it on the edges. Is that what is that what that company does? I I don't know what that that company is.
So Mercor is one there's like you know this is this is what Scale AI did. This is what like a bunch of a bunch of companies uh a bunch of companies are doing you know some in like for particular uh use cases like coding or you know coding has a lot of information on the internet uh some for you know physical engineering of stuff because like a lot of those you know that that knowledge is not captured. You'll get uh you know you'll hire a bunch of experts to like say like here's a particular question here's like how I would answer that question here's a rubric for like what the right answer will be here's me like training the the the model on like what the right answer looks like and talking to people not at Mercor just like generally in that you know kind of category they're going to like I think you would underestimate like how much new LLM knowledge is just like us getting a person to like teach it that thing.
Uh and so again like all of this there I'm sure there's somebody very smart out there who's like yeah but we have you know you know recursive self-improvement is coming and that actually also not just applies to models but applies to knowledge and if you have reason for long enough then you can actually figure and maybe that is like all true. It seems very weird that we've gotten to like this point uh and it hasn't been like all right cool actually thank you I have all the like all of the information you guys have. I've stored it all here. I'm just going to go figure everything else out. Like, you'd think that it would do that at some point, right?
Right. Yeah. And also, I mean, I don't know why it just came up, but like the people who seem to know the most like do well on Jeopardy, you know, like that's that's kind like that's like what that is good for, you know, is like to be good at trivia, you know? It like I don't know. It just all of this stuff to me has this feeling of like drastically underestimating what humans actually do, right? Of like it's just like the all of the debates of like are we is it going to be like creative like it's like dude you're not giving yourself nearly enough credit like you are it's way more weird and complicated and like it's just you know believe in yourself a little bit you know it's like it's you know.
Yeah. I think this is my problem with it because like again I use it. I love it. I think it's amazing. Like I I my I don't if you took it away from me like I don't know what the day-to-day would look like. I'd go back to Google and I'd be like clicking on links and trying to and all of that. And I think it's just very weird that like there's like this am I allowed to use the word cuck here? There's this like cuck fantasy that people have where it's like a this thing is so much this thing is so much better than me. Oh my god. You know it's like it's so weird. It's like really [__] weird.
That's a good way to put it.
Yeah. Yeah. It's a real powerlessness thing of like it's it's it's over. I give I give up. It's like no, not really. You know,
I know. And maybe, you know, maybe I don't spend enough time in school. When I walk down the street in New York, I like I'll think about it. I'll catch myself sometimes like just being like there is literally nothing about this experience of walking down the street here that has changed like one iota because AI exists now. Like my this is exactly the same as it would have been 5 years ago. Uh so like I don't there's there's clearly something harder than just like make you know a machine that can that can like do math problems and totally changes.
And it's the simplest things you know on in New York it's like trying to get buy that t-shirt you know of Mike Tyson or whatever it's like that's so complicated you know or like figuring out if you want to buy it or all of this stuff is like yeah and that's the beautiful weird stuff too.
