Dwarkesh Patels Quest To Learn Everything Ep 27

How is AI integrated into your work and your life right now? I’m a host of a podcast where I try to ask good questions. My upcoming guest is a geneticist and I just upload the ePub of the file. Then I’ll just have a bunch of shots where I’m like, “Well, how does he explain what groups made up modern Europeans?” It has all the context in there that’s it that ends up being incredibly useful. First thing I do is just what are the key ideas and concepts I really need to understand. It’s given me a bunch of question answer pairs that consolidate the key things I need to understand about this post. So, a lot of topics I just find I’ve had a vague sense of what’s happening, but I don’t really like get it and it’s super helpful to chat with Claude to make sure I’m on the right track. Wait, step back. Why is this necessary? What’s going on? How do I think about the broader context of what’s happening here? Because I really can’t ask good questions unless I have a good mental model of what they’re talking about, where all this is fits together. I have come up with a couple of different workflows and tools that help me really interrogate and make sure I’ve reinforced what I’m reading about or learning and language model is very helpful cuz like it gives you the the content in another context. It can quiz you if you want. I really just want to know everything. It seems like every company these days is rolling out an AI chatbot in the bottom right corner of their website, but they’re mostly still pretty annoying to use. But what if a chatbot could act less like a chatbot and more like a conscientious human being? It could know what you’ve done, look up account information, and even perform actions on your behalf. It could take over your mouse and guide you through the site in real time. That agent exists and it’s called Command Bar and they’re a sponsor of this show. Command Bar’s user assistance platform is an intelligent agent that companies can embed on their website. It looks like a regular chatbot, but it isn’t one. It can perform actions on the user’s behalf. It can fetch data. It can even co-browse with them through the website. And it’s not just reactive. It can actually be proactive. It can nudge users when they seem confused and help them through the site. It’s sort of like a friendly human assistant standing by and helping a user when they need it. Not just an annoying barrage of pop-ups. It’s a power-up for product and support teams that want to drive engagement and activation, encourage conversion, and of course deflect low-value tickets. And it’s trusted by teams like Gusto, HashiCorp, FreshWorks, and more. If you’re interested in integrating Command Bar into your website, check out the link below or in the show notes. And now, onto the show. Dark Horse, welcome to the show. Thanks for having me, Dan. I’m so excited to have you. I mean, for people who don’t know you, I assume everyone knows you, but for people who don’t, um you do the best, most like honestly the smartest interviews in AI that I’ve found. Um you have like really incredible guests like Mark Zuckerberg, Demis Hassabis, Patrick Collison. Um yeah, you you you’ve created like the go-to show for smart people to to learn about AI, but you also kind of branch out into lots of other things like geopolitics and history and stuff like that. It’s it’s really great and you’re like you’re just one of the one of the people that inspire me to like make smart content. So, I appreciate you coming on the show. Yeah, that’s really kind of you to say. I mean, um I’ve I’ve always been sort of trying to make I have the conversations that I would like to have if I was getting dinner with one of these professors or CEOs, what would I want to ask them and um I’m I’m I’m I’m glad other people enjoy them as well. Yeah, it comes through. Um and I I think like it’s it’s really fun to get to like turn turn the tables on you a little bit cuz you’ve done some interviews, but but mostly you’re interviewing other people and I think um like it’s probably on people’s minds like how you use AI in your work in in your life and so and so that’s what we’re going to talk about today. So, um maybe just start start by giving us a little bit of an overview like um how is AI integrated into your work in in your life right now? Yeah, so it’s actually changed a lot. I remember a year ago, this was after I think it was after GPT-4, um I somebody asked me, you know, do you use AI to help you with your research or prep? And I was like, “Not at all. It’s completely useless and mid. It gives you these banal um you ask it like what should I ask so-and-so professor and it’ll give you these banal where did you grow up? What’s your book about? Whatever.” So, initially it was, you know, um terrible. I think recently the models have gotten to just the point where with these like, I don’t know, 4-0 or especially with the Claude new Claude models, they’re intelligent and interrogative and they consider the context which you provide to them. And so, they’re still not that good at like, “What should I ask this person?” because obviously that’s why I have a job, right? So that I can help with the questions. But for the research itself where you’re for me at least, I try to like ingest everything they’ve ever written, um all the rebuttals to their ideas, all the other considerations, and there’s often a lot involved, especially given there’s like many different fields I try to go deep into. Like the last interview I just did was with Dylan Patel who writes SemiAnalysis. It’s a um it’s a publication about semiconductors and AI hardware and so on. So like there’s a bunch you have to learn and uh it it’s I I mean, I can go through my workflow, but it’s incredibly useful to be able to like have this thing where I’m like, “What’s going on here? Can you help me explain this?” Um and then I I guess one bigger thing I’ve been thinking about is ever since I interviewed Andy Matuschak, if you’ve if your audience is familiar, he’s the guy who did um who talks a lot about how space repetition and other tools can enhance our ability to learn and how the normal mode of learning you’re actually not picking up that much. If you pick up a random book and start reading, you’re not getting that much out of it. And I really have found that to very much to be the case to the extent that if I’m just like casually reading a book, I think I’m basically wasting time or entertaining myself. And I have come up with a couple of um different workflows and tools that help me really interrogate and make sure I’ve reinforced what I’m reading about or learning and the a tool like and language model is very helpful cuz like it gives it it gives you the contact and the content in another context and you can like it makes it can quiz you if you want. So, it’s it’s super helpful with that kind of stuff. That’s really cool. I think we should start I want to start back to front like with the the stuff you’re using to read cuz I think the reading all that reading is sort of the input like one of the inputs to the interviews and then we’ll get into the interviews um and I’m really excited for both. So, so let’s start with with using AI to read and to learn. The um you know, so as I was I was talking about like the what one of the main things I think is important is if I’m studying a topic over the course of a few weeks, it’s um especially if it’s a difficult topic, it’s like new to me, it’s incredibly important that I’ve I’m not just casually reading cuz if you were just casually reading, it’s like every day you reread the same key terms, the same concepts and you’re you start over from scratch. Um so, one of the things I like to do, for example, I was recently interviewing Dylan, right? So, if I go to his publication SemiAnalysis, there’s just a ton of lingo and things you have to understand. Um so, the new one was pretty interesting. It’s talking about why nobody has built a huge training cluster yet. And then first thing I do is just like, “What are the key ideas and concepts I really need to understand?” So, I made myself um a hugging face space. You don’t honestly need to do anything like this. You can just It’s pretty simple to have Claude build you a hugging face space or if you prefer, it what it literally does is like apply this prompt to everything I paste in. So, just you can you just copy paste that prompt into Claude yourself. But basically it just has a I copy pasted some of the things in Andy Matuschak’s post about how to write good prompts and now I just ask Claude to make those prompts for me. Space repetition prompts. So, um when I do this, hopefully in a few seconds we’ll get something back. Um initially this will give just give me some ideas of like, “What are the key ideas here I need to understand?” So, um super useful, right? I can even zoom in a little bit so it’s more helpful. Um so, for for the audience who’s listening like it’s given me a bunch of question answer pairs that consolidate the key things I need to understand about this post about um you know, we can go through the specifics here. I’m sure that the actual specifics of AI hardware will bore people, but um a lot of the things where it’s like, “Okay, if you don’t get this, you’ve like totally missed the boat here.” And so, you can start with something like this. I add it to my space repetition app um or I can just look through this and I’m getting a sense of like, “Oh, okay, here’s what it would take to train a GPT-4 level model on a 100,000 H100 cluster. What are the three main types of parallelism you need to use to train on a big cluster?” Whatever. And this is on for a technical post. On other kinds of posts, there might be different kinds of cards that come up. For history, it might be a different kind of thing. For philosophy, it might be a different kind of thing. So, this gives me a lay of the land. I love this. This is super interesting. I like I feel like I can go in like a a bunch of different directions, but what where I want to start is like um how are you reading and when are you reading? So, is this like are you using this specifically for um for reading that you’re doing for the show or are you just doing this for any reading that you’re doing that you feel like is serious and you really want to learn? Um both. So, just this weekend I was reading um I forgot the author’s name, but it’s it’s a book called Medieval Technology and Social Change and it’s about how different things that were developed through um the last 1,500 years technologies like the stirrups, how they affected society. And it’s like you can it’s entertaining, you can read it, and then one of the things is like, “Okay, what did I really understand what’s going on here?” Um what the relationship he’s trying to elucidate. So, afterwards In fact, I have some Claude chats where I was just going through um while I was reading it. Um let’s see if this recollects. Do it. I I want to know. I’m I’m on the edge of my seat cuz I have this book. It’s like sitting on the desk in front of me. Um and so, I want to know what what you got out of it. Okay, so um first I just asked it to make space repetition prompts for me. First of all, I was just like, “I read the chapter. I’m not sure I got it.” So, just explain to me the chapter about how he says that stirrups created feudalism. Like, what exactly was the connection here? So, it it’s a much more condensed like here’s what’s going on here. Basically, if you understand this, it’s a useful scaffold so that when you’re reading the rest of the chapter, you understand where the pieces fit together. Have you tried like one one of the things that’s that I’ve tried with this is like cuz sometimes it doesn’t know, especially for a book like that where it’s like not that popular. Have you tried like one of the one of the one of the things I do is create a little Claude project and then upload the text if I can find it. Have you tried that? In fact, let me just Claude AI uh projects. So, if I go to um I literally just I’m like I’m a I’m a host of a podcast where I try to ask good questions. My upcoming guest is a geneticist and I just upload the I get the ePub of the file. I convert the ePub to a text using an online converter. I upload it to project knowledge. Then, I have only just started prepping for this guest, but I’ll just have a bunch of chats where I’m like um you know, how how does he explain what groups made up modern Europeans? It has all the context in there that’s that ends up being incredibly useful like you were saying. Yeah, that’s so cool. I I I love that. I love that feature. Okay, wait, let’s go back to let’s go back to stirrups and in this chat you’re having with um with this book or about the book. Yeah, so you know, it explains that the reason stirrups created feudalism is because you needed um you needed a lot of land basically to support the kinds of people who become heavy cavalry, the knights. The knights need a lot of land in order to um have the income to have, you know, like uh armor and uh lances and other kinds of equipment and to train themselves. Um but you a knight is only possible if you have a stirrup against which you can brace yourself as you’re attacking with a sword cuz otherwise you’re just like a Mongol who’s shooting bows and arrows. So, um but then then there’s a bunch of stuff that’s confusing here like uh why is it so that expensive to have have be a knight that you need to like completely confiscate church lands in order to subsidize this knight lifestyle? Um and then on these kinds of questions, the author is dead. But it’s just like I’m just like murky about it. I don’t know what’s going on. So, I can just um these kinds of things I can the book didn’t even talk about, right? But I can always just continue the conversation with Claude and have it explain what’s going on. Um And so, this is just like a recreational reading that Claude ends up being super helpful with. I think that’s really interesting. What what do you think about um like books like this? Like in your you know, as a as a person who likes history a lot, books that sort of single out like a specific thing like the stirrup and then are like, well, you can trace all this stuff to that like one thing where it like makes so much sense, but then there are things like I don’t know like Guns, Germs, and Steel where like Jared Diamond had that whole thesis about I I can’t remember the the exact thing, but it’s like people in warmer climates or I can’t remember the exact thesis. Is this going to turn out to be like totally wrong? Yeah, how do you feel about uh things like that? Yeah, so my feeling on these kinds of books there’s there’s one I mean, the sort of concise answer is like, yeah, there’s ones that do it poorly, but let just don’t read the ones that do it poorly or something. There is a failure mode for public intellectuals where they initially start off with a discipline and they do some exemplary work there and then they write an initial broad book that’s about how this idea explains a lot of the world and it does incredibly well. And now they’re in like public intellectual mode and now that the next book has to be like, here’s my theory of everything and it’s just not that satisfying. So, I do worry about those kinds of things. But um presumably the reason I don’t know. I’m not I’m I’m not into like reading 500-page books about like uh just how the stirrup physically worked. Like, what’s the point of that, right? I do want to understand the implications and maybe they’re wrong, but um I mean, what else are we trying to do here, right? We’re doing We just care about or maybe you would just intrinsically care about how the stirrup physically works. There I I I will point out a couple of examples. So, there there’s a lot of interesting topics where you really can’t get at the heart of the matter without just considering the whole story. And in fact, so a couple of biographies especially stand out in this way where if you look at Caro’s biography of LBJ or Kotkin’s biography of Stalin, it’s basically a history of the 20th century or in the case of Kotkin even before the 20th century. Um I think this Caro books on LBJ start off with the Comanche raids on uh frontier settlers in the mid-19th century or something. And it goes through rural life in Texas, why electrification was such a big deal, a whole bunch of other things, right? Now, um so it’s basically a history of the 20th century, but it has a very specific point of view or a specific locus, a character that’s moving the story along. And I find those to be incredibly helpful in getting a full picture of what’s going on in an era. Um there’s a couple other books where they really aren’t trying to write a theory of everything. Like I don’t think Caro’s trying to write about like, what is the history of the 20th century? But they just can’t help themselves. They feel like you really cannot understand the very specific topic I care about unless I tell you everything about everything. You know, like um Kotkin’s story biography of Stalin starts with the like Bismarck’s career as a military general and how that changed the way that different powers thought about um colonialism and the need to modernize. And that’s where it starts, right? It’s a biography of Stalin. So, uh yeah, I I love those kinds of books. I think there’s like a very just deep point about the universe being interconnected there, but there’s also like a really interesting point for um people who want to make stuff like make writing or make podcasts or whatever cuz like there’s this deep fear that everyone has about like being pigeonholed and it’s like, well, if I pick this like really specific topic, I won’t be able to like bring all of myself to it or I won’t I won’t be able to be like multifaceted. And it’s like, no, no, no. If you just pick one guy, Lyndon Johnson, and really get deep into him, you have to explain everything else about the world in order to explain him. And I I love that and like as a creator myself, like that’s the thing that I think about when I’m like, oh, maybe I’m getting too narrow here. It’s like, no, no, the narrow is actually good. You can find the entire universe in the narrow. Yes. Yeah, yeah, yeah. I I couldn’t have said it better. Yeah. Um okay, so so basically what I’m what I’m seeing right now is you’re using Claude when you’re reading books that you care about, you care about like learning from, and you’re using it a little bit to like prepare your prepare the mind your mind for like what you’re about to read, which I think is particularly good for like difficult books or for like thinking through a particular argument before you like go through it. You’re asking questions you’re asking questions. You’re kind of like it’s a reading companion. You’re getting more out of the books you read from that. But then you kind of take take your what you’ve read and throw it into this this card generator. Yeah, and so that that’s mostly it’s just chatting with Claude. And so, let me see if I can find a better example. Um So, I mean, a lot of topics I just find um I’ve I’ve had a vague sense of what’s happening, but I don’t really like get it. And it’s super helpful to chat with Claude to make sure I’m on the right track. I one was like um oh, there there you know, Dylan has a couple of posts about how why packing as a technology is super uh necessary for these advanced chips. I I’m not trying to make this podcast all about AI hardware. It just happens to be the last podcast I did, so that’s what you’re getting. But um then I’m I was it’s like a confusing it’s like five series posts about how advanced packing works and how, you know, what what the technical specifications are. And I’m like, wait, step back. What what why is this necessary? What’s going on? Um all kinds of other questions about it when there’s questions about how I’m worried about where I might get too deep in the weeds when I’m just explaining. Yeah, basically I’m just like how how do I think about the broader context of what’s happening here because I really can’t ask good questions unless I have a good mental model of um what they’re talking about. Like, I really get where all this is fits together. Hm, that makes sense. And so, and and and so Claude is kind of like the first thing you flip to when when you when you want to know that. Are you using it on mobile or using it on desktop? Desktop. Okay, interesting. So, you’re doing most of your like reading and research stuff on desktop. Yeah, that’s right. Hm. And what do you think about like just Claude being really great right now and like like I assume your ChatGPT usage is lower than it used to be. Yeah, I think these things will keep getting better over time and I um I I you know, I I yeah, I I think we’re just like getting in the practice of using these tools. I’ll talk a little bit about how these tools relate to my post-production process. Initially, it was kind of useless, but I did spend a few weekends trying to write a few prompts and keep create a workflow. At the time, it was basically useless. Now, it’s actually ended up being useful and I can use the same, you know, Jupiter notebooks or whatever to get things done. So, um it is worth investing in getting even if they don’t work perfectly now to get them part of your workflow so that as they keep getting better, you’re getting the returns from that. Yeah, that makes sense. So, so I want to just go back to the the flat the the the Anki card generator, the the spaced repetition card generator. So, as part of this, once you once you’ve done all this sort of like conceptual um you know, clearing the ground conceptually for yourself to like kind of understand the the basics of of what a guest is talking about or an idea that you’re interested in, then you’re kind of um you’re adding to your your flash cards so that I I guess so that it you retain the information past even when you talk to that guest. Is that right? Yes, that’s right. Um I mean, I I I think of the larger mission of the podcast is to cons Why does the podcast get better over time? And it’s because basically I’m getting smarter or learning more things. I’m reducing my ignorance around a bunch of topics and so to the if I don’t do that, I mean I think about all the episodes I did before I interviewed Andy and started using spaced repetition and I just like really regret it cuz I talked to all these world experts in a ton of different domains and to be honest like in many cases I didn’t take that much away. I vaguely remember some things and now that I use I can walk you through the kinds of cards I make in the spaced repetition tools I use. But it’s like totally a game changer in terms of what I can retain. And in fact, I think it’s not even about making sure I remember what I discussed in a previous episode or what I learned previously. It’s more about future learning because I’m sure you’ve heard the saying about um the you know uh learning compounds because you can use what you’ve learned in the uh past to um learn future things cuz they all interconnect. Well, you can’t do that if you basically forgotten most things you’ve learned in the past. So I my learning has for future other things has become much faster because I have cashed all these different um uh concepts and figures and facts and so but future things are just like I I understand how everything fits together much more. It’s not even about the past. It’s really about future learning. Mhm. Can we see your uh I don’t know what what you use for for spaced repetition Can we see your your deck? Yep. I I will point out by the way uh as a side note so one one use case of Claude that ended up actually being pretty useful. You can sometimes you read obscure like I was reading Nick Land’s selected writings about AI and his accelerationism and I was like, what’s going on? Like I genuinely I’m like, what is his argument basically? Like why does he think that the AI takeover and the whatever thing it creates in the aftermath will be good cuz he’s a smart guy. I’m assuming he has an interesting argument. So I pulled the PDF of his selected writings. I just asked Claude like, okay, so why does he think it’s a good thing that AI is taking over humans? And he uh the it offers a summary. Like initially this isn’t necessarily that helpful cuz I kind of did read this in the essay. But what’s what’s helpful is that when you go through and I’m like I respond like, I don’t get it. Like what does he think is wrong with human society that you can’t uh you have to erase it? And then it gives an explanation and I’m like I still don’t get it. Like what exactly are you talking about here? And then like here’s what I do with the podcast, right? I have the guest on and ask them, what do you mean here? Like I I I I disagree. Here’s a contradiction, whatever. And going through their writings with Claude and like have I have I actually found a sort of blind spot in their thinking or is this just me being confused by their ideas? It’s super helpful. That’s that is really interesting. It’s like um you can you can get you can get down to a deeper level before you talk to them so that you can start there with them as opposed to like starting at the surface. Exactly. Which is really cool. I I use that too for like um for for difficult books like not necessarily for like interviewing the author of those books, but like for example, I interviewed Reid Hoffman like I don’t know a month or two ago and I wanted to talk with him about the kind of like intersection between philosophy and AI and he was like um he like almost became a philosophy professor like at Oxford wanted to like was really deep into Wittgenstein. So like I read a bunch of Wittgenstein um which I hadn’t read in a while and I I just used Claude for it and it was like so much better um because I like I haven’t taken a Wittgenstein class or maybe I took one in college like a long time ago, but I’ve read him a lot and it just there always these those points in those kinds of books where you’re like, I think I I know what they’re saying, but like I’d probably have to go to a a grad graduate school and like get a master’s in this to like really know. And Claude is actually like makes me be like, oh, I don’t need that anymore. Like I I any book I want to read like this like I basically know and it just helped me so much in that in that interview cuz I could just ask Reid like really deep Wittgenstein related questions and he could answer them. Yep. Yep. I I think that’s totally legitimate. I think some people would be like, oh, you need to read it in the original, blah blah blah. I think if you care about the ideas and you think the ideas are timeless and not the ideas are not about the specific kind of prose that the original author used, but just generally like what is the essence and the gist of what’s happening here? If you care about the ideas then I think this is totally valid, right? I I don’t disagree with the people who are like, no, you need to read like the specific syllables that Wittgenstein used. Yeah. I mean, I’m also just saying like I’m I have the book open and then I just take one of his like statements and just throw it in there and then it’s like here’s what here’s what it what it means or whatever which which I think is really great. Um okay, so you’re going to you’re going to show us the the the spaced repetition cards. So what what app is this? This is Mochi. Um Interesting. It’s like Anki, but I this is the one I use. Um Why? Uh so that actually I don’t have any cards today cuz I just went through them this morning. But let me give you a sense of Let me give you a sense of what kinds of things I have, right? So Yeah. I have um if you go through history, recently I’ve been planning on interviewing David Reich who is a geneticist who explores human origins. And these are especially cases where you just like reading the book I’m like I would have totally forgotten. He names all these different ancestral groups and how they combined and in what eras, you know, when did the Yamnaya people come through Europe, when did the Anatolian hunter-gatherers, you know, wash over uh Eurasia? All these things that were just like you you read it in one ear goes out the other one unless you make cards for it. And so I made a ton of cards about this kind of stuff. Um I and you you know, so there there’s examples of that here. It’s especially useful for uh hardware and um technical things. So here I feel like if I don’t make cards, I’m just constantly relearning the same things cuz I didn’t learn the lingo the in the right way first. It’s not just about learning the terminology. It’s about understanding the underlying concepts. Let me give you a good example of that. So um maybe I’ll I’ll step back and I’ll explain like I go through these cards in the morning. If I maybe you can see what it kind of looks like if I do the cram cards thing. Please. Um I can go through them and right now I’m like, I remember this, right? I remember it’s uh this is the first one that came up randomly, but it’s um multi uh query uh attention to not have to use KV huge KV values and then sharing KV values between layers and using local attention. And that’s the answer. Now um it seems sort of trivial right now cuz it’s just like three things, but like I would have totally forgotten about this if I hadn’t made a card for this as soon as I read the blog post and then it’s just like I wasted my time in the future if I’m learning about these technologies in a different context. I just like don’t have the connection to what was happening here to connect it to, right? If I go to a different um category, if I go to cram cards. Um this is the uh the white thing. I would have totally forgotten about it if I hadn’t made these cards. Um I yeah, I I’m just a big fan right now. I I sort of I’ve become a spaced repetition fanboy these days. How do you think about like the the usefulness of spaced repetition in a world where like any of these questions is possibly is like pretty much answerable like with Claude with like a you know, one one search? Yes. So I think it’s about not necessarily remembering this information, but when a future thing comes in, you understand like the conceptual In fact, let me give you a good example of this, right? So I remember sometimes I actually make cards about facts that I don’t even understand in the moment, but in the future the as I learn more about the field as as the sort of territory becomes more clear, the things I said in the card make more sense to me. So if I um I was reading some of Colin Burn’s papers and so I made this card about like why Colin Burns thinks that uh alignment is a tractable problem or understanding what model thinks is a tractable problem. And um at the time I wrote things down about like uh you know, features are in a linear space. What the What does that mean? Um or like we can sort of see features in other sorts of categories. And at the time I’m like, I have no idea what this means, but I’m just going to write it down cuz I read the blog post and there was no point of reading the blog post if I’m not going to make the card. Later on as I learn more about how uh the residual stream model of how attention works, you know, uh how what that is and so forth, this card made much more sense to me in the future. But I would have just like totally memory holed or not even memory holed. I would totally forgotten this content which required future understanding if I hadn’t made a card of it. And then when I see the card again in the future, I’m like, oh, this is what Colin Burns meant. Now that I understand how attention works, this is what it means, right? Um Yeah. This is this is really interesting to me. So I want to get into some of the like the ways you use AI for for interview prep um cuz I think we we’ve mostly covered the reading stuff, but before we do that, I just want to like understand like what is driving all of this? Like it feels like you are just consuming massive amounts of information and turning that into knowledge in your head. Like you have this in sort of just overdrive of curiosity which I actually resonate with a lot. Like I’m I’m surrounded right now by books um and I’m just sort of like um curious for you what what that. What do you think that’s about? Mhm. I think I um I’ve uh I I I I really just want to know everything, right? I It’s I I don’t know how to express it. There’s a beautiful passage in a Will Durant book as he’s turning 90 where he’s writing a memoir basically of his main ideas um called Fallen Leaves and there’s a passage on philosophy where he says uh you know, as as you get older, maybe with all the philosophy and history I’ve done, I can I’ve reached some plateau of higher understanding and clearer insight, or at least I’ve understood that such a thing is possible.” And I I something like that just resonates with me. I don’t know. I just like that I find that idea really appealing. I’m nowhere close to it, but I just hopefully in the years to come that will just be a thing that um I also really admire people I’ve had on the podcast who do have these self-consistent and really deeply interrogated world models. You know, I’ve interviewed these guests and some of them um people couple names come to mind, people like Carl Shulman or Tyler Cowen or Burn Hobart. They It feels like they’ve really read everything and you know, everything you know is a subset of what they know. And I just um I I I find that them to be super compelling as thinkers. Of course, there’s many things they can still be wrong about. I don’t I’m not one of these people who buys like there’s like a thing where you just know everything and you can never be wrong, right? You always have blind spots. But the ability their ability to which you can see when you talk to them to connect anything you ask them about and they’re like a Claude 6 in the sense of like you start talking about why has a fraction of finance as a percent of GDP ever amassed in Tyler there and he has a right off the cuff just in a super interesting answer that connects a bunch of different disciplines. Um you ask Carl about like how fast um AI hardware could grow and just like done this sort of Fermi estimates on how fast algae bloom and how much solar power they consume and um how many fabs TSMCs make It just like I I I I I find the that sort of compression of the input they’ve ingested over their lives and they can Not only do they know that stuff, but they can really connect it in a really interesting and compelling novel way. I I I I find it super compelling. And and in terms of developing your own worldview, like do you have that anywhere where it’s like you’re creating like some sort of living document or is it just is it just all in your head, like all the stuff that you’re learning? Obviously, you have the you have the cards, but that feels like more like um dots in the space rather than like uh the ways that they all connect and and how you think about everything all together as a system? Hm. I think um I’ve been trying to do more of this recently. Uh and now that I’ve sort of built up an underlying maybe vocabulary or understanding because of podcast, it makes sense to do more of this. Something I’ve been doing recently, if I let me pull this up. I’ve only just started. Hopefully, there’ll be more by the time people are looking at this, but um I’ve started writing, you know, riffs on different books or things I read. And if I go to um It’s basically on my website and um so just like I can read a book and I have questions or I connect it with other things I’ve read. I can I remember for example in when I was in the Steven Pinker’s Language Instinct, he was writing the book before the FOXP2 gene that can help explain human uh language was found. And so, he has all these observations that are then later explained by the FOXP2 gene. Um and so, I can just sort of the the sort of connection that uh you’re talking about I can do by riffing on other people’s ideas. I I actually am curious, do you have suggestions on what I should be doing? Maybe I should be writing more blog posts or what what do you suggest I should do? Um let’s see that’s a good question. So, uh well, before before we get there, like one of the things that this reminds me of it I think Claude is so good for reading old science books cuz it can tell you what’s outdated and what’s what’s not. Um I do that all the time and I I love that I love that little thing. But yeah, I mean, I think basically like developing a worldview is like you have to just try. You know? Um and you try over and over and over again and I I do think like blog posts are uh really good for that, especially like, you know, for me like I have to write every week. Um And so, I’m like forced to take a view on something and generally if you’re like intellectually honest, you like want one post to like somewhat agree with the last post and your audience will call you out if you’re like just disagreeing with yourself all the now, I’m actually like my big thing this um this quarter is like I I just have these like ideas that are simmering that are like sort of the relationship between like um language models and like some deep philosophical questions that we’ve been like talking about since like Plato, which is like the appearance reality distinction and like how do we know what’s true and what’s knowledge and all that kind of stuff? Like I think there’s a lot of overlap there and it requires like it’s going to be like a 10,000 word post or something like that. Um and so, what what I’m doing is I just have like a Claude project with like I have all these like little notes and riffs and like and stuff and I’m just like going into Claude and being like, “Hey, like what’s what’s the thread here? Like what’s going on? Can you help me like figure out like there’s something in me that I have all these little ideas for, but I can’t quite like put it into an argument that all makes sense.” And I think just honestly like sitting with that for like a couple months, I will I will know what’s in there, but there’s something in there. Um and yeah, it’s cool. Do you do you make a Claude project to like here some of the things I’m thinking, how do they connect? Or like how do you keep track of those things as you over those months? I’ll I mean, I’ll just show it to you. Um let’s let me just pull it up. So, okay. So, if I go into Claude um I have a couple different projects. Um one project is Seeing Like a Language Model, which is the title of this big post, what whatever it is. Another is Zen and the Art of Motorcycle Maintenance. So, this is like a book that I’m reading as prep for this um this piece that I’m writing. Um I’ve read it a bunch of times before, but like now I’m like doing a little bit of a deeper read and so, I have like the I have the uh the whole book uploaded and then I can like ask questions. Then I have another one that I love called My Psychology, which has a bunch of like journal entries, uh goals I’ve set for myself um over the years and then also like things I’ve observed about my about my psychology or things I’m working on, like little aspects of myself that I’d like to grow or change. And so, when I’m making decisions or thinking something through, I just go in there and it can reference all that stuff, so it knows who I am, uh which is really cool. Um So, in Seeing Like a Language Model, let me see if I can pull it up in the projects uh directory. Um So, I have like uh basically I have this one note in Apple Notes, which is like uh every time I have a little thing come into my head, I’m like I just I just put it in there. Like let me see if I can find it for you. Um Um I just throw it in here. And this is like huge and messy and it’s like different quotes from different books and like just different ideas that like come to me off the top of my head as I’m like walking around. And I think that there’s like a thread here in all of this stuff. Um they’re all like I can see how they’re all related, but but like I can’t quite pull it out and so, like what I’ve been doing is I just like throw it all throw it all in in here. We have uh we have this like all the all the quotes and all the ideas and fragments. I have a little bit of a draft, like an intro. And then I have a chapter of a book um uh by Richard Rorty that I think is is is really good um called Pragmatism as Anti-Authoritarianism that like kind of sparked this whole thing. Like I read that I read an I read a chapter of that book and then I was like suddenly like rereading like a bunch of Plato and like Aristotle and and like I was just down this huge rabbit hole. And so, um like what I did for example is I put all that stuff in here and I was like, “Hey, I have a bunch of notes and some fragments for of text for a long 10,000-ish word piece I want to write called Seeing Like a Language Model, but I need to understand what I actually think and make a bit of an outline before I get started. In order to do that, I need to understand the patterns of what I’ve been thinking and writing down and reading about. Can you suggest some ways that you can help me do this? I want to get from where I am to an outline. You have access to some fragments, notes, and early unfinished intro.” And it just like has a bunch of ideas, like thematic analysis or argument mapping or chronological development. And I’m just like sort of going down the rabbit hole with it where it’s like um you know, I asked it to do concept clustering. So, it’s like, you know, one of the concepts that I’m playing with is the philosophical divide, Plato versus Aristotle, which I think is like not quite right. It’s actually Plato versus the Sophists, but like it’s it’s it’s close or the evolution of Western thought. It’s like how does Plato like um uh how does Plato ladder up into the rest of Western thought and and into into science and into and just the way the Western mind works. Um and then how how do language models sort of like differ from that from that um paradigm? Um So, that that’s the that’s the basic thing that I’m trying for cuz I I I I do have The reason I asked this question is cuz I’m selfishly like I feel like a little bit I haven’t done the like big idea thing as much as I really want to cuz I’m I am writing every week. I am sort of like reacting to stuff. And so, I want to be a little bit more thoughtful. Um and this is this is my like attempt to do like to put all of it together into something that like makes sense. Yeah. You know, as as you’re going through this this really actually makes me want to write more because now that you’re talking about it, now that you asked the question, I’m like, “Yeah, I I should be sort of consolidating the things I’m learning in a more comprehensive way and in a way that’s also more useful and accessible to other people as well, right?” I you know, I spend weeks like learning about some random what not random, but like the things I care about about to prepare for Daniel Yergin, the guy who wrote The Prize, it’s a history of oil or a geneticist um you know, AI researcher or whatever to the extent that I’m getting something out of these research processes, I should consolidate it in a way that’s not evident in the podcast itself. Yeah, I mean, I I selfishly want you to do that cuz I’m curious what you think. Yeah, I appreciate that. You will send this out through your newsletter, right? Is is that the main way you Yeah. Yeah, this will this will go out through through newsletter. I might do something special for it like you know, maybe we’ll maybe we’ll make a little mini site for it when it when it launches, but that’s sort of in the future. I have to actually I have to actually write it first. So okay, so so so let’s I want to move on. I want to I want to talk about the how you use AI to to sort of like do the interview prep. So so let’s let’s move into that and then we can also um uh we can also um maybe even like prep for an interview together. Yeah. Yep. Okay, let’s do it. Um I honestly the interview prep is like it requires a lot of work, but fundamentally what’s happening is not that complicated. Like I can just show you a document I might have made in the past. I’ll share my screen. So honestly it literally is just like I’m going through I come up with a bunch of questions and I sort of group them together in relevant categories. If I go to um if I was interviewing David Patel I’m sorry, this is not the right one. Um It’s like a bunch of different Yeah, it it just like a list of questions basically. It’s not complicated. But the process of coming up up with them is you know, very research intensive. So we can go through like if I’m I’ve got some like only barely started preparing for them. We can go through the process preparing for them. Yeah, can can we can we I just want to stop at those questions like again selfishly cuz I think it’s really interesting. Like you have these like long lists of questions that are organized by theme. Are you like going down the list or are you sort of jumping around Yeah, so it’s really interesting because um I come up with these list of questions, but it’s like it really never ends up being I ask question one and I ask question two and I ask question three. The it I you know, I start off with the interesting question and if you listen to the interviews hopefully it comes off more as a almost conversation because I spend so much time preparing that I have these questions basically memorized. And so the next one that is appropriate to their response. If they say something about um if you know, memorization in LLMs, I’ll have a question prepared about that or related to that. And I’ll I’ll just ask them next cuz that’s what fits in together. And so you know, I’ll have a list and this is what I’ll send them if they ask for it, but like really it’s just sort of me off the cuff like here’s here’s a question I remember that was relevant to this in in the actual interview. That makes sense. So like the the point of the doc is like it’s almost like writing the doc is is the prep itself and it’s you don’t even necessarily need it in the interview. Like maybe you have it just in case, but yeah, that makes a lot of sense. Yeah. Um and then yeah, we can we can even go through Let’s see. I’m doing a couple of interviews in the future. David Reich and David Reich and Daniel Yergin. So the one the former is a geneticist with about human origins. The second wrote The Prize which is the you know, the famous book about the history of oil. Which one sounds more interesting to you? We we can do that whichever one. Uh I want to I want to do the famous geneticist. So let’s go to Claude. In fact, I I I do have his book uploaded as a project, so we can just use that. That’s great. And so basically what we’re going to do is like we’re going to we’re going to watch you and I’ll I’ll do it with you. We’re going to prep for an interview with this guy. What’s his name again? Um David Reich. David Reich. Okay, cool. Can we get like a little bit of background on David Reich? Like maybe we can even ask Claude cuz like I’m you know, obviously I’m a I’m a newbie to David Reich’s So he is a geneticist at Harvard and over the last decade or so their research into how have human populations across the world been formed? Basically like how do they Who are the Europeans? What groups make them up? What ancient migrations and genocides and population replacements made them? Same with the Indians or Native Americans or Africans. It’s completely changed I mean they’ve basically sort of like made many academic disciplines irrelevant because they actually have empirical data on like here’s actually what historically happened. You guys are completely wrong about what you think your theories of what happened. Um If you’re familiar with Naftali Minsky’s challenge, you know, like you you have these like bird poop squirrels, but with some advanced techniques you can get some useful information out of them. I feel it’s in a similar vein. Obviously they’re not the same kind of project, but like it’s a similar vein of like once we develop the advanced mathematics or genetics or whatever to understand what’s latent in the genome, we’ve like we’ve just uncovered a ton of insight about what’s been going on in human history basically. And I’m I’m sorry, I’m just going getting nerds and just going on riffs here, but like one of the interesting things is uh you you can see when one population replaces another whether it was just like oh, we met and like we were like now intermingling and trading and whatever or is it like we’re committing genocide against you? And you can tell that because if in the case where it’s genocide or population replacement, it will be that the the male line of the population that is invading will overtake the male line of the existing population, but the female line will remain. So mitochondrial DNA is only comes up in the female line and you’ll see like the female line because they’re getting you know, like the new men who are coming in are taking them as wives or something. And then the Anyway, so you you can just like learn a lot about like what kind of invasion was it? Did they like conquer or was it were they just like mingling or something? One of the many things you can see from that DNA. That’s really interesting. Wait, and so so this is like basically re-examining DNA evidence of like old settlements and like basically he’s uncovering new ways of being able to analyze the DNA. Like what’s the what’s the what’s the new methods that they’re using to like draw new conclusions from existing evidence? One of them is just that, right? Like seeing how the Y chromosome and the mitochondrial DNA cuz like you can just you learn a lot about population based on how the female versus male male line is propagated about like what was the social structure like and so forth. Another is you can even tell how the level of inequality in a society because if there’s a lot So for example in India um one of the things I was super surprising is that the the amount of endogamy which is to say that the um a certain caste in a certain village would just like not there wouldn’t be any sort of intermixing with another caste in the neighboring village. Like it to the extent that’s true of nowhere else in the world and they were able to find this in India where the amount of social stratification you can see that in the genetic catalog over the last thousands of years where for thousands of years these two neighboring castes haven’t mixed with like 99% or something which is like even from sort of infidelity or rape or something you you would expect there to be more than what actually ends up being the case. So you can understand modern culture in India based on what has happened over the last few thousand years. That’s really interesting. So I want to like I feel like you’re doing like such a good job of summarizing his main ideas. But I kind of want to I kind of want you to do the same thing with with Claude so we can see how see how you stack up versus Claude cuz obviously you’ve you’ve you’ve input his his book into into this project. So it has it has that as as reference material. Can we ask it to just like summarize like a few of his main ideas? idea. Can you summarize and maybe like the techniques he used to come up with with his perfect new. And so what you’re writing is can you summarize the main ideas from the book and the techniques he used to come up with new insights. Cool. And one thing that’s like really cool about this is like you’ve been able to do something like this with ChatGPT for a long time, but ChatGPT’s context window isn’t that long and so it chops it up and like it’s not going to it’s not going to really be able to summarize the entire thing because of that. You know, it has to like find the right parts of the parts of the book and the embedding search in it is not very good and all that kind of stuff. And Claude you can just like throw us throw a ton of stuff in the context window just like makes a big difference. Okay, so it looks like it looks like we’ve got some some some answers you want us that the ancient DNA has revolutionized our understanding of human prehistory. And then we we’ve learned that populations today are the result of multiple waves of migration and mixture. And then you know, just like a bunch of other genetic stuff. Then it talks about the key techniques about whole genome sequencing and how they’ve enabled these sorts of new new discoveries they’ve been making. Yeah, yeah, but but anyways, there’s a bunch of interesting things about their research. Well, now I’m interested in like okay, so the key techniques that he’s is using are whole genome sequencing of ancient DNA samples. So is whole genome sequencing like a new thing that you can do it on ancient DNA samples? So it’s saying like by improved extraction sequencing technologies. Is that that’s Yeah, but that is an interesting question. So we can even ask Claude cuz I’m not sure. How How exactly do you sequence an ancient um or you know, a prehistoric uh genome? Like can you do like what what what is How does that work, right? Okay, so they grind the bone and they have then techniques to get the DNA out of that. Now another thing we can ask is like one thing I’m curious about, let’s see. I don’t really remember the chapter on um Native Americans. I could ask about what exactly happened with Native Americans. Here here’s one thing I’m curious about. Like how would Okay, so I don’t even know if David Reich addresses this himself, but like how would David Reich’s theories help explain why civilization suddenly emerges so rapidly and that too concurrently in the new and the old world. After 10,000 BCE, aka the end of the last ice age. And then I’ll I’ll maybe I’ll just like ask Claude why I think it’s an interesting question. So like this seems like a really remarkable coincidence given how long humans have been around, you know. That is interesting. Well, coincidence given that human I’ll correct my spelling. Given that humans um have been around for hundreds of thousands of years. I didn’t realize that if that we believe that it emerged like at the same time in diff in geographically disparate places. that’s totally new to me. I thought it I thought it was like just in Mesopotamia or whatever. It’s actually there’s a really good book by um Peter Turchin called The Great Divide and it’s one of the most interesting books I’ve read. This is a side note. It’s about Yeah, it’s comparing the emergence of civilization in the new world versus the old world. So in the old world um uh in the sorry, in the new world the Caral is like civilization in 3000 BCE and it’s based on fishing and um not on like conventional agriculture like Mesopotamia. And he talks about how that changed the evolution of the culture in the new world versus the old world. But anyways, um Mhm. that’s really interesting. Okay, so major population movements and uh mixtures exchange of ideas um Shh. So it says maybe there were genetic adaptations during that time. Um There were no major biological changes. So he’s saying that I guess human population like yeah, we we weren’t like a different kind of human after the ice age. Um So we’re sure common answers to that were probably wrong. Maybe I’ll tell it like be more specific. I I still learned your answer doesn’t help me explain your answer doesn’t help me understand why the end of the last ice age led to all led to civilization. What changed from before? And then you this is really helpful to interrogate LLMs in this way because their initial instinct maybe because of our lazy chef is to be sort of summarizing mode and comprehensive and like I don’t know, just like give me the answer, you know. Um Mhm. Totally. Okay, so maybe climate stabilization and like uh increased food sources population growth. Okay, interesting. Um Yeah, so this is super interesting. Claude didn’t really give me a full sort of full answer, but this is here’s why this is still super useful because now that I know Claude doesn’t have a good answer, this makes it all the more interesting for me to ask David Reich. Because otherwise if it just like given me the right answer, I’m like, “Okay, this is like a known thing. I’m not going to waste his time with this, right?” Now that I know it’s like not really clear, now it’s going to be such a fun conversation with David Reich. Yeah. And and you sort of know it’s not in his book because you’ve like asked the whole book, uh which is which is pretty cool. This is Okay, so here’s here’s what I kind of want to do. I want to see if it it is good enough at picking up patterns in how you ask questions that it can help you come up with questions that you think are actually pretty good for for for David Reich. And I don’t know if it’s going to work. Um are you are you down to try? Yeah, yeah, that’s a great idea. Let’s do it. Okay, so basically who have you interviewed recently-ish that is sort of like in the same vein as David Reich, like the same kind of person? I think the closest would be Tyler in the sense of like a more sort of polymatic less techni less like AI focused. So if I look up Tyler Cowen questions What about like any anybody that like is specifically deals with like genetics or like sort of population changes? Is Do you think Tyler like you you covered that in your interview? Oh, no, not at all, but I I just haven’t interviewed somebody about that in a while. Okay, so that’s new. So Okay, so So Ty- Tyler’s good then. I mean I I think I think Tyler’s a a good enough like broad enough range of has broad enough range of ideas that it that’ll probably work. Cool. Okay, let me pull up the questions. Okay, so here are the questions I asked Tyler when I have had him on the podcast. And it you know, it was like about basically all these different economists. So I read, you know, Keynes’s theory of interest, money, and employment. I read I read Adam Smith. I read like Hayek’s all all of Hayek’s essays. And I This was actually a super interesting interview. Um Um and you know, I was like asking about the contradictions between Hayek and Keynes. Anyway, so I think this will be a sort of And Mill, obviously. Um Um so this will be a super like hopefully it’ll get what kind of thing I’m doing trying to do if I add this to Claude. So we go back to prep. Um and then I add content upload from device. Or sorry, I I guess I should just add a text. Um Tyler Cowen questions. Yeah, did that work? Sweet. So one thing though I want to do first with this is like can you just ask it like given all the questions you asked for Tyler Cowen, like just ask it to pull out the patterns in how you like to ask questions and how you like to conduct interviews. Um do that first. say like the host of a podcast. Right. I want you to find the pattern. Maybe I don’t want to like prompt it with self-congratulation of a podcast. Um I want you to find the patterns in my questions. Um here’s uh questions from an interview of the economist Tyler Cowen. Um one thing I like to add to these is like um the like the results you give me should be so detailed that another AI who is impersonating uh can follow your output to generate um questions like this for a new guest. That’s a great prompt, yeah. Uh there’s patterns you give me to generate questions for a completely different guest. Different Nice. Okay, I like that. Yep. So you want to hear them out? It says you often ask about comparisons between different kinds of thinkers. You pose what if questions. Um you ask how historical economic theories might apply to current issues. Uh you present counterarguments and alternative viewpoints to test the strength of theories. You draw connections between economics and other fields. Okay, so maybe I’ll just ask it like I don’t always interview economists. Um Maybe you should add a couple more transcripts and see if yeah. Yeah. Find patterns that are not domain-specific. So let’s go to Demis. Oh, shoot. Okay, then we can do um Dylan. Sweet. Eh, it’s all right. So it’s telling me like um I I feel like it’s uh over um it’s being like specific to the kinds of interviews I’ve done cuz it’s asking about like timelines which I asked Demis about or technical bottlenecks, resource allocation, industry dynamics. I feel like that’s more about what I asked Dylan about with semiconductors. Um Yeah, I’m not sure how good this was. Uh But we we can try it. So Hm. What if we Yeah, yeah, yeah, what if we what if we said I’m going to I’m now going to interview a geneticist. He’s a geneticist, right? Um Can you write a guide like for another AI to prepare questions um Basically what I what I want it to do is like write the guide specifically for interviewing a geneticist based on the patterns in your um in your previous questions. See if it can do that. Yeah. this is much better. Great front. Okay, so it’s about um uh adaptability, um fundamental concepts, visual predictions, comparative analysis. All right, sure. Well, you know what? Let’s just try it. Um so, let’s put this in. Let’s put this in um Let’s try it. my Yeah, we’re a lot of a lot of different Um was So, we had Yeah, okay. So, we just go back to the If you’re just going to just go back to the project, yeah. Help me come up with questions. Oh, yeah, cool. with questions I might actually start it at start a new one and just be like And maybe I will delete this so it’s not over biased to that or You know let’s keep it. Maybe it has like maybe learn some context. Okay. Oh, yeah, it’ll have examples. It’ll have examples. That’s good. Examples are good. for David Reich based on these guidelines. Um Mhm. that another AI generated from other question list. And then um You can also reference Yeah, you can tell it to also you I was going to say you could also tell it to reference Tyler as an example, but Maybe one thing I’ll point out that I’ve that’s been sort of a big part of my um I’ve like noticed is that there’s a big hesitate like writing these prompts and reminding the AI like here’s who I am, here’s why I care about this, here’s the larger purpose of this project. It just ends up being a sort of big um big uh big sludge and that’s often what keeps me from using AI tools that I know would increase my productivity cuz I’m just like I don’t want to retype here’s you’re an AI and you’re trying to help me come up with questions. I’m interviewing blah blah blah and like write this as you would write to another AI. So, um I hopefully in the future we get models that are just like know all this about you and like you don’t even need to remind them cuz they they would have just been listening to this call and they didn’t know what we’re trying to do, right? Um but in the meantime, I just think it’s like a just like don’t be lazy and just like do the do the prompting. Um Yeah. Yeah, I mean I think that the chat GPT memory feature is like kind of getting there to some degree. I think like Claude projects you like it it can it contains custom instructions from one chat to another. So, that’s kind of nice, but yeah, I agree. I mean it’s there’s a lot of typing to do. But what do you think of the of the output from this? think so. Let’s see. So, it’s asked about uh it asks given the rapid advancements in the ancient DNA sequencing, what breakthroughs do you anticipate? Um How do your methods for analyzing ancient DNA differ from those using contemporary genetic samples? What are the current So, I one one thing I’m noticing is that um they’re very generic questions and that’s kind of what makes it not that useful for me is like just come up with questions for me because I try to ask um more specific like here’s It’s a more of like after having read this passage and like considering the research in this area, here are some thoughts I have like the 10,000 year BC thing was a perfect example of this of okay, we know that civilization emerged rapidly. Like what was going on there, right? It’s a very specific question and I don’t think that if you just in gen- generally these things just don’t aren’t that good at like come up with the specific thing from a large context. It like really wants to you know do a summary level or high level kind of um uh uh uh questions, right? I I I think you’re totally right. I think this is like a sort of common failure mode with um with using AI tools is like people end up having they’re like, okay, maybe it can do my entire job for me. And then it’s like no, it’s like too high level and it doesn’t it doesn’t really work cuz like ultimately are about what’s interesting to you and it doesn’t have enough context on you to like know what you are specifically going to be interested in just from the patterns in previous guests, especially especially if you don’t have guests that um cover the same sorts of topics. Like I think if you had like five other interviews with other geneticists, it might it might do a little bit of a better job. Um and I I think like usually the solve for that is like backing up and thinking about what are the like different micro tasks that I do as part of getting to those questions um that this could be helpful for. So, like instead of just doing the whole thing all at once, it’s like backing up to like the the the the particular context that I’m interested in currently. Like what are the micro tasks that I could I could replace myself And for example, the things we’re doing with like research is pretty helpful cuz like then you can narrow down on like I want to research X. Um but not like completing the job where yeah, totally. Yeah. Um so, you still have a job. We’ve We’ve We’ve determined that that as as far as Claude 3.5 Sonic goes, uh not replacing Durkesh yet. Um before I before I let you go, I have a couple I have a couple other questions for you. Uh really quick. I I mean I have to ask you what your timelines are. Uh if you have any updated timeline. But but but before we do that, should I show um So, I’ve been working on coming coming This is like the research, but also in post-production, there’s a ton of like we got to find the clips in this interview or whatever, right? So, Please. Yeah, yeah, I forgot about that. Let’s go over that. Um Claude 3.5 Sonic good for research, not going to not going to replace you, but you are also using AI not just for for reading, not just for research, but also for like helping you sort of in the post-production process of like putting putting episodes out there. Do you want to tell us about that? Yep, totally. So, um obviously there’s a bunch of things that go into the post-production of the episode, which I’ve been trying to come up with workflows to help AI help me out with. Um we uh before I started having a human make transcripts of the podcast and before I had ads to help me subsidize that, um what I would do is I would have a text-to-speech or speech-to-text assembly AI is the API we use or I use um have it come up with the transcript uh a first draft transcript and then I have um I have uh GPT-4 just literally writes this prompt transcript. I came up with a couple of methods to do this and it just says like um so, you’re and then I came up with some guidelines so that like here’s some uh prompts to make sure that when you rewrite the transcript, you’re removing filler words, you’re removing you’re making the thing more readable, you’re cleaning it up. And it actually worked decently well, but it just wasn’t good enough that it beat out a human yet. So, and then it’s like worth it for me to just have a human do it. Uh but other things we’re doing is we’re I’m trying to come up with a workflow where I can just upload the MP3 of the episode and the transcript it makes an auto-generated transcript. And from that transcript, we can just have it come up and generate um different ideas for titles and clips and highlights. And it’s a work in progress right now. Um but you know, we just do a sort of a few shot learning in the prompt of like here’s what good titles look like, here’s what good clips look like. And this is just a sample of two minute random interview, but like um we’re trying to I’m trying to like figure out ways to do get this kind of workflow going. That’s that’s really great. And and the reason I think it’s great is cuz we actually built this too at Every and we use it all the time and then we just like released it as a as an app. Um it Have you Have I shown you spiral? So, if you go to spiral.computer. So, basically we just built that into an into an app where you can create these things called spiral where spirals are like basically few shot prompts for repetitive creative tasks that you do. So, like for me when I do this podcast, I always have to come up with a tweet for it and this just I just do I put in some few shot examples. Um it creates a little style guide for itself and then you get a form that you can share with your team um or just use yourself or share publicly and then every time you have a new podcast, you just paste it in there. It can pull out tweets. It can do um it can do transcripts. It can do um like highlights and all that kind of stuff in your style and in your voice using using the few shot stuff. It’s it’s actually it’s it works really well. I use it a ton internally and it’s going kind of viral. Like we’re we’re about to pass 3,000 sign-ups for it. It’s been like a couple weeks. Um so, I would love You should Yeah. Can I set custom um prompts? Like I I want to generate a tweet using these guidelines or something like that. Yes, you can. Okay, I would love to use this. Okay, cool. You should use it. Um like I’ll hook you up with an account after this. You In fact, if we had to edit out this stuff I was saying about my drinky workflow cuz It’s like I’m trying to describe this to Steve Jobs. Like I you know, I I have this like little device and I’m trying to like it can I can open up different I can open up my cell phone Oh, I love it. Um Anyway, we’ll we’ll hook you we’ll hook you up with a we’ll hook you with a license and I I would love to see what you think of it. I I think I think it’s just like a problem that all sort of creatives face is like there’s there’s all this like like kind of drudgery of creative work that’s like not about the core thing that you’re doing and it’s like Claude like just got to this place where it can like um where it can like uh do a lot of that. And it’s like yeah, I just needed a for I just needed a little builder thing and you I sounds like you have you have the same problem. It’s like I don’t want to prompt it every time. Exactly. Yeah, yeah. I’m so excited to use this. Thanks. No, I Yeah, I’m just Yeah, I’m actually I’m going to experiment with it later today and it would save me so many hours per production of an episode to have this kind of stuff cuz you just like it’s hard to overestimate how much time these kinds of things take cuz I’m sure you know. Totally. Totally, I do. I I I love to hear that. Let me know Let me know if you have any feedback. Um So, before I let you go, I have to ask you a couple a couple important questions. Um so, what are what are your AGI timelines? Um if I I would give like a 25th percentile 70 75th percentile sort of um uh bounds and I would also say what do we mean by AGI? It’s like not just that it’s productive or can generate trillions of dollars value, but really it’s uh you can replace a remote worker. Any remote worker you can just replace with AGI. For that, I’d say like 25th percentile maybe 2029 and then 75th percentile like 2050. Mm, okay. Interesting. And then what about P-doom? It’s hard to um yeah, it as far as definitions go. So, something along the lines of like the thing that’s taken over doesn’t really have any sentient experience and it just like doesn’t have culture, it doesn’t have individuality, it doesn’t really um it’s not just that humans are disempowered, right? Cuz I think like humans disempowered chimps and I think we didn’t doom the universe. I think it’s like um it it just has to be like some the paper clipper that uh and what is the odds of like something like that? I don’t know. Uh 10% or something like that. I It just like it thinks that it could go really wrong. Um That That makes sense. Um that actually uh that’s I can’t I can’t decide whether that’s higher than I than I thought it would be or lower than I thought. I think it’s higher than I thought it would be. That’s like That’s a kind of pessimistic think it is like fair to be like that’s crazy, you know, like 10% odds that everything we care about is going to not exist in the far future. Um I think that’s like the right reaction to that, honestly. Um cuz often it’s easy to bandy about these numbers as abstractions are like I don’t know what what are the odds of Patriots winning or something and no, we’re really talking about technology that will be alive in our lifetimes um and might result in something really bad. And it’s worth taking that seriously. Yeah. Um So, I really appreciate you coming on. For anyone who has been listening to this and does not know you yet and wants to find out more, where should they find you? I have a podcast called The Dwarkesh Podcast um available YouTube, Spotify, Apple Podcast, wherever. And if you want to there’s also the newsletter as you saw, hopefully. Thanks to Dan’s prodding, I’ll be doing more writing and um that is at dwarkeshpatel.com. As for um and then you can also follow me on Twitter or something, so that’s the dwarkesh_sp or just look up Dwarkesh Patel on Twitter. Amazing. Thank you so much. This is awesome. Thanks so much for having me on, Dan. That was That was super fun to go through that workflow and also get your tips on how to, you know, use these tools better. I’m actually pretty excited for people to see how um yeah, how how these tools have been useful in my workflow. So, this I’m excited to for this to be out. Oh my gosh, folks, you absolutely, positively have to smash that like button and subscribe to AI and I. Why? Because this show is the epitome of awesomeness. It’s like finding a treasure chest in your backyard, but instead of gold, it’s filled with pure, unadulterated knowledge bombs about ChatGPT. Every episode is a roller coaster of emotions, insights, and laughter that will leave you on the edge of your seat craving for more. It’s not just a show, it’s a journey into the future with Dan Shiper as the captain of the spaceship. So, do yourself a favor, hit like, smash subscribe, and strap in for the ride of your life. And now, without any further ado, let me just say, Dan, I’m absolutely, hopelessly in love with you.