A $10B Hedge Fund’s AI Playbook (Best of the Pod)

It's becoming a meme that CEOs are like writing basically like the where AI first memo, you have the best example of that memo that I've ever seen. Would you mind just like reading, I don't know, maybe the first paragraph or two of what you wrote, because I think it's amazing. Using ChatGBT is not cheating. That's a non applicable idea from academia. I use ChatGBT to write this email, you should be using it too, and be proud of it. As a hedge fund, we should be ashamed to leave money on the table by ignoring tools that make us faster, smarter, and more effective. From the very top, we are building a culture around AI. Not using these tools is like refusing to use the internet in 1995, because it wasn't perfect. Will, welcome to the show. Thanks, Ben. Thank you. So it's great to have you for people who don't know you. You are the benevolent dictator of walleye, which is a close to 10 billion AUM hedge fund. And you're an every consulting client, we're working with you to help you do AI training and implementation inside of walleye. And honestly, like regardless of the stuff we've done, you're, I think one of the most impressive examples of someone, a CEO who is like pivoting their entire organization around AI, and you're sort of like leading by example. And so I'm psyched to get to talk to you. Maybe I can just give a little more context to that. I guess my technical title is CEO, CEO of a managing partner, but I'm both the owner and operator of one of the larger hedge funds out there. I don't do many press, I don't typically speak at conferences. I've only done one other podcast. So this is very deliberate. It's very deliberate for me. And because of the relationship that we've developed on this subject, I do feel a great sense of both purpose and conviction about where our industry is going, where our firm is going. I do believe that we're already a leader in that is going to continue to be the case. So sometimes I listen to a lot of podcasts. I was curious about some of the motivations for why people do this. I just want to be very clear up front, because I have that conviction, the skills to lead us into the next phase. We'll get into that, of course. And ultimately, the power to enact that. I really feel that it would be irresponsible of me not to go after it with the maximum amount of discipline and intensity I can muster, which is a lot. So for your main audiences here is the people at Walleye. We have about 400 people, not big for a normal company, but in our corner of the world. That's not a small amount. I'm second also speaking to people that aren't at the firm today, but may join us in the future just to understand how serious the firm is about AI. And that ultimately starts with me and then others. And then third, finally speaking to people in the ecosystem, whether they be companies building products that firms like us can use or in other ways of work together. That's why I'm here. But it does all start with me having, as I said, an enormous sense of conviction about where the world is heading. How did you develop that? What was that journey like for you? I actually don't even know this. So I'm a bona fide nerd by background. I was an engineer at Princeton. I went to Oxford to do a PhD in math. I started my career writing code all day for algorithmic strategies. So we as a firm and me personally have been using, let's just call it advanced statistics, not even AI for years. A large part of what our firm does is in pure quant trading. So I've thought for many, many years about how machines increasingly can both either augment or do the jobs of humans in finance. That's nothing that's new for me. What has happened, of course, in recent years is just that a lot of these tools have become more powerful and they've also become more accessible to non-technical people, particularly on unstructured data. And so part of it is just me personally being curious, being interested. A huge part of this is about curiosity. I've noticed the productivity improvements that I see using these tools, just even from a writing perspective, let alone some of the sort of more advanced things you can do. So it's a combination of having been deeply versed in technical background, as I said, bona fide nerd, and then noticing what's out there and just the sense of responsibility to the people at the firm, to our investors, to say, look, we understand where this is going. If we ultimately get disrupted by AI, that's on me. And so let's get after that. So even two years ago, this isn't a sense of, oh, yeah, we're going to do all these things in the future. Two years ago, we started an internal AI program. Today, we're already doing a lot. So it's not just a story about what's to come. And yeah, a lot of that does start with me. I mean, a lot of times, leader of organizations won't sort of say that how much actually matters in the leadership. But this is one where I feel like I absolutely needed to lead from the front. So tell me about that a little bit more concretely. Like, one of the things I appreciate about you as a communicator is you're just kind of like no bullshit. We were talking like, I don't know, three or four weeks ago, and I was telling you something about Everee and our strategy and direction. And you were like, you sound kind of afraid. I was like, yeah, I'm a little afraid of this. So you're not afraid to kind of like put your finger on the nerve. And you're also not particularly like hypey as a person, I don't think. So tell me about the moment where you're like, oh, we have to take this seriously versus it's just like, yeah, tech people, nerds are psyched about it. But like, it's something that we really need to understand. Yeah, and I appreciate you saying that. That's definitely the way I try to operate. Many people tell stories. Sometimes those stories get hyped. And for whatever reason, I think a lot of it comes down to being comfortable with yourself, self-confidence, and not feeling like you need to convince something of someone you're not. So that's been a hallmark of mine. I would never have done that. So I don't know from experience that that's a problem. Well, really, that's a core belief. In terms of AI, there have been a couple of moments over the years that have led us to continue to go down and recently accelerate this journey. The first was two years ago called our most advanced internal AI project is called Current. That's led by someone who was a former analyst and one of our TMT for technology launch for stock picking teams comes to me in March of 2023 and says, look, I've used these tools that are now available to make myself way more efficient, eventually trying to replace what I was doing as an analyst. And I was pretty skeptical at first, because that's not typical that someone would just go out and do that. This is with GPT-3? Yeah. Yes, it was like two years ago. And it was early. But the point was, just sat down, gave me a demo and was like, wow, this is where we're going. And yes, you need to have a sense of belief. You need to imagine. You need to be able to dream. But if you understand the problem you're trying to solve and the tools that are available, this is absolutely where we're going. And so that's when we started an effort to say, OK, for fundamental investing, absolutely, there should be agents that are ultimately helping with analysis and ultimately to provide meaning to what's going on. So people talk about that now, but that was two years ago. That was a real moment for me. And then more recently this year, I've seen a lot of podcasts that consume a lot of information. And I was listening to one, actually, with Chris Saka, who's a very entertaining individual, sort of talking about, with, frankly, a bit of hyperbole, but entertaining hyperbole, just about where we're going. And I was like, you know what? That's so true. What these machines can do now is incredible. And it does take a little bit of imagination. But in some ways, it's kind of like you can see the endpoint more easily than you can see the steps to get there. I was using this analogy recently. It is a little bit, well, I'll just tell it to you. But I really did use this one. So my favorite movie growing up was The Sword in the Stone, made in the 60s by Disney. And I have three little kids, and that's one of their favorite movies. So I was watching them recently. And so Merlin, in the cartoon version of The Sword in the Stone, he lives backwards in time. And so he can see glimpses of the future, but he doesn't see the steps in between. And that's kind of how it felt like here. It's impossible for me not to believe that in five years, firms that do what we do won't be heavily, heavily integrated with best of breed AI technology across the firm, not just for investment, for non-investment purposes. And I say that in our industry because it's one of, if not the most clear associations between information and ultimately money, the value of having an edge from an information standpoint is huge. And so that's why finances is really going to pick up and actually use these tools and say, OK, what applications actually give me an information advantage? So that's very clearly where we're going. But the steps to get there can be a bit hazier. And so actually, it was after this podcast with Chris Saka that I wrote the team. This is how we ultimately met you. It's like, we're going to train mandatorily every single person at the firm. Doesn't matter what department you're in, how technical or not technical you are, you're going to have the base case level of proficiency in AI tools. And as part of this actually does come from a sense of responsibility. People are anxious about what are all these things going to mean? How does that impact my job? What skills do I need to have? And me saying to the firm, OK, we are going to be a place that is going to be leading that, that we will actually train you. We will give the tools available. You still have to learn them. We'll make it accessible and make it accessible to everyone. So that was when we just started doing a lot more. Not just building, call it, tools using advanced AI. And as I said, we've been doing that in our quant business for 10 years. And not just building, essentially, digital analysts to replace the work of humans in long-term stock picking. But then over the past, really, this year, I've said to everyone across the firm, even in accounting, finance, compliance, legal, yes, you should be using either ChatGBT or CROC or some LLM to assist in anything that you're doing that's analysis or writing. We record pretty much every bit of information that flows through the firm. A huge part of this is actually having a proper data strategy. And then also just having the culture of, yeah, we don't really know all the answers to this, but let's just start talking about it. Let's make it accessible. So simple things like having weekly emails where there's leaderboards of who's using these tools the most. Actually, I sent the entire firm an email and said, hey, if you suggest a tool that ultimately we end up pushing out across the firm, just like there's incentive systems for employee referrals, we'll do something similar here, too. We have weekly meetups, informal weekly meetups internally just to talk about AI, be it prompts or other use cases. One of your previous guests talked about the social nature of AI and how hard it is just to even discover best use cases, which I couldn't agree with more, but just even internally trying to make it a bit more social, a bit more accessible. And I'm involved in all of this. In fact, I'm sort of the chief AI evangelist. And a huge part of that is just my own personal curiosity, but you can already see it working. People are doing things that they weren't asked to do. It's so cool. And the productivity coming from that, it's real. This isn't just paper. So yeah, we're definitely going down this path. One of the things I think you've been so effective at doing is basically, you can think of companies, companies even 10 people, but 400 people bigger than that. The bigger they get, the more hard to steer they are. Maybe a startup is like a canoe and a 400 person company is like a cruise ship and a 10,000 person company is like a battle cruiser or whatever. And I think you've done a really incredible job of pointing the cruise ship really quickly. And that's something that's happening a lot now. It's becoming a meme that CEOs are writing basically like the, we're AI first memo. Yeah. And I think you have the best example of that memo that I've ever seen. Would you mind just reading, I don't know, maybe the first paragraph or two of what you wrote? Because I think it's amazing. Yeah. Yeah. I can read that. And on that point of, as I said, words are cheap, particularly in a world of AI. You can create great words very easily. To your point around being a bit of a cruise missile, I do think that's an advantage of ultimate governance. I'm the owner operator. So I don't worry about getting fired. I worry about doing what I think is right. I very much believe this is right. And so we can just go and do some of these things and large organizations just can't operate that way. So as I said, a huge sense of responsibility because we can act that way to do it. So yeah, I can read a couple sentences. Here we go. So yeah, so this is an email that I wrote to the firm and the entire firm. And the subject is AI at Walleye, a challenge to all of us. It says, I use chat GPT to write this email. You should be using it too and be proud of it. I'm writing this as a follow up to the comments I made at the town hall at the start of the month and after our AI Senate meeting earlier today, which is a group of forward thinking AI users from departments across the firm. And the message is simple. Walleye is all in on AI. Not using these tools is like refusing to use the internet in 1995 because it wasn't perfect. That's just dumb and something I can't understand. As a hedge fund, we should be ashamed to leave money on the table by ignoring tools that make us faster, smarter, and more effective. Using chat GPT is not cheating. That's a non-applicable idea from academia where using AI to do homework or take tests is actually cheating. In the real world, using AI is like taking a magical elixir that makes you 20% smarter instantly or a lot more. So why wouldn't you use it? From the very top, we are building a culture around AI. This is not optional for anyone at Walleye. If you write, research, analyze, build decks, process data, or think for a living, you should be using chat GPT and or other AI tools every single day. Managers, this is now part of your job. You need to be pushing this across your teams, starting with being fluent yourself. So here's what's coming. And I talked a bit about some of the things we're doing recently that I just mentioned. And then ended, you know, this is the beginning. The edge is real and we will not fall behind. Let's lead. I love that. Like what was the... You hit on a couple of things there that I really love. Like one is this cheating question that a lot of people have. I have that too. It's just like an internalized sense of like, oh my God, this might be too easy. And when something's too easy, you're like, am I cheating? And also sort of like addressing this fear that I think a lot of people have, which is like, am I going to be replaced? And I think the way that... going to be replaced. And I think the way that you're talking about it is very like, actually, this is now part of your this is part of your job. It's not that your job's gone. It's that your job is changing to include this as an expectation. Yeah, yeah. So I have very strong views on both of these. You know, as I've gone on in my career, you know, my, my role has changed. And I believe that for all effective leaders that as they go on, they think they should evolve as well. You know, I, I love the Jim Collins quote, you know, build, build a clock, don't be a timekeeper. And so this is an example of that where if you can have a tool that makes you more effective, it's the same thing as sort of hiring someone to replace part of what you were doing so that you can move on to the next task. So your context level shifts up. And so I'm not embarrassed at all, that I can write emails that used to, or, or long memos that used to take me hours that maybe I could do that in a half hour or less. And that to some extent, people have this insecurity that all if I didn't put my blood, sweat, and tears into it, that somehow it's not, it's not real. But at the end of the day, you know, results are what matter. I think having I spent a lot of time as an athlete, and I still do a bunch of stupid stuff in the gym as my hobby. At the end of the day, either you pick up the weight or you don't. And I think that's just the attitude that people need to have in the business context, as I mentioned, and there's this huge difference between academia, where people like, Oh, my God, it is ruining school, which in the call of existing paradigm, you can certainly argue that, but that's not business. And just being very clear about that, you should be trying to be as efficient as possible, not so that you can just leave work at 2pm. So you can actually spend time thinking about next level tasks, higher level contexts, be a bit creative. And ultimately, think of yourself, even if you're an individual, and you don't manage any human, you're still going to be managing employees, a lot of those employees are just going to be AI robots. And that's a skill in of, in and of itself. So just the reason I wrote that to the entire firm is just making people feel comfortable and not kind of embarrassed, or it's like, because what I was seeing before is people were, you know, using chat, dbt, or another similar product to create an email, and they were trying to like, dust it up, it's like, Oh, I don't want to be seen as doing that. So like, it's just stupid, you should be doing that. And if you aren't, like, why did you waste three hours writing this thing? So, again, very, very strong feelings about that. At the same time, you're coming around this anxiety of what is my job going to be? You know, when spreadsheets came out, or email came out, you name it, like, yeah, you have to learn how to use these, these tools. And I see this all the time, very deliberately, we hire people or you say are an inflection point of their, their career, typically mid career where they, you know, know enough to be dangerous, the competencies there, but the hunger is still there. And the ability to be dynamic to learn new skills is also also still there. That matters a lot. And I don't think there's any different. You know, if you're just using example, if you're a long short stock picker, and you can't use Excel to write a company model, you're totally obsolete. And in future years, you're going to have sort of the same thing where if you don't know how to use these tools, or you're not at a firm that can give you access to the tools to give you operating leverage, then you're also gonna be obsolete. And so that phrase is what I use a lot is just this, this concept of operating leverage and not being afraid of that. But yes, absolutely. I see that all the time. And it goes back to my earlier comments around feeling a sense of responsibility to put people in a position to say, like, don't be afraid of this stuff. Embrace it. And another thing that I was seeing, Dan, a lot of these tools aren't, they're not perfect. I mean, none of them are actually perfect. They all have their flaws, they all those stupid things. But the direction of travel is very, very positive. And so I wanted to set culture and environment, or instead of people being like, oh, it didn't do exactly what I wanted. So I'm just going to ignore it until it's perfect. It's kind of on the opposite way of being like, yeah, let's have, you know, a demo with a third of the company joining, which is real, that's how many people join. And if that demo screws up, who cares? Like, I basically had one last week that the demo gods were against me. And having people accept that embrace that and be like, yeah, this some of this stuff isn't perfect, but you can see where it's going. As opposed to just almost using that imperfection as an excuse, excuse to ignore it. So all those things combined into one, there's there's no other person besides me. And I think this is true of all leaders, the organization, like you have to lead from the front, you know, no one, no one can set that tone, besides, besides the leader. And once that tone has been been set, it's, it's kind of incredible how much you can unlock people to be like, yeah, have at it, and it's okay. So setting that tone, and flipping it outside be sort of, you know, you need to be afraid, or a lot of people afraid not to, or to make a mistake, kind of need to be more afraid to getting, getting left behind. And what have you seen, like, for someone who's watching this, and maybe he's in finance, or maybe he's just running another company with a lot of people and is thinking about, okay, like, but really, what, what productivity gains has it actually unlocked for you? Like, concretely, what are a couple of things that have been useful for you or for the fund? So there's a couple things, you have to put this into categories of what's, what's recent and what's not. So as I mentioned, you know, we, we do run a big quantitative trading business, you know, as a multi strategy firm, we run many different strategies, but quantitative equity trading is a big part of what we do. Those models have used, you know, nonlinear statistics, AI, or some of the underlying models in AI for years, more recently, with the advent of large language models coming about the ability to process, you know, unstructured data, of course, and incorporate that into signals, which historically was called sentiment analysis, you know, the ability to do that at scale has gone dramatically. So that's one improvement, just in a pure money making standpoint. And you're doing that, like you're using language models to do some analysis. Yes, we get we absolutely are doing that and have been doing that for years. And frankly, all world class quant firms are doing that. But it's hard. That's why quant trading is one of the things that definitely benefits scale. Some of the things that are called more recent or newer. And I do believe that the explosion that we've seen, it empowers the less and less called technical people that the technical side where really you just have to be creative. So some of us were doing like, you know, 75% of the firm, you know, is an active called chat, gbt, chat, gbt, like user, you know, every single week, like, actually, almost every single day, that's pretty cool. About a third of the firm, you know, use AI coding tools, such as windsurf. That's sort of very real, are, as I said, an internal product for fundamental launchers, stock pickers, you know, that's, that's also a big part of our business. Every single team uses this tool is called current, it spikes dramatically, you know, during the user spikes medically during earnings. We have people that that come here from our competitors and tell us that this is, you know, it's both way better, but but really an essential part of their of their job. And I believe that and I think our competitors probably do have good products as well. I think we've just been doing a little bit longer and are further down the path of using these tools actually to provide provide meeting and real analysis as opposed to just summarization. But yeah, you know, you kind of can't go through earnings period now, as a long for stock picker, without some of these tools, if you're going to be competitive, because, you know, what one of your competitors, like someone here is going to be able to process all of them in real time, and then have a machine go and basically impute things that humans can't do as fast. So how do we actually measure that, of course, having benchmarks, like, it's kind of nonsense in a real, real company. But just sort of seeing the level of level of adoption, just seeing people suggesting products like, okay, we want a, we thought we talked about last week on our one of our meetups to have a product that can do summarization via podcast to make it more accessible to people that want to listen to something. And then that day, we had, you know, five different products in beta, and then the next day was pushed out across across the farm. So some of these sort of cultural elements of just having actually set up a process where we can both incorporate third party products and build some of ourselves. You know, that's pretty cool. And I don't know how much smarter, more efficient that's, that's making people I certainly could speak from for personal experience. A big part of running an organization is communication. My communication is dramatically and which a lot of in a lot of cases, really, that's dramatically more efficient now. When I was when I was using these tools, and I think it's probably true for literally every single person in our firm. What does that actually look like for you? Like, when you're using it to communicate? What are you doing? I think best when I write out my thoughts. I tell people that my education is very expensive. So I better be able to write, or else what the hell is all that for? And so historically, you know, I really would write to convey what I'm thinking where the firm is going, why. And I just believe that leaders should be able to really communicate. And I believe in the power of high quality pros. And it's actually one of the things in general, before LLM came about that, it was pretty pressing that younger people just were terrible, terrible writers. So for me, when I want to, you know, write an email, as an example, write, write a memo, I do it in bullet point form, I type out, here's what I'm thinking, here's why, you know, I work on my prompts. And maybe I'll give a bunch of context of, hey, you know, here's all the stuff that I've written on a similar subject. And I want you to give something in my my own voice. And that can be, you know, no exaggeration, like a 15 minute process that would have taken me four or five hours. Historically, that's why I have such, you know, almost religious views on this stuff, because it's, it's, it's pretty wild. So that's, so that's one example where I'm using that personally, you know, another thing that we do. And yes, there are always the questions on what can or can't be recorded, it's kind of going back to the governance element where I could say, this is what I believe is the right thing we're to do. So internally, with a few exceptions, we really record every single zoom every single call. That's just the nature of our industry. And frankly, I think the whole world should get used to doing that you're seeing these articles about people wearing wristbands, recording every everything they say for, for months and months at a time, you extend that forward, without sounding too much like a nut job, I think people are going to have essentially recording devices implanted, you know, their bodies that record everything. So just getting comfortable with the fact of, yeah, like all this data is going to get get captured. So we're trying to do a bunch of that internally. And then just being able to go back and process that, because so much of the power of this is, do we actually have a data strategy, get all the data into a lake, which you can then put a straw into and get it out. So, you know, a big part of my job overseeing the risk of the firm, the chief investment officer title, you know, every single morning me and my, my risk teams are like in the in the control center of running this, this giant process, you know, we have a risk calls, and those are all recorded. And we can go back and say, Hey, you know, what were we talking about at this time, and continually have LLMs that are processing those transcripts and helping, helping us to both remember and provide insights and ultimately be a bit, a bit predictive, which has been hugely helpful just just in that exercise, which is, you know, we haven't sort of talked about where I think this is going and the power of all this. And I mean, like, we're, I do believe that we're, that we're a leader, I don't want to say we're the leader, because I definitely don't know what other firms are doing. But I certainly think that we're a bit more advanced in our thinking of how to use these tools. But we're just scratching the surface of what what is possible, once you actually start connecting all bits of information within the walls of the firm. And this is not just, you know, not just Walleye, not just hedge funds, really any company of saying, Hey, let's actually put all of our data together into effectively a collective and then that that information get processed. We're totally just scratching the surface there. But we're certainly working towards that. Everyone's talking about how AI is changing how we work. For anyone working in sales, you felt it. The whole job is changing under your feet. But one thing isn't changing the when the deal that finally closes the moment the hard work pays off. Maybe you've got a gong in your office or confetti machine or something else we don't judge. Every team has its rituals. That's the moment audio is built for it's the agenda CRM that runs the work behind every win. It turns every customer signal into context you need to move at unmatched speed and scale with revenue agents and automations that build pipeline advanced deals and grow accounts around the clock. You'll get all the context and velocity you need to scale any go to market motion so your team can raise the ceiling on what's possible. Start with audio today and just focus on the win from here. Go to audio.com slash every and get 15% off your first year. That's a TTI o.com slash every I want to go back to something you said earlier about writing as thinking and using language models to turn like a you know, four or five hour task into a 15 minute task. Yeah. What is your like, one of the things I worry about, for example, is maybe I'm not thinking it through as clearly if the language model has like written a bunch of stuff that it's coming from my bullet points, but I haven't like really gone through every single thing and been like, I stand behind that. Yeah, sure. I think this out in terms of, you know, thinking through the concepts versus the linguistic syntax, what I was noting, at least personally, and I do think a lot of people do this as well. When they're writing there, they're trying to be both consistent to some extent clever and to some extent, unique to their own style. And so a lot of editing can be I think less about the concept and more. What are some of the nitty gritty details of how you stitch sentences together, even simple things like it drives me absolutely nuts when someone ends a sentence in a preposition and everyone at the firm knows that. But you don't you don't have to spend as much time again, I think on the the important, but not as powerful tasks of writing, like a lot of it is just sort of stitching, stitching these pieces together, the tying your shoes part. So the principle, the, the elements of writing, you know, what are the concepts that I'm looking to, to convey? That's what I spend my time on now. So what I found with these tools is really trying to be clear, like, this is the concept that I want you to get across. And this is how and then yes, it will suggest, you know, a string of words that, that convey that. And particularly with the way the recent models are architected, it's as everything that I've, I've written, so it can do it to some extent in my own voice. But I just don't have to spend as much time like, frankly, trying to be clever. And that's what a lot of writers do. They try to say a lot of relatively straightforward concepts in a clever way. I just don't think we need to waste time on that anymore. I would never, never do that. As a writer. I'm curious, like, but let's, let's flip the table a little bit to like, when you're reviewing someone else's work. So for example, for me as a manager, I think like, let's accept, for example, like, if someone's going to publish something on Every, that sounds like it's AI written, I don't. on every that sounds like it's AI written, I don't, that's just out for different reasons. But like, if I get an internal report that looks like it's written by Chachi BT, and I did this actually last week, because everyone internally is using these tools all the time. It's not that I care that the voice sounds like Chachi BT, it's that it's not clear to me that the person has thought through the thing that is being presented to me. And I don't want to spend time reading something unless I know that a commensurate amount of time has been spent thinking about it first. So how do you like deal with that? Look, these tools don't negate the necessity to think. I say that all the time, if anything, they should just give you more time to think like, if you say, okay, you have an hour to complete this task. And it used to be historically, I don't know, 50 60% of that time, it was just going to be mechanically typing out. And now, 5% of the time is me doing that. So you have more time to think just in a fixed fixed amount of time. So you should really think you should you should read your proofread and say, does this make sense to me? Is this what I'm trying to convey? And so I can definitely tell as well, when something is written by a machine. Sometimes that's just the way that the text appears like the bold, like clearly a human didn't go and bold it in exactly this way. But that's fine. But it's not enough. It's not sufficient. You still need to convey the concepts clearly. In an ideal case, someone has clearly used these tools, but the concepts still come from them. And I can tie it back to that person. And there's a why of like, okay, why are you doing this? Why does this make sense? And you didn't waste your time doing something that wasn't necessary. But at the same time, you didn't just outsource all of your brain to a machine. And sort of there's that optimal point on the curve that we're trying to get to. And that's, again, why I don't I don't think people should be totally, totally afraid of using these tools, because by themselves, I don't think they're sufficient. I think that, you know, it's like having a very powerful jet, jet engine, excuse me, and you use that analogy to me as well, like a jet engine won't fly by itself, you still got to hook it up to the plane. And there's a hell of a lot of things that matter when it comes to aerodynamics and you know, that make a plane efficient or not. So humans can kind of design the plane a little bit more and someone else brings the engine, you can use that engine in very powerful ways. But you need to be a part of the process, for sure. I want to talk about that. The thing you brought up next, which is sort of this data lake idea of like sort of recording everything, you've been calling it the Borg as a Star Trek fan. So like, where, like, you're recording all the meetings now, which I think is awesome. And you said it's already helping like in your, for example, in your risk calls, you can tell like, how you made a decision, like, can you give us a concrete case where having all those recordings has actually been helpful? So yeah, the Borg, which just come from Star Trek. And I'm kind of sad now that when I say the word Borg, even some real nerdy people don't even know it means I'm getting a little bit older. And the collective, which is you don't get get all the information together. That's, that's at least our spirit animal, our spirit guide for the future. It is really hard. I think all companies are going to have this and some of this is not at all particular to finance. It's just like, there isn't even a great way to process all the firm's emails right now using AI, which I'm sure will be solved soon. So the most salient example of where we've done a miniature example of this goes back to set our internal product current, which, you know, it takes analyst notes, all you know, information's coming in from, from brokers and these PDFs that get emailed around all the time, earnings transcripts, really any bit of information that's, that's germane to, to a stock. That's, that's our, our most advanced call it Borg example. And that, as I said, that really is real. Like all of our, all of our PMs view this as, as an indispensable tool that saves them a ton of times, particularly when information flow is, is very fast. Again, quantifying that exactly, you know, there's no perfect metric. A lot of it is, is definitely subjective, but I can see the internal use case numbers and I can also see like firms, all external firms know that we're, that we've built this and are doing it. And over 50 of them have asked like, Hey, can we, can we be a beta user of current? We'll give you feedback to help make the product better, which we, which we've done in some cases. And it just also makes sense, right? That a huge part of the job of a human is synthesizing information. And until recently, like when it comes to reading documents, like machines can do that very well. Reading documents or listening to the voice, essentially other text forms. But now machines can, and now the servicing the second order, the third order effects of that, again, not just summarization. That's what machines are starting to do. And that's the said, that's our main use case. And so this, the broader idea of the Borg, I look at what we've done to just help, you know, our long, short stock pickers and say, that same concept should be able to help every single department at the firm. That's, you know, generating text is a simple example, through emails, through through Slack messages, through, you know, live, live calls, live conversations. And then ultimately, the ultimate goal is to tie that back to numerical data, whether that be market data, internal data, accounting data, you name it. So the applications, as I said, do take time to to imagine, to design. But as I mentioned earlier, it's not that hard to see where this could go in the future, when you do have these sort of miniature collectives across all departments of the firms and other firm of any firm, and linking them together. Like that that will happen. It's just a matter of how to get there that I think everyone is still trying to figure out. I know that you're a student of history. Do you have any historical periods or examples that you're turning to to kind of help you navigate what's going on right now and this transition that you're going through? I'm a student of history. I do love basically the period between the Civil War and World War One is a time when I think the whole world changed dramatically. And that part of that is that my my office looks out on a train station built by one of the robber barons. So I do think about it all the time. So it's certainly not the only period of history. But, you know, definitely a time period where things changed dramatically. I'm not a VC, thank God, because I think most of them don't know what they're doing. Certainly this idea that when you look at a, you know, an exponential curve, you know, humans sort of nose hits, it's up against that don't realize how, how fast things can change. So there are periods of time, like how fast the railroads got connected, or reddened, how fast you had transatlantic cables, and what that actually meant. You know, that, as I said, the period between Civil War and World War One is just huge, huge amounts of change. And people within their own lifetimes sort of went from having relevant skills to obsolete skills, that is going to happen faster, this go round. And when I said, I think I'm not a VC, because you're a lot of people, investor types provocating about the project case thing about this, but they aren't actually involved in any operating companies and don't realize that someone still needs to go out and build all this stuff. But I do think the sort of first principle arguments of, yeah, things are going to change dramatically, you know, corporations, collections of humans, let's just say, companies in the future that want to operate in a world class manner at scale are still going to need many, many thousands or more of employees. But if you have those employees are going to be humans, a lot more of them are going to be to be machines. And so you've certainly seen that level of disruption and other other areas that just have a little bit faster this time. But at the same time, I'm an optimist in general. Like that's very important for the for leaders actually to have an optimistic tone. I don't think the world is, is going to end because all of a sudden, people are gonna have their jobs disrupted by AI that they need to adapt. It's sort of having a level of realism around that, if that's that's what I mean, our firm is a microcosm of that, if, yeah, you have to learn these tools, or in whatever time period, you're not gonna be competitive. And we are at the tip of the spear from a competition standpoint, just given the nature of what other industry and what are what are types of firms actually do, but I think it's gonna be true across a bunch of water swaths of population where you got to be trained to be to be efficient. And it's important that if you don't, that's your choice. And that just is what it is. Yeah, I think that period of history is so is actually is actually really relevant. And coincidentally, I've been I told you this already, but I'm sort of in my cowboy era. And I've been like reading a lot of cowboy stuff. And we watched the Netflix thing on wide open cowboy war. No, should I? Yeah, it's really good. It's really good. It's it just came out. Yeah. Okay, I'll check that out. I just finished Deadwood, which I was telling you about. I've never watched like any TV, but this one came out. And I also love sort of the old old West. It was a good story. Do you know why the why cowboys disappeared? I just learned this and it was a really interesting fact. You tell me I have a hypothesis, but you go first. Barbed wire. Yeah, I would say the broader count. Well, in the documentary that the cowboy war what? Basically, the answer was civilization kind of came in, you had the railroad come in, which brought a lot of people and then, yeah, eventually barbed wire and you couldn't steal cows. I mean, the cowboys were a gang in Arizona, in the 1870s and 1880s, stealing cows, especially I didn't know that. Oh, yeah, like the cow. So it's fascinating. But the cowboy war, and wider, which is this historical figure of legend, you know, gets in a this huge fight, like the movie Tombstone, which is kind of historically accurate, but not really. It was sort of the wider posse versus the cowboy two words posse. And it was this big, it ended up being this big deal, like that's the gunfight, the okay corral, because it started up all this sort of North, you know, North versus South sentiment 20 years after the Civil War. But the broader historical context, there is people sort of wanting to bring about change, because Tombstone where, you know, okay, corral was and where I was was a silver mine. So it brought in all these, you know, people from across the country, both North and South, but there's this huge tension between those wanting to modernize, and those wanting to get stuck in the ways of the past. So, yeah, it's, it's that period of time is, I find it fascinating, too, because you had sort of land with no laws, all of a sudden becoming civilized at various different paces. And a lot of cool things are interesting things, at least happening because of that. Yeah, I think that's, it's, for me, I love that, because I think it's such a good metaphor for technology and technology frontiers, and kind of this trade off between, you have like the individualists who are going out and exploring, and there's no laws, and there's a lot of creativity and all that kind of stuff. But then you kind of need the civilization that comes behind them. But that sort of, it's at odds with that frontier spirit. So there's that, there's that, always that tension between structure and creativity. And I think there's something very similar there about technology. I mean, that's even true with sort of, at least historically of the, you know, why does VC investing exist? You know, why is it that you go and read about the story of, you know, you pick any company, you know, from Nvidia on down, where they're like, I can't do this at a big company. So I'm going to go and push the frontier in a world in which I'm less constrained, you know, there's no barbed wire, and I'm going to go and build. And then at some point, you know, those, those companies become successful, they become institutionalized, and then someone does that again. And so that's not a geographic frontier, but but a technology frontier. And as I mentioned earlier, like I feel a sense of responsibility, because we can kind of do both. And they're just not the only companies that can say that if we want to push the frontier, but with resources. And those ultimately are, I think, the businesses that when you look at any technology transition are able, you know, not just to adapt, but, but to thrive of having that mentality, but not being, you know, not using single action rifles when other people have machine guns. How do you think about how the past informs the future? And I'm I'm asking this both from like a kind of, I don't know, late 1860s to now perspective, but also from a from an investing perspective. And, and, and, and I think this this layers into the AI stuff, where it strikes me that a lot of I'm curious what you think about this, but a lot of investing has to reliance to some degree on the idea that certain things that happened before are going to happen again. And and part of the investing part of being good at it is knowing which which one's going to happen again, and which one's not. Do you do agree with that characterization? And how do you? How do you think about when to when to rely on past patterns to help you understand the future? I don't think human nature has changed in 10,000 years. You know, maybe it's evolved a little bit, but you go back and my son who's eight is really into to Egypt right now. I love Roman history, because I took Latin, and I'm sort of a real nerd. So I know a lot about Rome. But you can read a lot about the two like human nature hasn't changed in a long period of time, you know, you go and just use a famous example, like you, the old always wants to hurt that the new always wants to replace the old, you know, the, the son wants to outdo the father, these, these are timeless, you know, when Alexander the Great was conquering Persia, with their eyes, the king of Persia sends him a note, and he's like, Hey, how about we have a truce? And he responds, like, I'm coming for you. Yeah, no, that's a good story. Right. But this concept of the new wanting to, you know, replace the old and this continuous evolution, you find it in nature. You know, everyone sort of knows the analogy of a forest fire burns the trees, so the new new brush can can grow like this, this, this concept renewals is always going to be there. It's in human nature, it's nature, I think it's just a pattern that we're going to, to see. And it's the same place here, like there's going to be a group of people, a group of firms, collection of individuals that are going to look what's happening and embrace that. And then there's going to be groups of people that are going to hold on to the past. And there's, there's going to be conflict to varying degrees because of that, that that hasn't changed a lot. So when it comes to investing, when we look for patterns, you know, quant, quant investing, of course, is built on this concept, that there are patterns in history, in stock prices and information that was predictive. And there's a structure to that data, even if that structure is so complex, that humans can't understand it. And in world class quant investing, you know, we've moved past what humans can understand, you know, many decades ago, people forget how long and what the book on Renaissance came out. But you know, some people still don't realize that the quant investing has been going full bore since at least the early 90s. And so absolutely, that whole class of strategies, and which is a huge part of the markets today, is built on this concept that historical patterns do repeat themselves. When most people think about investing, they think about sort of human driven intuition and investing. driven intuition in investing. I also believe that's timeless. It's just timeless on a different scale. And that is where on a go forward basis, I still very much believe that human investors will have an edge, particularly on low, law of large numbers situations where a machine hasn't exactly seen all the priors, the precursors that would lead it to make an informed decision, but a human can be better at dealing in the fuzzy mess. And so all these tools that we're building can be built to enable that human to make a prediction more accurately. Yeah, there's patterns that come with history. History doesn't repeat itself, it rhymes, like that's absolutely true for humans. So I guess summing all that up, just saying, oh, this is the way things happened in the past is silly because they never repeat themselves exactly. You have to sort of more go to the underlying mechanics, particularly around human nature. I think human nature is a constant, is stationary across time if you just look at it in the right way where you're thought, and why is all the other crap that's happening? Well, we'll have to debate whether human nature is static or not, because I'll take the opposite on that one, but I have another direction I wanna take this, which is, I didn't realize, now I'm nerding out, I didn't realize that there are things happening in quant trading that are in principle not explainable, not like humans can't understand it at all. Is that what you're saying? Or they could, but it's just so complicated that no one takes the time to do it because it's not worth it. Well, there's so many different flavors of quant investing. But yeah, generally speaking, the higher Sharpe ratio or the more consistent strategies that you get that aren't pure arbitrage, it's not just based on speed. It's just like, why does an LLM do what it's going to do? People kind of understand that, but not really. There are all the nonlinear relationships. A lot of those techniques have been used on structured data in these models for years. I mean, you go back and read the book about when the guys from IBM came over to Renaissance, what they were doing, they're doing speech recognition, like a lot of that sort of pattern matching and sort of predicting what comes next. Yeah, for a human to actually sit down and like walk me through all the different layers of the neural network and why did a machine do what it's going to do? No, the dimensionality of that problem, just it's way past what a human mind can understand. So I'd say generally speaking, world-class quant models, while the signals, what goes into them, I'm saying, this is something that I ultimately, this feature makes sense, how those features get, and which I do think is important. And again, there's various different ways in which firms go about this. Generally speaking, we want to understand at least some rationale to our features, even if there are many thousands of them. But how those ultimately get combined and understand the nonlinear relationships from that, that's very complex and that's totally fine, so. It's interesting that you brought up intuition a little bit earlier as something sort of separate from the more algorithmic and a helpful addition to the more algorithmic quant decision-making. Because I actually think of neural networks in any way, I think of neural networks and intuition, human intuition, as being analogous and maybe even helping us to understand how valuable and important intuition is despite being unexplainable. Because it's kind of unexplainable in the same way. You're working on a very high-dimensional basis with problems that you can kind of talk about why you make a decision here or there, but it's really just kind of a feeling that you've built up over many, many experiences. And I think neural networks are the same way. I do think intuition is very important for building a process. And really, another term that you could say is analogous to intuition is first principles. And there's a very famous guy that uses first principles all the time. But I very much believe that, is actually sort of understand, and it's really a math term, right? If you can understand some of the core concepts of first principles in math, then you can take any test and get 100%. So I was a math guy. I like taking from first principles. And I think it's the same way. But when you interface that with machines, what machines are really helping to do is say, I might actually help you come up with something intuitive, but you wouldn't necessarily come up with yourself. Kind of like having a coach where they could watch you doing a movement or lifting a weight. Like, hey, did you actually realize that these things are connected? So I don't think they're antagonistic, but it's entirely consistent that a machine can give you a very intuitive answer that you wouldn't necessarily have been able to think of yourself, if that makes sense. Do you think that first principles, and this is a leading question, because I have an opinion on this, but like, do you think that first principles in this, like, I don't think that first principles in the math sense are the same thing as first principles in the like, let's say, decision-making sense. In the sense that when you use the word first principles in a math sense, once you adopt them, you can just like work out all of the implications of those principles without, like you said, you can take the tests and you can just figure out what all the answers are. But first principles in a kind of a decision-making sense, like, they're not at the bare level of like, you know, axioms in math. They're already like many layers up above and can be filled in many different ways. The difference is in a mathematical model, and math is just a model for the way that things operate. And within a model, you have rules. And so there's real objectivity to what are the rules. Sometimes those rules can be very complicated, but there's an underlying structure in organic systems that sometimes there's structure, sometimes there isn't. So I agree with you when it comes to decision-making. That's why the word subjectivity exists. And what could be first principles to one person might be totally the opposite of someone else's first principles. They both call them ground truth and axioms, whatever it is. So there is an element of subjectivity in the real world. You know, when it comes to decision-making and what I believe and what I believe really good decision-makers try to do is in their own decision-making process, both to discover what's led to good decisions and what are areas and what's their decision-making has been subpar. So in some ways you can kind of apply it more to your own closed system, but there's a universal truth around how decisions are gonna get made. Like, yeah, that kind of falls down at some point because people might just totally be schooled in first principles that are so different that they would lead them to make completely different decisions. What are those for you? Like what are the, as you've learned to improve your decisions over time, what are the things that you've learned to take as good, reliable first principles and what are some of the ones that you've thrown out? At the top of the list would be the power of incentives when you're dealing with humans. And not in a negative sense, like everyone is greedy, although, you know, a lot of times that does drive behavior, but actually understanding the right incentive decision for why an individual or a group of individuals is doing something and trying to align that as much as possible. Whether you're running a company or making an investment, that's extremely powerful. It's just to say like, are the vectors, are the incentive vectors all pointing in the same dimension? So that's a big one. You know, you mentioned earlier, like I hate fluff, I hate, sometimes it gets me into trouble, but I just, I wish that people would be more that way, whether that's an axiom or rule, but whenever I sense that my antenna go up and you know, that's sort of a negative sign and really just trying to be intellectually honest and so not trying to reduce everything into a math problem. But a lot of times there is structure that you can pull out of a situation and it might just require hard work. So, you know, you can't manage what you can't measure. So try to measure a lot of things. And, you know, personally, I try to do that. I, you know, evaluate myself. I keep a journal, you know, every single day with AI now, by the way, which is great for efficient. Highly recommend everyone do that. You know, every single workout that I do for years, I've, I track, you know, I have a log and there's a lot of numbers in that just to see these trends, you know, evolve across time. Again, none of these are perfect, but I feel like people, because it's hard, a lot of people just don't do things like that. And I actually, on a go forward basis, I think that's one of the things that's so exciting about this world of really data and making meaning out of data is it's gonna be so much easier to do that. And a lot of it's gonna be around, are you collecting right data to help people with that? So yeah, you know, the biggest first principles I said, power of incentives and intellectual honesty. And those would be the top for me. What goes into your journal? Well, there's three components of my life, which would be, you know, my family. You know, I have three kids and I've been with my wife who's amazing for 15 years. And I love being with them. And it's not just a platitude. So my personal life, and now I'm interacting with the family and I feel a huge sense of responsibility to my family to provide them with a great life. So there's a section on that. There's a section on, and by the way, none of these are mutually exclusive. I don't believe in balance. I believe in harmony across different areas of life. But there's a section on that. There's a section on work, of course, which, you know, for me at this point is this great, great big challenge that's fun. I think sometimes people don't even use the word of fun. And I guess one of the things with AI, like this doesn't need to be scary. This is fun itself, if you look at it the right way. It's really important that people have that view. And you know, I'm working because I get meaning out of it. I enjoy it and it's different than, and I feel fortunate in that than most people where I can say like, I'm doing this because I certainly want to. So there's things on work that I talk about. And then there's things just on my personal health that I talk about, you know, how I'm feeling. Was it a good day or bad day? How'd my, you know, how'd my deadlifting session go in the morning? Stuff like that. You know, I'm a bit of a crazy person on that subject. So if I tried a new supplement or tried some new technique or something, I'll write about that. But I'm just trying to capture what's going on in my mind. I started doing this because, you know, I'd look back and in finance, right? We're dealing with time series. And so we could say like, oh, on this day, you made or lost money or something such happened. But I really wanted to remember like, what was I actually thinking on that day? And just trying to keep it all in your head, even for someone that can have a lot of, a lot of hard drive space is impossible. So getting that out of my head a bit more was why I started doing that. And it's been really helpful, both as an exercise in itself, even if I'm just writing it, but then going back and saying like, oh yeah, this is what I was thinking on that day. That's interesting. And help me understand that more, like you're maybe now, now you're like typing this into ChatGPT, you're speaking it, you're writing it into Google Doc and then putting it into ChatGPT. I'm trying to capture the concepts of what I felt on that day. So I can either speak or just write bullet points of, here's what's on my mind in these three categories. This is the date, this is what's on my mind. Each of the three categories, here we go. Some days that's a lot, some days that's a little. And then because I've been doing this and the machine knows my voice, that can be a one minute, even a 30 second exercise. It's just so easy to do. And I feel the reason most people don't journal is because historically, sit down and it would take time. But this is a perfect example of, you know the thought that's on your mind, just get it out there very easily and then it can be captured, processed in a way that is accessible once in the future. It's just this tiny little use case that's sort of so powerful with these tools. And there was no way that could have happened in the future, so. One word that you brought up a lot in the context of work, but also you just brought it up also in the context of family that I'm curious about is the word responsibility. What does that mean to you? I do think about this word a lot in the context of work and family, slightly different, but overlapping of course. You know, I feel that people that either are endowed with certain skills, like their clock speeds really fast, or they have a lot of resources, or they have, you know, great networks, just generally speaking people with, that the world has entrusted upon them skills or capabilities, have a responsibility to themselves and to their community to use that to the maximum extent possible. It's like, if you can run a hundred meters in under 10 seconds and you don't, it's a f***ing travesty. And, cause not everyone can do that. So historically I, and this is, you know, before I met my wife and years ago when it was really just me, I'd say that sense of responsibility was to myself to try to get the most out of my own capabilities. And then as I went on in my career and felt that I was able to do that, and I was like, okay, well, my kids are 10 and eight and two and a half, and we haven't even talked about what it's like to be a parent in a world of AI, but having a sense of responsibility to have them grow up and to flourish and have a relationship with them that can evolve as they become adults, but ultimately to set them on a path, you know, huge, huge sense of responsibility, you know, as a parent, it's just amazing how much kids look to you for guidance. And then responsibility for the firm, like in our world, in the investment world, you know, you have sort of two sets of responsibilities. You know, one is, and this is by far and away at the top and take this very seriously is people give you money or typically in our case, institutions give you money and because of our type of hedge fund, we're a pass-through structure, which is not an operational, you know, nuance that gets buried in a document that means that we have a blank check from our investors to spend money whenever we want. There's only, there's a very small number of firms that have that type of structure. Biggest ones would, or most famous ones is Citadel, but there's really only maybe a dozen, probably less than that that are real. So huge responsibility for our investors to do what's right, to not abuse that privilege and ultimately to, we're in the money-making business. Let's just be honest about that. But also responsibility to our people too. And I think this is what's generic across all leaders and all companies in the world of AI that leaders have a responsibility to the people working at their firm to prepare for what's coming next. And then I definitely feel that. And all of it's tied together. I have a responsibility to our investors to do the best job as I can for them and I have a responsibility to our people to be as efficient as possible. And those two coming together mesh very, very nicely. But yeah, I think for me, and I'm 40, so I'm not that old, do want to be doing this for a long time. But as I say, as you go higher up as far as responsibility, I think that the notion of being a little bit more of a steward and helping others accomplish what they want to accomplish, like that's something that successful people talk a lot about, and I very much feel that. And it's very much aligned with what we talked about today in the world of AI, so. I love it. Well, always a pleasure. Thank you so much for coming on. I'm excited to have you back maybe in a year when we have more results on how everything has been going. I always learn a lot from our chat, so thank you. All right, you bet, man. Thank you. 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 chat GPT. Every episode is a rollercoaster of emotions, insights, and laughter that will leave you on the edge of your seat. And if you're a fan of the show, you know that I'm a fan of the show. And now, without any further ado, let me just say, Dan, I'm absolutely, hopelessly in love with you.