Speaker 1the entire industry is focused on earnings. If you get the earnings right, you get the stock right. Maybe not. It's less of a predictor. Everyone's really good at earnings now in a way that they weren't when I started. A lot of the differentiating factor is this amorphous multiple. It is a little vibe of this is just going to work. This is the regime we're in. Structurally, that benefits the single manager because the individual idea becomes more important. It's hard to systematize multiple. You can systematize earnings. Multiple is harder. How do you systematize and create a process around multiple versus the individual knowing that's going to work and the multiple is going to expand?
Speaker 2I'm Ted Seides, and this is Capital Allocators. My guest on today's show is Paul Enright, a guru on equity hedge fund investing. Paul, who manages his family office, Kranos Capital, that he started after a long career at Viking Global. For the last three years, he went back into the business as chairman of fundamental equities for multi-manager pod shop, Jane Global. Our conversation is a deep dive on single manager and multi-manager hedge funds and on equity research in the age of AI. Capital Allocators is brought to you by AlphaSense. Every allocator sits at their desk for hours preparing the work around investment decisions, getting up to speed on a new theme, prepping for the next investment committee meeting, and tracking what changes across an entire portfolio. AlphaSense just launched something built to help. It's called Super Analyst, an always-on agent that works like a power user on your behalf. It draws on over 500 million premium sources, including broker research, expert transcripts, company filings, and earnings calls, and keeps working long after you've gone home for the day. Point Super Analyst at a theme, company, or market, and it learns the universe so you can easily turn new insights into the briefs, models, and memos you'd otherwise build by hand. It was designed to work like an extension of your team, so it remembers your projects, follows your templates and processes, and links every output to its source so you can see where each conclusion came from. The best way to judge its quality is to give it a whirl. Pick the research task on your desk before your next committee meeting and hand it to Super Analyst. Start a free trial at alpha-sense.com slash capital and see what it has ready for you. Capital Allocators is also brought to you by Ridgeline, the investment management platform built for firms who are rethinking how work gets done. You're at your desk looking at a line graph of your firm's growth. The line is moving sideways. Linear growth, and you know exactly why. You're constrained by your operational bandwidth that's been the same for so long, it just feels like the cost of doing business. Ridgeline is built to change that line's trajectory, to be the inflection point that takes your growth from incremental to accelerated with the team you already have in place. Ridgeline is the only AI native investment management platform that brings together the front, middle, and back office with one set of data and powerful AI intelligence that automate away operational drudgery, so your talent can shift from managing old technology to materializing new ideas. Ridgeline helps asset and wealth managers understand and improve their firm's growth trajectory. To see how your firm compares to 16,000 of your peers, visit ridgeline.ai slash trajectory. If you file Form ADV, your firm is already there. Then talk to Ridgeline about how you can do more with the team you already have. Capital Allocators is also brought to you by Juniper Square, the operations partner trusted by the leading private market GPs. One theme that keeps coming up on the show is GPs that build the strongest and most durable relationships with LPs are the ones that have both strong returns and strong operations. AI raises the stakes on operations, and having Juniper Square as your operations partner helps ensure your data is organized, accurate, and accessible so you can unlock operational alpha. That's why thousands of GPs trust Juniper Square. Learn more at junipersquare.com slash capital allocators. Please enjoy my conversation with Paul and Wright. Paul, great to be with you.
Speaker 1Hey, Ted. How are you?
Speaker 2Great. I think the best place to start is going to be talking about what you've learned the last couple of years to sitting inside a multi-manager pod shop.
Speaker 1There's a lot of things I thought I knew that were true, and there are some things that I thought I knew that were probably different. The analogy that I've used, if you think about a hedge fund generically as a factory, and you said, single manager factory, what is your chief unit of production? You would say, it's an exceptional idea. As a single manager, my job is to create and produce an exceptional idea. The factory as a whole hopefully can create and produce a portfolio of exceptional ideas. That's the chief unit of production. If you're a multi-manager, the idea is a component, but it's not the product. The product is the portfolio itself. The multi-manager factory is producing a portfolio of portfolios. Individual and idea is to a single manager as a portfolio is to the multi-manager. Your top 10 ideas at a single manager are going to drive the returns of the single manager. Your top 10 portfolios will drive the return of the multi-manager. Those are the equivalencies. The interesting thing then, let's talk about quality control. If we're talking about factories, the quality control layer in a single manager comes in the creation of the idea, the cultivating of the idea. It's more similar to a multi-manager than a multi-manager. Apple spends all this time, Legacy, Apple, Steve Jobs, Johnny Ive, we're making the iPhone. We spend so much time. They've tested it and retested it, and you've obsessed over it so that when it is delivered into the market, it's not being recalled. It feels like there's less risk. Sure, you're sending software updates. We have an S and an SE, or maybe it replaces it. We don't sell the iPhone 2 anymore. That's the upgrading of the idea. On the quality control side, on the multi-manager side, because the unit of production is the portfolio, individual ideas will not be as quality controlled ex ante. You have to be that much more aware and monitoring of how it operates once it's out in the market. If a portfolio is out there, you have more layers of risk management to say, hey, we might need to recall that portfolio. You bring it back, then you test it again. Maybe you put it back out in the market, or maybe you say, we're going to discontinue that line. We're not going to have that portfolio anymore. Let's replace it with another portfolio. That's how I've come to think about it. It's a crude analogy. There's a lot of merit to that framing of the two different types of businesses.
Speaker 2At the portfolio level, what does that imply about the ideas that go into the portfolio that becomes the unit of production for the
Speaker 1multi-manager shop? Can I torture this analogy a little bit more? Sure. Like if we switch from a factory analogy to more of a comedy analogy, you said a single manager is like a standup comedian. They're going to get up on stage. It's scary. They're there by themselves. Maybe they have a couple of analysts who are joke writers. Your job as a joke writer is to give me one great joke that I can stand up and I can stand behind. That's the equivalent of a single idea. A multi-manager portfolio manager is more like a showrunner. I need to produce a bunch of shows and you write the show. An analyst has to come together with a packaged show. It's not one joke in the show. How does it all fit together? You've got to tell more stories. I need a package of ideas that go together that will fit into this portfolio because one single idea isn't that important. Part of the reason why that is, is because multi-managers are more premised on managing risk under modern portfolio theory, where you want breadth, you want diversity. A single idea has much less relevance in the portfolio of a multi-manager than it does in the portfolio of a single manager. Great. You like this idea. I need six around the same theme. It's a package. It's a sleeve. Different people call it different things. You need a lot of those. You put those all together and that's how you get a portfolio, whereas on the single manager side, each individual idea comes together. The risk happens here because all the ideas might look and be the same. Then you have hugely high correlation, a lot of volatility, a lot of risk. You've thought about each one of them independent, that you don't realize that they're all the same thing. Over here, the risk is that you have five great ideas in there. You treat them like all the other ideas. It gets diluted down by everything else. This is less risky, but maybe you don't capture your best ideas all the time.
Speaker 2If we move away from the analogy and go to the day-to-day work, it feels like there's two different functional roles. There's the analyst coming up with the idea on either side, and there's the portfolio manager putting it together. How do each of those different at the multi-manager and the single manager shop?
Speaker 1If you had three units of time and you said, someone's got to do the research. Someone's got to do the idea generating and someone's got to do the portfolio management. At a single manager, an analyst does almost a hundred percent of the research. Maybe it's 90% and a portfolio manager does 10% or maybe a portfolio manager mirrors some percentage of it so that they feel like they're in the sweet spot. It's almost independent. You do a hundred percent of the digging. It's your job to cover this, own this, come up with something and say, I think I have something. That starts the conversation. The idea generation process depends on how senior the analyst is. It could be 60, 40, it could be 70, 30, but it's a shared process. At some point, your analyst is super independent and it could be a hundred percent idea driven and they're only pushing ideas up, but it's still a pretty shared process. The portfolio manager's job is to a hundred percent own the portfolio. Once we decide that's a good idea, I'll take it from there. I'll figure out how to do it, size it, how to get it in, how it fits with everything else. In the multi-manager research, idea generation, portfolio management, same three things, but the portfolio management piece becomes a little bit more shared because you're not just asking for an individual idea. You like that idea. What are you going to pair it with? The analyst job, I would argue at a multi-manager is a lot harder because a lot more is put on them. They have to do the research. They have to come up with ideas and then they need to package it together to help the portfolio manager. Some amount of portfolio management is pushed down to the analyst. In a multi-management layer, that is not done in the same way at a single manager.
Speaker 2So the extent that the single manager PM is focused mostly on the portfolio construction, you'd think that maybe they have time to say, there are these individual ideas, how can we package them together so it delivers a robust portfolio?
Speaker 1The single managers that do it best are focused on exceptional ideas and exceptional portfolio construction. The enemy of exceptional ideas is mediocre portfolio construction or solid initial good portfolio construction, but not as sophisticated risk management around the portfolio construction as you face volatility.
Speaker 2As you stepped into this, you said there were some things you learned along the way. What were the most surprising aspects of being inside the multi-manager shop that you didn't appreciate from being outside?
Speaker 1The execution of modern portfolio management theory and the square of the sum of the squares math, the diversification benefit. I think that there are a lot of things that you learn along the way. I understood that conceptually. A lot of people will talk about factor neutral, beta neutral, all this. Breath, having a lot of ideas, a lot of diverse ideas is incredibly important. It's a batting average business. Single managers, more of a slugging percentage, best ideas for most. I had no idea. Totally clueless. I thought I'd walk in like, oh, this is the magic model. It's the Wizard of Oz behind the curtain. You get some sense of this is how idiosyncratic risk is. No, it's just a model trying to backtest and predict things. It's the emphasis on breath because volatility and sharp drive returns a multi-manager. At a single manager, what you care about is long short spread, how you apply your gross and net to that potential long short spread. That was surprising to me. I've never been a high velocity of ideas person. I'm interested in the things I'm interested in. If I'm not interested, my eyes will gloss over. When you hear pitch after pitch after pitch, and you're like, that's fine. That's fine. And then you're like, that was awesome. That's exciting as a single manager. When breath is your goal, you need all the ideas. It's hard to get excited about all the ideas. You're like, yeah, that's fine. I always had the luxury of being like, that's not great. I don't want to do that. I'd rather be in cash. And it's like, oh, that's a great one. Let's do that. I'm wired that way. I'll always be wired that way.
Speaker 2As the pod shops have scaled, clearly, the breadth benefits from scale. What are some of the other ways that scale has accrued to the benefit of the multi-manager shops?
Speaker 1The thing that we can say definitively about the hedge fund industry when I broke into it was there was more talent than capital. If you wanted to be an investor and you were talented, Tiger was the only game in town in the late 90s. The idea that there is all this density of talent there in this industry that had not yet scaled makes sense. Then, like any diffusion into the economy, 10 of those great investors go and do 10 different things. They hire 10 people. So now you've got 100. It makes sense that the talent would get diffused. One of the great things at Viking was we had great talent density there and we were always competing for capital at various points. And then you get bigger during the early years. Wow, we have way more talent than we have capital. And then after a while, you felt like you were in equilibrium. As the multi-managers have scaled, the benefit is they have more capital. They have more capital. Then they have talent. Therefore, they become the magnet for the talent. That is a huge benefit. The second thing is that when you have scale, you can create a center book behind it. The opportunity to have breadth of ideas, to see all the different pockets of ideas and run a center book on top of that, that's an amplifier. It's almost pure margin on top of that. The other way that there's real scale is in the data layer. Having the access to be able to acquire and try and error as much data as possible and have teams that can analyze that data so that it's not necessarily put onto the investor or the analysts themselves. So much of the challenge with data during my career, whether as an investor and looking at companies that extract data or transform data or anything along that continuum, most of the time was spent getting the data into the usable format, putting it in the right place. Very little time is spent analyzing the data. There's real scale. And there's real scale to being able to throw a team at that. We're going to get the data, scrub it, figure out what structure, what's unstructured. We're going to put it into place. Then we're going to deliver it to you in a form that's usable. When you layer AI on top of that, you have better access to capital. That's a benefit. All of those are huge benefits of scale.
Speaker 2So much of the concept of alpha generation in equities is finding the better idea or a portfolio of ideas that's better It's often hard to get your head around that more breath, more ideas somehow can be better than the market than a smaller group of ideas. When you saw it in practice, how did you wrestle with that? Somehow there's still a unit of alpha that's being monetized, risk managed, supersized through leverage.
Speaker 1This is probably another area that I have better visibility and better understanding of now than I would have had five years ago. A unit of alpha has a volatility component and a risk component to it. You might have a better, more exceptional idea that has a different volatility profile and different risk profile. Therefore, you don't want to apply a lot of leverage to it. On the other side, you might have something that has less alpha content to it, but it has a type of alpha content that, when it's packaged with other ideas, reduces the amount of risk. The way you get your risk is you add the leverage. On an unlevered basis, Apple's. What you would find is that those returns would not nearly be as good as that exceptional return because you've shrunk what you would consider as less measurable risk to a more measurable size, a more measurable unit of production. Then we'll lever the hell out of it because we'll take leverage as the form of risk. You can't have concentration, volatility, and leverage together. You're dead. Those are all strange bedfellows. You've got to pick one or two of them. If you're levered, concentrated, and volatile, I guarantee you, you're going to have.
Speaker 2There have been a few problems.
Speaker 1You're going to have some problems along the way.
Speaker 2I'd love your perspective on what this means at the industry level.
Speaker 1Big picture. Take a step back. If you think about 25 years of the attribution of stock returns, put aside dividend and dividend reinvestment. You say there's two components. One is earnings growth. One is multiple expansion or contraction. That is wildly inconsistent over time. People might give you averages. But if you looked on a year-by-year basis, I broke into the industry in 98. So I've looked back to 98 through 2025. You went through periods, especially early on, where most of the attribution to stock performance was earnings growth. It makes sense to me that the industry would gravitate towards, how do we predict earnings? If you can predict earnings, then you can predict the stock. If you look at 2011 through 2025, we switched. There's a couple of weird years in it. Most of it is two-thirds of returns in any given year on a pretty consistent basis for multiple expansion. What is the system for hedge funds, for multi-managers, single-managers, whoever, for stock pickers, for trying to understand what leads to multiple expansion or multiple contraction over periods of time? maybe not. Structurally, you have the entire world now focused on earnings, earnings, earnings. We're getting to the end of 2026 and the debate could be, I'm focused on 27 earnings and someone say, you're a chump. 27? 28. If you just look to 28, close my eyes, I'm going to make all this money. Now it's gotten a little bit more quarter by quarter. There's no debating and meta is a perfect example. Everything that's happened to meta in the last, two or three months, as this document really volatile, has been multiple. Muse is this amazing thing. Wow. Multiple goes up. I guess they're going to raise money and Muse isn't monetized. Multiple comes down. Bobby and I have talked about this. Bobby's always funny with the way he'll distill things. Bobby will be like, so it's just vibe? Multiple is just vibe? And I'm like, it's not vibe. There are reasons and you can study it. Historically, it changes. If we did a study, you would say how earnings move determines what the multiple will do. If you get a revenue beat, it usually leads to a more sustainable multiple expansion. Revenue growth seems more durable. If you get a revenue beat with gross margin expansion, that's a better combination because revenue growth with flat to declining margins, there's not quite as good. There are reasons why multiples expand. More important than that, a huge contributor to performance has been the derating ahead of some big secular fear. A big component of this is interest rates. Multiples dramatically went up because people started to say, what if the terminal is not five and it's four or three, now two, and now all of a sudden it's like, well, what if the terminal is not two, but three, but four, but five? How much of it is that? That's a fair question. Industry structure, it doesn't matter whether it's single manager, multi-manager, everyone's focused on earnings. That benefits the single manager because the individual idea becomes more important. We know what our models are doing. We know where our earnings are. We know we are versus consensus this quarter, next quarter. As the updates come in and consensus starts moving, we can track it. Multiple is harder.
Speaker 2How have you thought about or seen people trying to play with the behavioral psychology part of markets to get to a point where it's not just a one-off guess?
Speaker 1It's part psychology and it's part being comfortable saying this isn't just about stock picking. There's this identity in being like, I don't do macro. I'm a stock picker. We care. We ignore it. The stock picker vibe is I'm focused on this is a good business. It's going to earn $10 in five years. It'll trade. 20 times. The 20 times, who cares? 20, 22, 23, 2018, whatever. You just make up a multiple because the earnings are what matters most. Brad Jacobs has this great quote. You get all the little things right and get the big thing wrong. It's not going to matter. That's true in earnings too where you get all the little data points right. But if you are picking what you think is great long in a sector that's not going to get the multiple because of where we are in a cycle or in a regime, there's this overhang, that might be more important than what you think. I have this image in my mind of every sector is always flashing yellow, green, or red. If you are a traditional long, short, neutral fund, more in the multi-manager camp, you might always want to think that everything's always yellow. You can always find equal longs, equal short. In practice, everybody knows that's not true. Sectors are always flashing green or red most of the time. If you're picking a stock, in a flashing green sector, you're probably going to get multiple expansion way better than you think for some period of time. But you have to be aware that's at least influenced by the backdrop. If the macro backdrop is pro-cyclical, you're going to get even more multiple expansion. If interest rates are declining versus rising, you're probably going to get even more multiple expansion. There is some psychology to it. There are other things outside of the stock that contribute to multiple, which makes it more of a breadth of a skill set versus that earnings piece of it.
Speaker 2When you pull away from these components that are increasingly complex in trying to get a stock right in this period of time where there's a lot of money in pod shops, there's also a lot of money in index funds.
Speaker 1A lot of money in quants and CTAs too.
Speaker 2How do you think about the potential for whether it's a single manager or multi-manager to have enough of an edge that they can beat the market over time going forward?
Speaker 1I used to think that edge was, I'm going to see something nobody else has seen. I'm going to therefore make an investment and eventually everybody else will see it. That was more of the opportunity set in the past. Now that everything has picked over, the idea that you have some unique insight edge comes from having a deep crystallized knowledge that you possess, not that you're borrowing from an LLM and not that you're borrowing from your analyst. If you have a deep, long history with a series of stocks and you have this crystallized knowledge, it's when that thing that you saw years ago, other people think they saw, gets tested. That is where more alpha comes from. Danaher has been a terrible stock for five years. This summer, Danaher bottoms. Should you be going in and doing work on Danaher and saying, now's the time, this is the opportunity, this is the hedge fund darling for 10 years. It's been one of the worst performing stocks post-COVID. Your edge comes from being able to look in those and crisis inside of a business and say, this is a blip. They're going to figure it out. GE Aerospace bought a component company a few weeks ago and how it has gotten crushed. Multiples gone from 30 to 20, rolling up your sleeves, doing work there about disruption and inflections. There's real edge there, real controversy. Maybe the edge comes from zeroing in on what is most controversial. Can you be first? If you're a subject matter expert, you have a PhD in semiconductors from an investing perspective, you're first. Memory is the bottleneck. Yes, those opportunities present themselves, but the more readily fertilizable or available opportunity are those battleground stocks. If you can have a view on a battleground stock where the multiple has dramatically moved ahead of the numbers, all you're betting is the numbers aren't going to come down as much or the numbers aren't going to come down at all and we're going to get the multiple back. That is the best source of edge for idea generation in the market right now.
Speaker 2And partnering alongside clients to shape tomorrow's economy today. Invest in the businesses that shape tomorrow. Learn how you can own what's next at brookfield.com slash own. This is not an offer to sell or investment advice. Investing involves risk, including the loss of capital. And now back to the show. That concept of idea generation requires a fairly long tenure of understanding. Let's, then, say, if you were an analyst at a fund who is developing that knowledge over time, how does AI play into what the functional role of the analyst becomes, what the functional role of the PM is relative to the analyst?
Speaker 1This is such a great question. It's my favorite topic in the whole world right now. It applies to our kids' education, to any industry where you have to develop some sort of knowledge. There are multiple different types of knowledge. There are multiple different types of research that you should be doing. One is, I should be getting familiar with a company. And I should know the company and its products and everything it does. And then I should understand the sector that the company operates in, any adjacencies, maybe their suppliers. There's these Venn diagrams. There's these rings. Separate from that, there's the stock. I need to understand the stock. How does the stock work? Where does it trade? And is it a U.S. listed stock? What's the market cap? What's the liquidity? How much does it trade per day? Increasingly, you want to understand, how does it screen from a factor perspective? Is this a value stock? Whose radar is it on? I don't think you want to be predicting the macro, but you want to be aware of it. There's work that you need to do to understand the outside view. You're developing your inside view, but there's this outside view that's coming at you all the time. And how do you track that? So all these things I would put in the category of research. The thing that AI can do exceptionally well is all of it in some ways. If you want to be able to survive coexist and leverage AI to the best, my suggestion would be take the things that have the longest half-life, that have the greatest ability to get embedded in your long-term working memory, to be crystallized. I know things about charter communications, Disney, Qualcomm, Intel. I will never forget. The way that people should be spending their time, don't take a shortcut and ask AI to do what they want to do. If you want to be able to do what you want to do, take a shortcut and ask AI to do what you want to do. You should be reading every single document you can get your hands on. You should be buying books off of Amazon and learning the history. You should be doing all of that work at your desk. And then like a CIA agent, as my favorite analogy, then you should get out and get into the field. You should be having conversations, walking floors, testing products, going through all of this stuff that gets embedded. All of that has a pretty long half-life because it's company-related, it's product-related, it's management-related, it's industry structure. it takes a long time to change. You should do the work and feel comfortable that it doesn't have a short half-life. The stuff that you want to automate and that you want to give to AI is the stuff that has a pretty short half-life, where it's a complete waste of your time. Why are you building a spreadsheet that shows what consensus earnings are doing? They're moving all the time. Automate that. Figure out a tool that gets that one that you don't even have to fetch. It's sending it to you. You want to create a macro dashboard, which is you've decided, I care about what is the consensus dot plot saying that rates are going to do. What does my dashboard look like that I can glance at it and be like, huh, the market's pretty bearish. I'm not doing anything. It's got a half life. I'm not doing any work. I'm just consuming. What do I want to consume? Because there's no real benefit for me to create. On the other side, what am I embedding in my long-term knowledge that I can do the work, avoiding cognitively, cognitive debt is super important right now. If you go and say, oh, I did a screen on AI. I have Claude delivering this summary. Claude wrote the initiation memo for me on XYZ company. I guarantee you, you won't embed it in your memory and remember it the same way if you did it yourself. Claude can one-shot a model in a few minutes now. You should decide exactly how you want the model to build up with the KPIs, et cetera. Don't let AI decide what the inputs are. Don't understand the inputs. You won't understand the reflexivity between the inputs and the outputs if you haven't done it yourself. Once it's up and running, have it updated for you on earnings in two seconds. Don't short-circuit that initial point. The process of launching coverage, do it yourself. The process of maintaining coverage, get the help.
Speaker 2What will that look like within an investment firm that's using it effectively from a staffing perspective compared to what it's
Speaker 1been? In the past, to scale your self, you would always throw more people at problems. Now, you'll throw a combination of people in compute. What is it that humans should uniquely do? Part of the risk, too, is you hire a junior, you give them a lot of short half-life tasks, and they're not getting better either. You're treating this human being as a commodity. They'll ultimately leave because they'll be like, I'm not getting the quality work. Define the skills or the tasks that you need to be done. Give to the humans, the ones that have the longest half-life, the most opportunity to develop them. You're creating this crystallized knowledge inside these human beings and take all the other tasks and automate them. That's just the research side. Then you can point AI at portfolio optimization in a way that would be better than human beings, too. My second favorite topic is the word duration. I'll give you three different definitions of duration. Some people mean, I want duration. I want the permission from my LPs to not be actively managing this portfolio because I'm not really good at actively managing it. I want the permission from my LPs to not be actively managing it. I want the you're going to be happy that I absorbed all this volatility. That usually means one or two stocks that I love that I don't really know how to size. They have a disproportionate impact on my portfolio. From a multi-manager perspective, when increasingly you're hearing a couple of the big multi-managers talk about, maybe we should have a little bit more duration. What they're saying is we need to build portfolios that have greater duration. Not we accept more volatility. We need to build the portfolios so that they will endure through this volatility. How do we get greater breadth? We need to construct the portfolio in a way so that it's not being subject to the same volatility as the portfolios are now. How can we better optimize the breadth, the risk profile? That's where AI can help there. My view of duration is I want to have long duration on my view of the company. I want to have the appropriate duration. I think about them completely differently. My view on companies does not change as much as my view on the stock changes. The portfolio should reflect the duration of my view on the stock. Don't get those two confused. On the portfolio optimization side, if you have a concentrated portfolio or you have a wide breadth portfolio, optimizing the portfolio is something that needs to be done probably every day. Some people do it, never. Some people do it once a week. Some people tweak a little bit here and there. AI will be better at helping you to optimize your portfolio. If you get to a place where man plus machine equals, we're generating these ideas, we have these little packages, I need some help putting this together. That will create some edge around. We get optimal portfolio construction because we have this mapping technique that we're going to need to do. We're going to need a little bit bigger here, a little bigger there. You might even get to the point where you just press a button, it does it for you. It's hard as a human being to be that on all the time, that much kinetic energy of constantly optimizing for every move. That's probably where some source of edge will come over time. It won't be like quant where you're completely doing it. There's some mixture between the two where you are generating the ideas. These are your ideas. You're looking at it and you're thinking, I don't know, it's up 20%. It's now an 11%. You press the button. Based on everything that's happened, you probably should take 1% out and redeploy that 1% over here.
Speaker 2That optimization exercise, if the output is adjusted position sizing, what is the input that the machine or a quant model is looking at to say, based on this movement or that movement, therefore you
Speaker 1should take it down by 1%? The input you're assessment of earnings on a one-quarter basis, two-quarter basis, one-year basis, two-year basis, your multiple, it's difficult to dynamically adjust your multiple as time goes on. We tend to think about multiple like, I'm looking at 2027 multiple or 2028 multiple. You talk about that in August in the same way you talk about it in October, but two months have passed. Real-time thinking about how multiples relative to growth shift on a daily, weekly, monthly, quarterly basis, because this is a rolling nine-month thing. Weird arbitrage opportunities happen where you get back from Labor Day, everybody's focused on the end of the year. You think, I don't care. I've positioned my portfolio. I came into the summer. I'm fine. I'm looking at 27, and I guarantee you I don't absorb any volatility looking to 27 and next year. You make a bunch of money because you've shifted your perspective. Having a tool that helps you do that on this real-time basis on the multiples can be helpful. The second thing is the real-time updating. You tend to do point-in-time updates to consensus. Real-time updates to consensus as well would be another thing that could be automated. You can be behind a little bit.
Speaker 2So much of this evolution, that fine-tuning of the information you have that goes into a decision, that goes into a portfolio, is a whole level of science order of magnitude different from, I have a great idea, so I want to own a lot of it. I'd love to get your perspective on concentration.
Speaker 1The Kelly criterion and its application in sizing is interesting to look at if you're talking about a long-only, unlevered pool of capital. That's easy to do. You could look at your upsides, look at the. Rolling of the dice and figuring out, based on the number of ideas I have, the relative upside of those ideas, we could figure out how to size them and how concentrated we should be. Then you would be able to compete those against the marginal idea and say, no, why? When you put in leverage, then volatility, then maybe shorting, and where it's not the relative sizing of the idea to the capital, but the long ideas to the short ideas in your long-short ratio, it becomes a lot more complicated. There are certain things that I've learned instinctually over time that are red flags as volatility increases relative to position sizing, your real-time monitoring of your future expected long short spread relative to how much gross exposure you have. If you could train a model on all of the things you care about, have that model monitor what goes on in the market and what goes on in your portfolio, can you optimize around those things so you can say, you're wrong, you shouldn't be doing this, you're wrong, you shouldn't be doing this, you're wrong, you shouldn't be doing this, you're wrong, you shouldn't be doing this, you're wrong, you're wrong, you're wrong, you're wrong. Or 12% is the optimal size relative to 200 gross. If you're going to take your gross from 200 down to 160, then your optimal size is 8%. So much of position sizing happens passively. Stock goes up, you leave it, or our capital is 8 billion now versus 5 billion. And because it's 8 billion, it doesn't feel good to run with as much gross exposure as I used to have. If you are going to use a lot of leverage, concentration of volatility will get you in trouble. Where exactly those limits are, I can't tell you precisely. Because I'm stylistically biased to be concentrated, and I'm comfortable with volatility, then I'm just not going to use a lot of leverage. Part of my challenge in ever thinking about what is the right institutional product for me, that's maybe the hardest to raise capital around, concentrated volatile. It's a nice way for me to live. Not everybody wants to live that way. It's one of the things that I feed into Claude and I play with.
Speaker 2You have all these different cross currents about the way hedge funds have changed and the data and AI. If you were designing a fund of the future, from the analyst to the portfolio manager, what is a winning formula? look like going forward?
Speaker 1It's definitely man plus machine. Fewer people per dollar of capital you expect to deploy than maybe would have been the case a year ago or two or three years ago. But not so little that you can't do all of those long half-life tasks with a human being, and therefore you're relying on the AI to do some of the long half-life tasks for you. If the benchmark for a while was how many stocks can you cover and be productive, is that number change? That would be the starting point. We want to cover 200 stocks. The instinct now is for someone to say, we could do it with three people. We'll have a PM, two people, and each will cover 100 stocks. That's single managers might want to rethink their model a little bit in the sense that that's super doable for somebody. I think that's a good point. A lot of history and a lot of crystallized knowledge. As you get older, you have greater crystallized intelligence. You have less fluid intelligence. The fluid intelligence energizes you. It gives you ideas, and it's this kinetic energy, but it's dumb. There's not a lot of wisdom. Wisdom meets energy, and you go out into the world. It's a good combination. One of the reasons why I wanted to go back and do something institutional is I wanted to be around highly fluidly intelligent people who are going to do all this work. That's a great idea. I get that. You could do that now with AI. If you have a ton of crystallized knowledge, you could get fluid intelligence from the model because the model's monitoring things. Then maybe you only need one junior analyst. Maybe it's PM to senior people, to junior people who you're training, you're passing down real knowledge to, and they'll eventually replace you. They provide some fluid intelligence, and they're out matching with the energy. Energy out doing a bunch of different things, and then you're complementing that. Where in the past, it might be one senior one and three juniors. That's a real possibility. How you allocate the time and the tasks between throwing compute at it versus throwing the person at it will determine how that works. That structure will change over time. How people think about, you just do 40 stocks, go to the corner and don't talk to anybody. It becomes a little bit more of a continuum of more collective conversation over time. That's how I would consider approaching it today versus what a high-functioning team would have looked like a few years ago.
Speaker 2As you've stepped back again and are now overseeing your family office, how are you thinking about deploying your own capital into what is today?
Speaker 1Part of the reason why I wanted to step back was being inside an SEC-compliant entity, seeing all of these amazing ideas, knowing I couldn't act on them. I felt I had withdrawal. You shut off. You no longer have access to that information. I'm completely free to do whatever I want. I've got to rebuild it all. What's super interesting is playing with the models and trying to figure out, what do I know? Where would I start and pair it with there? I haven't hired anybody. What I'm doing now is probably not the right answer long-term. What's the product out there that doesn't exist that I would invest in? Is more of this public-private hybrid fund where you are looking, for the best ideas? It doesn't matter what sector it's in, what it's exposed to. You're truly a generalist and you're comparing these public ideas to the private ideas that are close to becoming public. What I've learned about myself and what I find interesting is there's a continuum of things that you can spend your time, energy, and money investing in. People, not me. That's not my thing. I can't just look at a person and be like, me and the Elon Musk business. God bless the venture capitalists that do that. That's amazing. Sometimes it's not just the person. It's the person and they've got a product or an idea, but it's still not a business. All I know is I would like to spend time with people that do that stuff because I'm in awe of people that do that stuff. I would like to invest in that stuff when they become businesses. Once it becomes a business, I'm interested. I love analyzing the people, the product, the business. I love analyzing the stock and the market around in the economy. That's what I want to continue to do. The best returns are going to come from being able to do that and pairing it with the stuff before it becomes public. I'm focused on that. The most interesting innovative product that's been put into the market by public market investors is what Philippe did with CTEK. Brilliant. I would think that there'll be more of those, not just targeted at high net worth individuals or retail individuals, but could be an institutional product like that as well.
Speaker 2What about that product compared to other institutional versions of service strategy made it special?
Speaker 1A lot will depend on the execution and what ideas go in there, but structurally, being able to make one investment and know you're getting the best of both worlds. It could be long only, or it could be a relatively high net long short fund. There's both an informational opportunity as well as a performance opportunity to have both the privates and the publics in one vehicle. You can generate the return in one place. If you're comfortable making multiple bets, stylistically, I like being concentrated. One way to spend all of your time. If you had access to information from OpenAI, Ramp, Databricks, and Anthropic over the last year, I guarantee you would have done better in your public. Now that's cyclical. That won't always happen. But if you look at the Facebook IPO was a bit of a crossroads for the public-private market. Anything that IPO'd before Facebook went public somewhere between year four or five of their existence. There was so much value creation in the public markets. Now companies come public when they're 10, 11, 12, 13, 14 years old. A lot of the value creation is in private. If you're sitting there as a public market investor, you're interested in force ranking things the way I am. I'm always looking at it and saying, I don't want to own this. I want to own that. I'm not in the and business. It's not, I want to own this and I want to own this and I want to own this. I want to own the best things when I think they're the best things. What's attractive to me about this is you could say, we're comparing this to this. We're putting this together. We're putting this together in one portfolio. And now if it becomes a portfolio where you own 150 things, it doesn't accomplish it. But if you manage it in a way where you're like, we have this concentrated private book, we're always looking for the best ideas that are going to become the best public ideas because they have a path there. We're not doing early stage when they're still a product and we don't even know if it's a business. We're investing in real businesses. It just happened to be private. That to me is super interesting and blurs the line between public and private in one product. It's not a breath. It doesn't take a big VC. We put in 30 ideas. It would be relatively concentrated. You have a handful to 10 ideas in the private side and you have a dozen or so public ideas. It's a best ideas concept, but a public-private best ideas concept.
Speaker 2Have you been spending your time since you left Jane?
Speaker 1I've been spending my time with AI. I still spend some time with Jane. I'm helping and advising them there. I've spent a lot more time reconnecting with my models, using AI to update some of my models that haven't been updated in four years, reconnecting with a network of people who I can talk to in an SEC complaint. I have a spreadsheet that has three tabs in it. One tab is the evergreen portfolio. It's 150 stocks, 150 of the best businesses in the world. I'm constantly force ranking them and looking at them and saying, these would be the 10 that I would own at any given point in time. It's always fun to look at them. There's a second tab, which is cyclical companies. These are ones where supply-demand dynamics are always going crazy. You're always you would not say they're good business. You're not saying they're bad business. They're cyclical. There's a third one where anything that was in column one that died goes to spreadsheet number three. In addition to that, there's a bunch of stuff that was never on one. If you were going to construct and take the 10 best from tab one, a nice blend of long short from tab two, and always be short what you think are the 10 worst here, that would be a good portfolio. Refreshing those stocks and getting comfortable with that, trying to figure out what do I miss? What do I want to do? That's how I spend my time. Do I want to do that on my own? I don't want to have constant exposure. I want to have opportunistic exposure. I want to get excited. Would I be open to doing something else? I don't know. That's where my head's at.
Speaker 2How are you thinking about what's next?
Speaker 1Doing the work and being a little bit more open to people suggesting things to me for now before outwardly going out and trying to see if other things would be interesting to me.
Speaker 2All right, Paul, I got a couple of fun ones at the end. What have you changed your mind about the last few years?
Speaker 1I changed my mind a lot. I've probably sent multiple dozen emails in my life that I thought were brilliant. And I was going to wow someone with my genius. They weren't always received that way. And now being the recipient, a lot of those emails over a long period of time, my advice, write the email, write it for your own benefit. It's a form of therapy or it's a form of thinking. And call someone and talk to them. Don't send it. No one wants to read that email. You want to write it, but no one wants to read it. I've changed my mind on that. And I wish I could recall some emails I've sent over my 25 years.
Speaker 2All right, one last one for you, which is a version of what we talked about, but a little bit different. If the next five years are a chapter in your life broadly, what's that chapter about?
Speaker 1I would like the next five years of my life to be a healthy balance of continuing growing to grow and learn. share as well. We're going to have an epidemic in this country of science keeping us alive till like we're 90, 100, 110 years old. All of our skeletal systems are going to fall apart. I have a window in my next chapter to like make sure my skeleton doesn't fall apart. That's incredibly important. We have a unique window in these five years for myself, for our kids, for future generations to make sure we don't go too far and lean too much into it and allow too much cognitive debt to accumulate for kids or ourselves and what we feed into AI. At the same time, we also risk not being aware of and leaning into it enough so that a whole technological cycle of productivity and a tool that can help us. How do you get that dialed in right? It's something I'm focused on personally, professionally as a parent.
Speaker 2Paul, always fun. Always mind-blowingly insightful. A lot of random analogies.
Speaker 1Until next time.
Speaker 2All right, good to chat. Thanks for listening all the way to the end. If you made it this far, how about one ask from me? Tell one friend about the show. It's the best way to keep growing this incredible community. Thanks so much. And until the next one, stay curious and keep compounding knowledge, relationships, and capital a little bit at a time.
Speaker 3All opinions expressed by TED and podcast guests are solely their own opinions and do not reflect the opinion of capital allocators or their firms. This podcast is for informational purposes only and should not be relied upon as a basis for investment decisions. Clients of capital allocators or podcast guests may maintain positions and securities discussed on this podcast. I'll see you next time.