与 Caro-Kann 的 Artem Fokin 一起精进投资技艺

You're about to listen to the Yet Another Value Podcast with your host, Andrew Walker. Today is a kind of special episode. I have my friend Artem Fokin back on the podcast. Artem is one of the most frequent and popular guests on the podcast. He's come on seven or eight times, and we do something different. We don't talk about an individual specific stock. This is just Artem and I coming on and kind of rambling for about an hour, over an hour, about process improvements, things we're thinking about, and ways we've tried to improve as investors over the past 10 years. We start touching on AI and expert calls a little bit. This podcast is going to be sponsored by AlphaSense. We are doing a separate webinar talking specifically about process improvements, AI, and expert calls, and I'll include a link in the show notes. I think investing is a mental sport. To Artem and me, I know it's a craft. We're always thinking about how to improve. So I think you're going to really enjoy this conversation about ways we can improve and ways we think about investing. We're going to get there in one second, but first a word from our sponsors.

Today's podcast is sponsored by AlphaSense. AlphaSense and Tegus are two of my longest-time subscriptions and two of the podcast's longest-time sponsorships. I love them both, and I'm so glad they merged. The product is awesome. I consider it definitely the most valuable subscription I've got between AlphaSense's AI tools and the expert library. I'm always pushing myself to be a better investor, and one way I'm trying to do that is by doing an expert call once a week on a company or sector that I'm researching, come rain or shine. It's a really interesting way to tap into new ideas and talk to people who are actually operating in an industry. I do that myself out of pocket. AlphaSense doesn't pay. Tegus doesn't pay. That's just me. But I mention it because I think it's continued to help improve me as an investor. If you're a fundamental investor interested in learning more and diving deeper, I think you will, too. So, Tegus and AlphaSense, I love the product. They've been a longtime sponsor, and I'm happy to keep having them on the podcast.

Hello and welcome to the Yet Another Value Podcast. I'm your host, Andrew Walker. With me today, I'm excited to have one of my best friends in the industry, one of my favorite guests, Artem Fokin. Artem, how’s it going?

Hi, Andrew. Great seeing you. I was actually thinking about wearing the same Yet Another Value Podcast polo today for the recording, but I decided that maybe it would be one too many. I figured you would be exclusively wearing this.

Well, that is it because I told you last night I was going to wear this, and then you said I might wear my famous pink polo to go with it. You often wear that for recordings.

Sometimes.

But we’ve got a lot to talk about today. Let’s just start before we get there. Quick disclaimer: Nothing on this podcast is investing advice. Full disclaimer always at the end.

Artem and I are doing a different podcast here. Artem is one of the most popular guests. I believe it’s his 7th or 8th time on the podcast. We did a podcast earlier this year that got a lot of great feedback, where I interviewed him as one of the keynotes at Planet MicroCap.

This podcast, we’re just taking a step back—no real individual securities—and talking about the investing process, things we learn about, and things that Artem and I talk about while I walk my dog Penny 2 or 3 nights a week, every week. That’s the overall feel of this process. Anything you want to add before I jump into a first question and we just start rambling for an hour—or, knowing you, 3 hours?

Yeah, knowing me and you, it will not be an hour. It will probably run longer. I think if I were to summarize the theme, the topic, the subject of today’s podcast, I would call it “Perfecting the craft.”

Yes.

Investing is a craft.

Yes.

It’s a profession. It’s a craft. We, as craftsmen, would like to improve. You reach that improvement through iteration, through feedback loops, through talking with other people in the field and getting their feedback on your own process—on your own process of conducting that craft—and sometimes cloning or stealing their ideas.

I’m not talking about stock ideas. I’m talking about process ideas: how to play this game at a better level.

Art, it’s like you’re in my head. I’ve got a post—it’s in draft, and I’m going to put it up at some point—but if I can ramble for a second, the post is all about this: I recently ran a HYROX race, right?

When you run any type of race, it’s really clear who’s better than you, right? If they cross the finish line faster than you, they’re better. One of the interesting things in all sports is that when you train hard and run a race, and then you see someone run it faster, psychologically, you know that barrier can be broken and you can push yourself harder and beat it. The most famous example is Roger Bannister, right? He runs the 4-minute mile, and people think it’s physically impossible. Two months later, someone else runs the 4-minute mile. Now good high schoolers will run 4-minute miles.

I always think about that in investing. What is the thing that pushes you harder? I’m sure all of us are internally driven, competitive people. What is the thing that pushes you harder?

I’ll tell you what it’s not. It is not, “Investor A generated 300% returns last year, and I generated 4% returns. They’re so much better than me. I need to push harder,” because that has no context. They might have just YOLO’d call options.

For me, what it often is—if I can try to land this plane—is when I talk to people like you and I say, “Hey, Artem, walk me through this thesis,” and you say, “Oh, yeah. As part of this thesis, I did 5 expert calls on Tegus, and one of them uncovered this nugget that no one knew, and it sent me down this rabbit hole.”

It’s when I hear someone who did great research, and I’m like, “Oh, I wish I had that insight into an idea I had,” or, “I wish I had that in the ideas I had.” That’s what pushes me, and that’s the process improvement for me. There are lots of other things, but I’ve rambled. Hopefully, I nailed that plane. You can just respond or tell me if you agree with that.

Great investor track records do actually motivate and inspire me, but I’m not talking about a single year. I’m talking about an extended period of time with very strong returns over that period.

For example, if someone did 25% CAGR for 20 years, they’re probably one of the best in this profession who ever lived, especially if they do something conceptually similar to the style of investing I practice. In other words, if it’s Renaissance Technologies’ Medallion Fund, sure, they play a different game. It’s not what I do, it’s not what I know how to do, and I will probably never know how to do it. That’s not inspiration for me. I’m intellectually curious about how they did it, but it’s not something that will inspire me.

If you pick someone—let’s pick a great investor, David, who supposedly, based on publicly available information, had a fantastic track record for many, many years—that is inspiration. Joel Greenblatt’s track record, or part of it, I think, is published on the front or back cover of the book You Can Be a Stock Market Genius. That’s inspiration.

Single years probably aren’t inspirational for me and probably will not push me to work harder or think differently, but they will probably inspire me in the longer term.

Sure, I’ll answer your question. I generally agree with you, but I’ve gone a little bit back and forth on this, and I’ll tell you why.

I was talking to someone with a very good track record—borderline elite—who also runs a very diversified portfolio. They’re running 30 to 50 positions, plus long-short, so they’re running diversified, lower net exposure, and all of that. I was talking to them about someone who’s got a great track record—14 years, super concentrated, very good track record—and I said, “Hey.”

They said, “Look, I’ve talked to this person. I’ve interacted with them. I’m very skeptical that this person is a good investor.”

I was like, “Oh, well, track record—scoreboard, bro.” But look, this person is concentrated. Over the 14 years, they’ve had 10 positions, right? You don’t have statistical significance. They literally could be the famous coin-flipping monkey whose coin came up heads.

If 10 years ago you had plowed into Tesla, your returns would be incredible. I don’t know if they’re still there or not.

Now, again, stock price, bro—scoreboard, bro. But is that statistically significant? Ten years in a concentrated fund is not a lot of investments. I think the person is clearly wrong, but the person who said, “Hey, Warren Buffett still doesn’t have enough investments to be statistically significant”—I think that’s wrong. Fifty years is different.

But I do hear you: if somebody started in 2012 and invested today—if they had bought the Magnificent 7, they basically would be a legendary investor. I’m not sure: they had the insight to buy the Magnificent 7. Should we give them credit as a legendary investor or not? I don’t know. But I’ve come to be a little bit skeptical.

Part of that might be because you see some—I read all these business books; Byrne Hobart and I have our book discussions—and it’s a bit of a running joke: everyone you read about had a near-death experience. You’re like, they created incredible returns, but any return stream times 0 is 0. In the first 10 to 12 years of their return stream, there was this thing where, if things had broken just a little worse for them, it would have been 0. So, again, throwing a lot out there, but I’ve gotten just a little more skeptical of even longer-term track records based on that. How do you think about that?

I don’t think this question is answerable. Meaning, we cannot, through deduction, induction, or inference, get to a conclusion that we would feel 95% confident is accurate. I don’t think it’s an answerable question. It’s almost a question of belief: do you believe in X, Y, Z, or you don’t believe in X, Y, Z?

Now, if you believe in Santa Claus, probably that’s a little too much. But there are other things that are issues of faith, of belief, that people don’t need proof for. To answer this question—and remember, I don’t know the fund you’re talking about. I don’t know the fund or the person who is skeptical. I don’t know the person who delivered those returns. I have literally as much information as anybody who is listening to this podcast right now. So, I don’t know.

My logic is—what I’m thinking is this: very often, when people say, “John”—and this is a hypothetical John, so if any Johns are in our network, we’re not talking about you—“John is a great investor,” what it means, more often than not, is that the speaker and John are fairly similar in their thought process.

The speaker is thinking of himself, because all of us are above-average drivers and all of us are above-average investors and above-average runners and swimmers, even if you don’t know how to swim. So, that was a joke.

No, I was laughing because in college I was like, “I’m an above-average swimmer,” and then I hopped in a pool. I was like, “Oh, I barely know how to swim.”

What it means is that there’s a similarity bias at play. When the person meets Bob—again, hypothetical Bob—who does something different, and the person does not fully understand Bob’s style or Bob’s thought process or decision-making, they think that Bob’s not a good investor, even if the track record may indicate otherwise.

I’m not dismissing your argument. Maybe it’s not statistically significant; I get it. But that is not answerable. I’m just sharing how I view it, and as I said, I cannot prove that my way of thinking about it is right. But I think that it works for me.

Some people are superior business analysts. They’re just fantastic; they’re at the top. Sometimes that translates into money-making, sometimes it doesn’t. Some people are really good at idea generation. They may not be as good at actually doing analysis, but because they’re capable of finding great ideas, they make money. There will be people who are really good at risk management, and that’s how they make money.

That’s the one that I understand the least. But I have some people in my network who I think have this incredible money smell. I talk to them about a position, and I think that they don’t know this position as well as I do, even though I don’t own it.

But that person still makes money?

They do some analysis, and I’m not saying they are bad. I’m saying that’s not where they make money. There’s something else about them—their psyche, their process of thinking. I don’t know, but they make money. They have the money smell. And remember, I came up with the term “money smell” because I cannot come up with anything better.

So, if I were to assign probability, my base case would be that your friend who did not acknowledge the returns being deserved by another friend of yours—my guess is that they’re just very different. Remember you and I, when we were in Denver at Planet MicroCap, we spoke about the superpower. Different people have different superpowers.

I think it’s very important to have people with different superpowers in your network. Also, for people who didn’t listen to everything that Andrew and I talk about, when I define superpower, it doesn’t mean that you’re the best at a certain skill in the world. It means that it’s your relative strength. Andrew is probably better at lifting weights than running—or, as we found out right now, swimming.

So, in exercise, weightlifting would be his superpower. In investing, different people have different relative strengths, and that will be their relative superpower. That’s just the definition. I suspect that those 2 people you spoke about have different superpowers. That’s why they don’t recognize that, or at least 1 of them doesn’t recognize that. As a result, he doesn’t speak that highly of the investment talent and acumen of the person with a very enviable, very impressive track record. That’s my best guess.

No, look, let me just propose 1 more hypothetical. I’m laughing so hard because you and I were like, “Hey, it’s just Artem and Andrew rambling.” We’ve got this buzzy, like ESPN-would-lead-with-it first question that we’re going to get to in a second, but we said we were going to kick off with that, and then we got into a conversation on the statistical significance of investor returns.

Let me ask 1 follow-up question. Somebody who bought Bitcoin 10 years ago and held it until today, I don’t think it’s crazy to say they have better returns than every investor that we know, right? Bitcoin 10 years ago was $200, and as you and I are talking, it’s $115,000, right? So, in 10 years they’ve got—what is that?—a 500-plus-bagger, right? They bought in any size. They’ve got just absolutely groundbreaking returns.

Is somebody who bought Bitcoin 10 years ago and rode it until today a great investor? Because their returns would say that they are literally the best investor of all time.

I would say that there is a very high probability that such a person is a great asset allocator, which is different from an investor. That’s number 1. Number 2, if that person—and the answer, I think, will be very context-specific—if that person in 2012 had a $3 million net worth and put $2.8 million into Bitcoin, I would say, “Boy, Russian roulette also statistically works, but it’s a dumb idea to play it.”

However, if that person, with the same $3 million net worth in our example—and, by the way, let’s imagine also that the person is 30 years old, healthy, has a good professional career, et cetera—if the same person put even 10%, $300,000 in our example, into the same Bitcoin in 2012, I would say they’re probably a great asset allocator.

I would also say that they may be good—and I’m saying “may be” because here we’re speaking about a statistical sample of 1 in your example—at envisioning different future states of the world. And 1 of those potential future states did play out, and it played out in an incredibly favorable manner.

No, look, I hear you. It’s just, I’ve been working with a coach, and one of the things they keep telling me is, “Andrew, you want a definitive answer to all these questions, and there are no definitive answers.” I suppose it’s the beauty of investing—the art of investing. But the logical side of me wants to read a story and know the ending.

With investing, there is no right answer unless you’re Renaissance Technologies and you run 1,000 trades per day, and then you can find statistically significant ones. But, like I described, the Bitcoin thing: they were kind of a VC, and they were right. It’s scoreboard, bro. But it’s up. Okay, anything else there? Because I want to get to our first take question.

Okay. Before you get into what you wanted to ask me—and somehow 20 minutes later you still have not asked me, or 15—I would also say that when you mentioned that your coach told you, “Oh, Andrew, you want to know the definitive answers to questions that don’t have such answers,” I would make things even more complicated.

Very often, if you are investing in companies that are commonly called compounders, I’ll try to narrow the definition a little bit: a company with attractive unit economics, with a long growth runway, executing on that runway and doing it in a very good way because it has strong management that can get the execution done. Again, I’m not saying that this is the classic definition. I’m just oversimplifying here.

You are statistically betting on outliers. Meaning, most companies will not deliver, let’s say, 10 years of 20% revenue growth, right? Or 20% earnings growth—you pick the numbers. Statistically, it’s not going to happen. So you are betting on outliers, which means that your base rates, if you just take the whole universe of stocks, will be against you.

Now, if you start cutting and recutting that universe—and I’m making these criteria up, okay, so I’m not saying that this is a good criterion—if you say, “I want a founder-owner-operator with at least a 20% stake, and I want companies that have—pick margins, pick revenue growth, pick ROIC, pick whether they got backed by Sequoia or Andreessen Horowitz, or any other great VC that we think are great investors”—whatever your universe is, in that case, you may be turning base rates in your favor.

But even with those, I'm not sure they will be massively in your favor. So you definitionally, from the beginning, cannot get the definitive answer. And if you get a somewhat definitive answer—“somewhat definitive” in quotes right now—then you're basically against yourself. You're betting on outliers in this example, in this style.

Look, that's a really fascinating way to frame it. I love the Warren Buffett example where he had a University of Chicago expert who said, “Oh, Berkshire, you know, there are 3-standard-deviation outliers, 4-standard-deviation outliers, 5-standard-deviation outliers.” Eventually, the economist just put his money in Berkshire and invested in Berkshire.

What you're kind of describing is that, when you're betting on these great compounders, the average company—you know, what's the stat? It's not the 80/20 rule in terms of 80% of the market's returns being driven by 20% of the stocks. It's like 80% of the market's returns are driven by 20 stocks or something over the very long run.

It was a famous academic paper published several years ago by a professor—I forgot his name. I think he's from Texas, but I can be wrong—and it was like 4% of stocks generated 100% of the return, et cetera. There is this—

I think there was some survivorship bias there and stuff, but the point was—but I love what you're saying. The compounders are like you're betting on a company being one of those 4%, right? You're betting against base rates. It's a really interesting way to frame it.

But another thing that comes up here is that—and, by the way, to defend that professor, any paper of this nature will have some limitations. It's probably impossible to make it perfect, and obviously many people will say, “Oh, you missed this, you missed that.” Sure, we're talking about the 80/20 rule, meaning the intellectual 80/20 rule: do you get 80% of the insights even if 20% of situations are not covered?

Since we got into this topic completely unexpectedly, if you say, “I am betting on statistically less likely events from the base-rate perspective,” then the probabilities, if you do a random dart throw, are against you. So, number 1, you need to figure out how to improve the probabilities. Sure, that's obvious. But number 2, you need to have—I think many investors call it convexity. I don't know whether mathematically it's convexity or not; that's probably somewhat debatable, but let's call it convexity here. You need that convexity of returns on the upside. And this is where I think many investors can learn a lot from venture capital.

With their mindset and how they think about it. Obviously, the most famous one is the power law. I think the first time I learned about it was from the fantastic book Zero to One by Peter Thiel. Then there's another fantastic book that I read 2 years ago, I believe: The Power Law by Sebastian Mallaby. This is the same author who wrote More Money Than God. I actually like The Power Law even better than More Money Than God, but that's a subjective assessment.

So that's a fantastic book about many great venture capitalists. Now, I'm not saying that public-market investors should become venture capitalists. That's a different game. And, by the way, by the time public companies become public these days, they're way past those early stages where venture capital can invest.

However many bags you want, you can have.

So, yes, I think that's another consequence of that framing.

All right, we're going to hard pivot. Let's ask the ESPN First Take question. You and I were talking the other day. We're talking about AI again.

Look, I've talked to some investors who are a little bit older than me, a little bit older than you, and they're like, “Oh, I haven't used AI at all yet.” And I'm like, “Okay, cool.” It's like you're basically an investor in 2005 who's like, “Oh, yeah, I'm not using spreadsheets and email.” It's so revolutionary to me, and it changes so much.

And expert calls—I just think it's not the only way, but it's such a useful process. For so little upfront, you can talk to people in the industry, get real insights, and everything. So they've really evolved.

Anyway, we were talking about AI because I was saying some of this, and you told me, “Hey, I think AI over time is going to make markets—I don't want to put words in your mouth, but I believe you said—make markets dumber.” In my interpretation, that meant less efficient, or easier for active investors to generate alpha. I kind of disagreed, but I don't want to put words in your mouth. So I'd love to ask you: Am I remembering this correctly? Did you think AI will make markets less efficient and kind of easier for active investors to outperform over the long term?

First of all, I don't believe I said stupider.

I think you said stupider. I think stupider was the correct word.

I think I said dumb.

Dumber. Okay, okay.

First of all, I think I would have said “more stupid,” not “stupider,” if I had used that word.

Okay, dumber. Let's go with dumber.

I think that's what I said. So, let me explain that. I'm not sure whether it's a consensus view right now among investors. I firmly disagree, and I would think your view is quite nonconsensus, though I don't think it's completely out of consensus.

Okay. So let me talk about how I got to that thought, and I also want to caveat something. I am not a coder. I'm not a software developer. My typical solution to any tech problem is to restart my computer, and if that doesn't work, unplug it from power for 30 seconds and then plug it back in.

Yeah. I want to be very clear about that. I also want to say that this applies to anything that I say, but here it's even more so. I could revise my view 6 months later, or 3 months later, or 12 months later, or even tomorrow, and I am not a futurist. I am not predicting the technological future.

This is my thought process. Where we started with you today is that investing is a craft, and it takes a certain amount of time to master that craft. The best way to master the craft is if you have a great mentor and you are working directly for him or her, and you are learning and copying that. That's probably the best.

If you didn't have that opportunity, maybe you figured it out yourself. You probably spent more time figuring it out on your own, and you probably got more scars on your back from your mistakes, et cetera, some of which were self-inflicted. But that's how you arrive at your process for generating insights and reaching conclusions.

In order to make strong returns, you need to arrive at some insights into the business, the investment setup, or something else that is usually fairly nonobvious. And I think there is a danger. I'm not saying it will happen; I'm saying it might happen. Going back to our conversation about Bitcoin, it's one of those future states of the world.

It doesn't mean that it will be the state where people—especially younger generations of investors—will outsource their thinking to AI. I would suspect that your parents' and my parents' generations, on average—and “on average” is the key word here—are better at navigating a city without a GPS. I expect that our generation is worse. I also suspect that those who are now 25 are probably even worse than us at that.

It's a skill—I may have mispronounced that, so you'll need to correct me—that just goes away for people because it's so easy to plug in your GPS. And I think there is a real danger that may happen with how people think, how they do research, and how they develop conviction in investing.

AI is an incredibly powerful tool. There is no doubt about that. Everybody says that. That's not insightful. But what it will do to our brains is an open question.

But as AI stands today in early August 2025, I think those people who succeed are the people who figure out how to use AI as a supplement and who do not fall, intentionally or unintentionally, into the trap of outsourcing their thought process to—

Well, and also something else you said: you interpreted my point as more efficient. I think market efficiency, or efficient markets, is such a loaded term that everybody means something different by it, so I would rather not use it.

Okay.

What I think might happen is that the market will become more consensus-driven because people will be asking roughly the same questions to roughly the same large language models. There will be some variation in their prompting skills and variation in the particular model that they're using. Some of them will be better at any single moment in time because there's a race between them. Sure, there will be some variability, but I think the answer that will be given will be fairly similar in the big picture.

What it means, I think, is that markets will become more consensus-driven. If you figure out where AI is missing either important data points, or information is just not out there—meaning it has not been written on the internet, in articles, in expert calls, somewhere else—your opportunity to generate outsized returns, because everybody else is running left and you realize that you need to be running right, I think will likely improve. That's what I meant by dumber, and dumber maybe is too strong of a word, but it will be more consensus-driven.

So, honestly, as you described it, I kind of don't know if I've got huge disagreements with you. I'd kind of analogize it like this. I like to say, hey, if somebody comes to me right now with a P/E-multiple-based investment, right? They say, “Hey, Andrew, stock XYZ is trading at 8 times price-to-earnings.

I think it's a buy. I'd be like, “Cool. There's no alpha there. You've just bought beta, and probably negative beta, right?” Because 50 years ago, 60 years ago, there was alpha there. There weren't computers doing that. You could calculate it and buy something.

About 20 to 25 years ago, computers got so good that they were automatically doing that. It was kind of priced in. But what that meant was, in, let's call it, 2010, that got priced in, and the returns to finding things that were non-obvious values—and I would think about many of the compounders you talked about, right? Early Google, Facebook, all this sort of stuff, and Amazon.

They did not trade for 10 times earnings, so they weren't traditional value, but they had huge growth, wide moats, and lots of ability to compound. For many of them, I don't think Amazon was guaranteed, but Facebook and Google, once they kind of locked in—especially Google—it was kind of guaranteed that they were going to grow and continue taking share. If you could figure that out and invest beyond “Google trades at 25 times next year's earnings,” and you could see the growth runway in the unit economics, you could make a fortune.

I think what you're saying is, “Hey, AI is going to price out maybe a lot of the medium-term insights,” right? AI is going to have a lot of insights. Maybe it's going to price out the alpha in buying the compounders and buying the unit economics. But if you can figure out something beyond that, the returns to buying Google were exponentially better if you figured out that insight than buying U.S. Steel at 4 times price-to-earnings in the '80s. If you can figure out the next insight that AI is not going to generate—and we can talk about what that would have to be—the returns might actually be exponentially better. Am I saying that and summarizing that correctly?

I wouldn't say that you're summarizing. I would say that you are putting another angle on this issue. Some elements, I think, are fair. Some elements, I don't know. I don't have a strong answer whether AI will be pricing out medium- to short-term alpha versus long-term alpha. I don't know. I don't have an answer. I don't think I'm prepared to figure out an answer right now. As time passes by, we may figure it out.

If you were to ask me now, I would probably say that I think it will be more dispersed. I think it will be more consensus-driven. That's kind of how I think about it, because if AI will be giving roughly the same answers to a bunch of analysts, getting an answer from AI, in my opinion, is very different from reading 15 earnings calls or conference transcripts, or listening to them live, if you prefer listening to reading, or, better yet, conducting tons of expert calls yourself and slowly getting to that mosaic and putting it together.

I think it's very, very different in terms of the process—how you get to that conviction and deep knowledge ingrained in you—versus reading a 30-, 40-, or 50-page output from AI, which can be useful in certain cases but should not replace your brains and thinking.

And then feeling that I have this knowledge—I call it fake knowledge. You don't know; maybe you memorized it, maybe you haven't, depending on the quality of your memory and its capacity. But it's not the same as figuring out those things yourself.

So that's how I think about that. What I think AI will probably do is make different competitive advantages—or competitive strengths, which would be a better term—that people have obsolete, and then you need to reinvent yourself and figure out your new competitive strengths.

I'll give you an example. If someone was really good at writing 10 years ago, you were just like, “Brilliant writing. You're engaging. You're convincing. You can figure out how to deliver insights and punchlines.” That's a fantastic skill. Not everybody can do it. Guess what? Now, you still may be better than everybody else, but everybody else with AI—the gap between you and everybody else who is average with AI—has closed dramatically.

I think it's an overall theme, by the way, of technology and technology-based investment and research tools, including AI, expert libraries, and expert calls. It's the overall democratization of access to information and knowledge. I also believe it's closing the gap between investment firms, or investors in general, with different financial resources.

In other words, hypothetically, let's pick expert-library tools, for example. If 10 years ago, before expert-call libraries came out, a person at a $2 billion fund had 100 units of utility because they could access bespoke expert calls and pay $1,000 to a provider, that's very expensive. Now, if you're a trillion-dollar fund, you have the budget. You're fine.

If you are a $5 million fund, for example—I'm potentially taking something very, very small—you couldn't afford it. You're out.

Then, with the advent of Tegus—and now it's all under the AlphaSense umbrella—you can subscribe, have access to a library, and do your own bespoke calls at a lot cheaper price.

Yes.

That democratization of access to information, knowledge, and insights is fantastic. I think that if, before, the gap between a small fund and a very big fund was 100 units, or 99 units, the gap still exists—don't get me wrong—but now it's maybe 20 or 30. It's a lot smaller.

No. So I think that's interesting, because what you're saying is that one area, again, 30 years ago—probably more than 30, 40 or 50 years ago—your edge could be, “I can calculate price-to-earnings.” Your edge was basically, “I can do math better.” That gets arbitraged away by a lot of things: quant funds first, then Excel, so everybody kind of gets it, and then quant funds start running and automatically applying it.

I think what you're saying is, “Hey, over the past 10 to 15 years, one of the areas of edge might have been that I can ingest information either more or better.” As you said with the big funds—and you're 100% right—a big fund now, but especially 10 years ago, could afford to go do 10 expert calls and throw down $10,000 of research expense on a position, whereas a small fund just couldn't do that.

With expert libraries, that's moving away from AI. But a big fund might have had 5 analysts who could summarize every sell-side model, every sell-side note, every expert, all that. AI can do that for a small fund now. So that summary of information goes away.

The new question going forward is: What is the thing that will kind of get amplified by AI? I've talked to Artem about this, and, in my view—and you can tell me if you disagree or if I've completely misinterpreted the premise—as that ability to summarize big things of information gets consumed by AI and maybe simple theses get spit out by AI, I think the next wave is, “Hey, can you go find me things that are not on the Internet or that are not publicly available?”

Obviously, I'm not talking about MNPI, but I'm talking about, “Hey, the next wave might be that you're researching a retail company. Can you find non-obvious sources of data?” Or you're researching a lumber company. Are you the person who can go to a lumber conference and network with 15 people at the lumber conference, so you've got a read on how everything's going that literally no one else has because it's bespoke? You created it yourself. I don't know. Do you agree or disagree there?

I think we are looking in the same direction. I agree. I will add a couple of examples or wrinkles or nuances. By the way, “wrinkle” is an interesting word. We can speak about it in a second.

I think I unintentionally created a pun here, even though I didn't think about it before I set it up.

Wait, say it again. My wife is so much better with puns and wordplay than me. What's the pun?

No, I remember I said there may be a couple of other wrinkles on what you said, and then I realized that my example actually fits nicely into wrinkles.

Okay. Okay.

So, big picture, I agree with what you said. I think situations where information does not exist in written form at all—that's one of the areas where it will be fascinating.

Let's step back. As you know, Kakuna Capital mostly invests in small- and micro-cap securities, not exclusively, but that's what I mostly do. That's where I spend most of my time. Generally, if you go into, let's say, AlphaSense or Tegus—and again, I will use them interchangeably because they're in the process of converging the two—if you pick a $20 billion market-cap company, there will probably be a few dozen calls. That's probably pretty reasonable, especially if you pick TMT companies, which are more popular among investors for doing expert calls. There will be a lot.

My ideal setup is that I go to the expert-call library, I put the ticker in, and it gives me: “There are no expert calls available. Would you like to request one?” Or maybe there are 1 or 2, and ideally they're also kind of old.

In that case, let's pick a real company. Disclosure: Kakuna Capital LLC and all its affiliates own shares of Sofwave, which is listed in Israel. It's a company that sells medical devices for skin tightening and has a recurring-revenue business model—a razor-and-razor-blade model. One of the consequences, if you are a patient using that, is that you remove wrinkles.

I said it was absolutely accidental, and then I was like, “Oh, I should use this as an example.”

When I first learned about the company’s existence and started doing research, there was only 1 expert call on AlphaSense. That’s the library that I use—only 1, not many. By the way, it was somewhat negative, in my opinion, or at least it was pointing out areas where the company needed to improve itself.

By the time I read the call, I thought they had already fixed all the things that the former employee was mentioning. So I felt like, okay, I guess those were probably reasonable concerns, but as the company had grown, they had fixed them. Fine, but that was all.

Remember, this company is listed in Israel. The financials are in Hebrew, and I don’t speak Hebrew, unfortunately. How are you going to interact with AI in that situation? You have an expert call and filings, and different tools will happily translate those filings for you. In fact, I use them for that. But it’s very different to interact with AI in that case: there is no information. A large language model needs language to give you answers—something in writing.

As part of my work, I’ve done multiple expert calls. I spoke with patients—sorry, anecdotally, that’s not expert calls—but I spoke with doctors, who are the people buying the product. They engage with patients to explain the benefits of the treatment and the use of those devices. I also spoke with former employees about the company culture, the market, the sales process—everything.

If you now go into the library and put Sofwave in as a query, you can probably get, by now, 8 or 9 expert calls, would be my guess, roughly, at this point in time. You can use AI to ask thoughtful questions. Now, you also need to have thoughtful questions, so you need to have an investment process.

AI will not replace your investment process. If you have a good investment process, AI will make it more robust. If your investment process is not robust, or you don’t have a good process at all, then AI will make things even worse. That’s my subjective opinion.

Now you can interact with AI, ask questions, and get to those takeaways, conclusions, insights, and punchlines a lot faster. What I think it means is that original ideas with very little information in writing, and information that is mostly qualitative, will of course be more important to find early. But I think the discovery period for those ideas and the market repricing of those ideas will probably shorten.

Well, look, I think what you described fits into my thing, right? I was talking about finding nonstandard places of information. In my head, I was kind of thinking about a $2 billion company where you go and get gut checks at an industry conference or something. Obviously, what you described was finding a company that was small enough that there wasn’t much information out there, and by doing the expert calls, customer talks, and everything, you created the information for yourself.

Now, that doesn’t mean it’s a good or bad investment, but if you have 100 companies with not a lot of information out there, you go create it all. Eventually, if you’re using a public expert-call transcript, in this case, it goes public. But you create all the data and then feed it in, and you do have an edge because you created it.

Hopefully, you can create that data, refine it, and execute on it properly. I certainly know there have been some undiscovered microcaps that I’ve invested in where I thought I had an edge, and I kind of wish they had gone undiscovered by me in the long run. But you are describing the same thing: creating information and having a data or information edge there.

I want to ask you something, since this is supposed to be process-themed. I think about process all the time. What is something that you did 10 years ago, as you were prepping for the prime of your career—is the generous way I’d put it—that you’ve changed or improved on? Is there anything that you think has made a noticeable difference in your investing process?

A massively bigger focus on the customer. More focus on the company’s customers.

Okay, just give me an example.

Early on, when I was running Karakhan, I made a very painful investment mistake. I invested in a company called Agrify.

I know Agrify.

A significant loser. I think you and I spoke about it. Back in 2015, with the full benefit of hindsight, if I had spoken with a few customers, I think I probably would have realized that, while customers value the product, there is a difference between satisfied customers and raving fans.

Ideally, you want to invest in companies that serve a constituency, and that constituency consists of raving fans. I think I could have done a substantially better job and avoided a painful mistake, both psychologically and in terms of P&L, if I had applied that framework to understanding Agrify. I could have done a lot better.

What it led me to is that I need to understand the customer value proposition better. And guess what? What’s the best way to understand the customer value proposition? Talk to customers.

Ideally, you could visit them, right? Spend time with them. But sometimes it’s an intangible product, so you can speak with them and learn about why they consume it, et cetera, et cetera.

By the way, I think doing expert calls is very, very valuable for understanding the customer value proposition: why customers are not switching or are switching, what the competing solutions are, the buying process, et cetera, et cetera, and who the decision-maker is within the customer organization if it’s a B2B product. You need to understand all those things.

The old-school way would be to ask your network. That’s the best because you will get unfiltered answers, but I have a reasonably large network of people from business school and other venues in my life, et cetera, et cetera. There are natural limits there, so you need to ask for introductions, or introductions to introductions, et cetera, et cetera.

If you could find someone like that, that’s the best. But if you couldn’t, then either go on LinkedIn and start cold-calling—fantastic—or use expert-call providers. In the past, it was very expensive; the cost has gone down, as you and I discussed 5 minutes ago.

Let me ask a couple of questions there. I like that answer. I do. But, as I said earlier, I want 1 rule, right? I want 1 rule to rule them all for everything I do.

One of the tough things I’ve found with customer calls is 2 things. Number 1 is, you know, what’s this? I think it was Steve Jobs who said, “The customer doesn’t know what they want until we give it to them,” right?

Sometimes I can think of recent examples. I don’t quite want to disclose the company, but I’m sure people can think of something similar. A company is launching a product, right? It’s an improvement on the current product, but word gets out and people know. I talk to the customers, and they say, “Oh, whatever. We don’t really need that product. We don’t think it’s that much better. We’re not going to upgrade,” whatever.

Then the product comes out, and you talk to the same customers 2 weeks later, and they’re like, “We love it. We’re invested. We’re upgrading all of our equipment to it.” These are generally B2B products, but there are exceptions.

That type of stuff is 1 thing I worry about: the customer can be so fickle. Obviously, we were talking about something that’s yet to be launched versus something that’s launched, but customers can be so fickle with that. I worry that I talk to them and have 1 opinion, and then the next week, if I talk to them, I’d have a completely different opinion.

Do you want to pause the recording and rewind to the place where we spoke with you about base rates, how you’re trying to recut the universe, and how you’re figuring out that you’re betting on outliers? This is my response to, “I want 1 rule.”

I know, I know this. Andrew, this is my 1-rule answer to you: there are no 1-rule answers at all. That will be my response.

Look, at the end of the day, what we as managers are getting paid for is judgment and decision-making. Everything else—talking to management, attending investor conferences, attending industry events, talking to former employees, searching the web, talking to AI, reading, modeling, whatever you do—that’s simply an input. It simply feeds into the output, and the output is your decision. Your decisions are based on judgment.

So I think what you’re saying is this: How can I make the same judgment all the time in a similar set of circumstances and be right? The answer is, I don’t think you can. It boils down to—and remember, it’s a lot about nuances or wrinkles.

In your example, you started with Steve Jobs’s famous quote. First of all, it’s about B2C. Second of all, it’s about true breakthrough, disruptive innovation.

Mhm.

When the product does not exist, I don’t think that question is knowable. It’s unknown and unknowable. You wouldn’t know whether that product will take off with customers or not.

And by the way, if you had had the insight that this product would be fantastic when the iPod came out, you would probably be calling me right now from your BBJ, from your own private island.

No, let me give you the other side of it. I know. Yeah.

So that's B2C, and that's unknown and unknowable. Then there's another end of the spectrum where the products are already there. It's a B2B software company. Let's pick HubSpot. I don't own HubSpot. I admire the business and what they've done and what they've built.

If, a number of years ago, you spoke with 20 customers using HubSpot, you would probably figure out that it's a very, very good CRM, martech, or SaaS company. You would probably figure out that the product is good, customers are satisfied, and they're happy. Maybe they're even more than satisfied.

That's a very different thing from talking to them about a product that doesn't exist but has a particular feature. It's very different. So I think you need to calibrate the tools that you're using. Do you play golf?

I did when I was a kid, but right now, no.

Okay, fantastic. So I—

I know the rules just fine. Yeah.

Okay. So I don't play golf either, so that's why I'll give you an analogy from golf, because I have no idea how to play it. In golf, as far as I understand, there are many different types of clubs, right?

I cannot believe you're giving an analogy from a sport you don't play.

There are. Yes. Depending on the terrain and the circumstances—wind, whatever—you use a different club for each situation. I think with your approach, you have a one-size-fits-all answer to all situations. You're trying to play in a golf tournament, and you brought only one club.

You are correct.

It works.

You have all of them and need to figure out which one to use when.

I can't believe how good an analogy you just used with a sport that you don't play or apparently know the rules to, because that was excellent.

So now—okay, I'm glad to hear that, because otherwise it would be very embarrassing if I gave you an analogy that was totally misplaced.

Let me go to a completely different question.

Okay.

One thing you and I talk about quite a bit—I think people can probably hear that—is the small-cap value investor fund community. It is not large. We probably talk quite a bit to the same people.

There are a couple of stocks where, if there are 20 small-cap value investors and you read 20 of their letters, 12 of the 20 managers are long the same stock, right? I don't need to name names, but I'm sure everyone who follows the markets closely can think of something as soon as I say this.

And guess what? It doesn't just apply to small-cap managers. Bill Ackman, for a long time, would go long a stock, and then all of a sudden there are 15 other people who run medium- to large-sized hedge funds who are long the same stock. It happens. Tiger Cubs, Tiger Cub allies—one of them is long, all of them are long. It happens.

You talk to smart people. You share diligence. You all come to the same conclusion. That's great. Hopefully, you're talking to really smart investors who have really thoughtful theses and stuff. But one thing I always worry about is that it's very easy to start outsourcing conviction and outsourcing thought. It's very easy to get into groupthink. It's very easy to get into the same names.

I feel like I have a higher bar when a friend brings me a name, because I'm worried that, because I respect them, I will have a lower bar for it. I'm worried about groupthink. I know you mentioned Sofwave, and I looked at the chart recently, and I'm kicking myself. You and I talked about it quite a bit, and there were medical-device issues, which I can talk about later.

One thing I always worry about is having a lower bar when a friend brings me a name. I'll just call them up after I record and say, “Hey, what do you think?” and then kind of rely on that.

So I just want to ask you: I'm running Yet Another Value Podcast—330 episodes, 315 of them are individual stock pitches, right? When a listener listens to this and loves an idea—and obviously they're not investing, but they love an idea—when you and I talk, or when you talk to friends, how do you dial back and avoid the groupthink? How do you avoid thinking, “I'm outsourcing my thinking,” or, “This person has biased me in an investment,” even if you're going to run the bases and do all the work on it?

How do you avoid letting somebody you respect—who has put their money behind it and put their belief behind it—influence you?

So, first of all, when you hit 400 podcasts, we've got to do another interview where I come and interview you. I promised to wear the Yet Another Value polo that you kindly gave me.

There we go.

Another interview. It will be the second. Remember, we did one after 210?

We did. We did. And on this one, you're going to show off the Yet Another Value Podcast tattoo that you got as well, right?

Oh, yeah. Yeah. No, I would not. But I can show a bottle. I can show a bottle. It's a very nice bottle, by the way. Great size.

So, the one-line answer would be: in the situation that you describe, I want to make the conviction my own. Don't outsource conviction. That's a dumb idea. Make the idea your own.

There are some ideas in my portfolio that I probably got from a peer or friend whom I like and respect for their thought process. But by now, I may know some of those ideas—not all of them, but some of them—better than they do. Even if they still know them better because they've spent longer researching the company, I have my own conviction in them. That's what I think we need to be aspiring to.

Number two: remember we said “low bar,” right? I don't think that's the right way to frame it—a low bar to invest in an idea because it came from a friend, especially if you like the friend. I think the low bar should be applied, if an idea comes from a friend or respected peer, to putting that idea at the top of your research pipeline.

There are some friends in my network whom, if they call me today and say, “I have a great idea. Let me tell you about it. I'm so excited about this idea,” I'll probably drop almost everything and start researching that idea. I will not go and buy it, but I will start researching it. It will trump a bunch of what I had in my pipeline that I sourced myself or already got from someone else.

So I have a lower bar there, but my bar to invest could be the same. That's what I aspire to. I am a human being. I'm fallible. There are probably cases where I did not execute the vision that I just described, and maybe I did indeed lower the investment bar. It's possible, but that's my inspiration. That's what I force myself to do.

The way you force yourself to do it is by following your own investment process. If your investment process is, for example, to read 3 years of all earnings calls and conference presentations, do not shortcut it just because the idea came from a friend whom you like and respect. Go and do it.

Again, I'm using very simplistic, very rule-based tools. If your investment process rule is to read 5 customer expert calls—or do your own if they're not available—and, by the way, 5 is a pretty low threshold, especially if there are 20 in the database. It's a pretty low threshold, or 30 or 50. Go and do the same number of calls that you always do, either your own bespoke calls, if that's your process, or, if you're okay with reading someone else's calls, fine. Go do it. Don't cut it.

If your process is to talk to management and ask them the questions that are on top of your mind, do not cut that step out just because the idea came from a person whom you deeply respect. Again, it's one thing to say, “I want to have the same level of conviction in my own ideas,” but the way you get it done is through your processes and your routines. That's how you get it done.

And then, also, this is very important about ideas coming from your network. I'll use myself as an example. If I call you, Andrew, and say, “Andrew, I got a great idea for you”—first of all, I hope you will pick up the phone.

No, I said I'll call you when I'm walking Penny. I'm with Sylvie. I'll call you when I'm walking Penny.

So, if I call you and say, “Andrew, I found this company. It has a $200 million market cap. It's listed in a country where most people don't look, and there is no sell-side coverage and no Value Investors Club write-up. Or maybe there is 1, but there is very little information, no expert calls, and blah, blah, blah,” I would suggest that you listen to me carefully and thoughtfully and maybe put that idea at the top of your research plan.

If I call you and say, “Andrew, I found this great investment setup. It's a coal company on the East Coast trading at 4 times earnings, but they're going to pay a big special dividend,” I think you should hang up on me.

Funny you said “special dividend.” I was like, “I'm here for it.”

I know. But remember, I usually don't call you with this type of idea. So you need to calibrate when an investor friend mentions an idea. You need to calibrate whether that's a play.

Imagine a hypothetical investor with a very strong track record, okay? You need to figure out—in my opinion, I want to figure out—where those returns came from and what types of patterns and situations produced them.

If those ideas came from cold names trading at 3 or 4 times earnings, returning capital, doing buybacks or a tender, or whatever, I probably won't listen to that person when they pitch those types of ideas. If that person made money on software-as-a-service names...

So software-as-a-service companies that are $1 billion to $2 billion in market cap and are about to become $5 billion or $10 billion companies in the next few years—I want to listen to those ideas from another person. I think collaboration is very, very, very important.

No, that's super interesting. Especially because there's a company you and I have been talking about that fits one of the descriptions, and it just rips higher in my face every day. I'm like, “I need to be in here.”

One of the great things about being a podcast host is that I can ask completely simple questions under the guise of asking for my listeners. I want to end with one question that has really been on my mind recently in terms of process improvement.

We've talked a lot about ideas and stuff, but in terms of sizing, force-ranking—however you do it—your ideas, how has that changed for you over time?

That's a process, and that's a difficult one. It's difficult because you would never get everything perfectly right. That's difficult.

This is psychologically challenging. I don't know whether the people who listen to this will know this about me, but in my prior professional life, I was an international tax lawyer in New York, working for a big American law firm.

Think about this: I'm an associate, and I'm getting an assignment from a partner to research a certain situation. Let's say our client wants to do a tax-free spin-off.

Yes.

Do we have the facts that will support tax-free treatment, et cetera? We need to do the analysis.

If I come back to a partner and say, “Hey, this is my legal memorandum,” let's say a 10-page memo, and that memo is about 60% accurate, I think the tax partner should take me and kick me out from the 38th floor of 200 Park Avenue—that was our office—and they would be right for doing so. They're very nice people, and I hope they wouldn't, but you see my point.

If you're right 60% of the time in force-ranking your positions or managing your portfolio based on your expected IRR, you're probably doing pretty well. That's a very challenging mindset for anybody.

That's why I think people who play a lot of probabilistic games and come to investing may have a natural psychological edge. They're used to playing poker: they may have a good hand and still lose it. That's a very different mindset.

For me personally, it has been a challenge to adjust since the time I left the legal profession, went to business school, worked for another hedge fund, and then started Kakuna Capital. That's not easy psychologically.

What I want to tell you is that it doesn't mean I have the perfect answers, and it doesn't mean I always get it right. It's a work in progress. I think I've become more mindful of looking more carefully at expected IRRs. I can't say that I've never done it; it's more about doing it in a more systematic way and comparing them. But it's still very, very tough.

Expected IRRs are an interesting one because I was talking about this with someone the other day. I've gone back and forth so many times on it.

My issue with expected IRRs is that when you do it by expected IRR—and there are other ways to modify this—it would push me toward the most leveraged stuff out there, because the most leveraged stuff is what gives you the highest upside.

That could be the correct answer. Ted Weschler, who is one of Berkshire's portfolio managers—everybody talks about his IRR, right? His IRR was 2 or 3 bets: super-levered companies that survived and went up 100×.

It's possible that we should be taking advantage of the convexity of non-recourse leverage in the stock market by buying companies that are super-levered. But what I was finding was that when I would force-rank by IRR—and at some point I would cover about 50 companies—I would do a little, “Hey, here's my fair value.” I sound like a sell-side analyst: “Here's my fair value, here's where the stock is, and which one trades at the biggest discrepancy?”

What I noticed was, “Oh, crap. All of a sudden, my portfolio is just the most levered companies I follow in general.” That's why I've gone back and forth on IRR.

Then I had somebody say, “Hey, well, charge yourself a volatility factor,” right? Or you can do it on an unlevered basis. Judge everything on an enterprise-value basis and how much of a discount to enterprise value it trades at.

But when you go that route, what do you end up with? For the most part, you end up investing in companies that have huge net-cash balances. You've got almost no beta on the upside because, yes, their enterprise value is $50 million and they're trading for $25 million, but it's all net cash and you're just waiting for the cash to be returned.

I've gone back and forth so much. Go ahead. Go ahead, please.

You want a one-size-fits-all answer, as you and I—

I know, I know. But I'm also just discussing and being a very handsome podcast host.

Oh, absolutely. So, first of all, stylistically, I think you and I diverged the most philosophically. I do not like levered companies.

Remember, you and I were running an informal, no-recording book club that we need to resume. We discussed a particular book, one chapter per conversation, and our biggest disagreement on one of the chapters was, “Do not invest in levered companies.”

For me, that was very natural. “Yep, I subscribe to that view.” I rarely invest in levered companies. Again, I put the book down and never read it again.

You're like, “That's nonsense,” right?

So you and I disagree philosophically, and that's totally fine. If you invest in levered companies, or you've invested in a number of them in your portfolio, yes, if you rank them all purely on expected IRR, you will indeed end up with all those names at the top of the list.

What I found interesting was that you said, “Okay, then I tried to do it based on volatility, but then I ended up with a bunch of companies that have very little enterprise value because they have too much cash.”

The volatility was that you charge yourself a volatility factor to account for that. I struggled with that, so then I switched to EV, and then you just get a bunch of net-cash companies.

And I think, in my opinion, if I were you—if you hired me as your portfolio-management coach, which, by the way, I don't think you should, but if you were to make that dumb decision—in that case, I would say, “Andrew, you need to apply a penalizing factor to those.”

Simplistically, you would have a 2-variable ranking system. One would be expected IRR, but then you need to add another one that will account for what I will call—putting “risk” in quotation marks—risk. I'm not talking about volatility here. I am talking about the company going bust.

When I said volatility, I meant risk. I wasn't talking about stock-market volatility, but please continue.

Okay. You would most likely be applying a fairly significant penalizing factor, or downward adjustment, because it's a levered company.

If you tell me I have 2 stocks, I only imagine the world like this: now we'll play Microeconomics 101. There are only 2 products, product A and product B. There are no other products in Microeconomics 101. There are only 2 stocks, stock A and stock B.

Based on your math—and let's assume that you have really good math, really good estimates, and everything—there's a 25% IRR over the 3-year holding period on each of them. I will say, “Fantastic candidate. What's the name? Give me a call. I want to know the thesis.”

But then I will tell you: one of them, you tell me, has zero debt. Another one has, let's say, 6× EBITDA, 4× debt, and 2× equity. I will say that these 2 opportunities are not created equal.

Right now, I don't know exactly all those variables, but I think you need to force that and apply some kind of borderline common sense. The IRR shows that my math, my Excel spreadsheet, my master Excel file, spits out the same IRRs, but they're not created equal, and then you adjust for that.

For me, it's probably a little bit easier because I don't invest in as many companies. But I have another issue. A few months ago, you and I were discussing this, and I think you mentioned it in one of your rumblings without naming me, which I appreciate—the anonymity.

At some point, I think you and I chatted about whether we as investors take upon ourselves and our portfolios a certain level of existential risk, and whether we're properly pricing it. That's a fascinating question because it can be a 3% probability of something really, really, really negative happening.

Imagine if you were selling a product and then that product got outlawed by the entire nation.

Yeah.

That's a big problem. You all of a sudden go out of business, right? And I'm taking an intentionally extreme case.

Extreme case, right?

Huh?

An extreme case that happens. I mean, this does happen. There are companies whose products have literally been banned.

Yeah.

It would happen, right? Or imagine if you were investing in a natural-resource company—by the way, don't do that—in an emerging-market country with a not-very-stable government and a not-very-stable political regime that may not necessarily respect a bilateral investment treaty.

Yeah.

And you may own the best copper mine in Country X—I'm making this up—and then you wake up one day and there's a military coup in that country. Now the new leader, in his ultimate wisdom, believes that they should be running that company instead of your company, which is listed on NEO.

It doesn't even have to be a copper mine. I've said—I think I did a rambling on the podcast about things that you say no to—and Chinese companies I've really struggled with. I've never been to China, so the culture and everything that's happening there are difficult for me to understand.

I just keep thinking about Alibaba's Ant Financial. It belonged to Alibaba, and then it did not. How do you account for the 2% risk—or whatever the risk is—of the government just deciding, “Hey, this company, your assets now belong to us”? It has happened in China, and I'm just not sure how to account for that risk.

Artem, what I can tell you is that you are welcome to visit China with me. I'll take you there.

The second kid is on the way, but I've been planning a trip. Maybe after the second kid is a little older and Alicia won't throw me out if I leave her alone for a 10-day trip to China.

As you know, I've been to China many times. I speak some Mandarin. My son speaks massively better Mandarin now, so he fixes my mistakes when I say something incorrectly, which is hilarious. So, yeah, I can take you there.

I appreciate it. It's going to take the Chinese stuff off my no list. Anyway, did you want to have any concluding thoughts here? We're running pretty long. I'll say what we're going to do next, but do you have any concluding thoughts on that or anything else we've talked about today?

Going back to what we discussed briefly, I want to ask you: do you remember when you spoke about borrowing conviction—relying on someone else's conviction—and, as a result, made an investment mistake? You spoke about how there is a relatively small community of, call it, small-cap managers.

How many are out there, do you think, in the U.S.?

It's tough because there are bolt-ons in there—everyone's doing it a little bit differently—but I think there's a core group of 50 people who have similar investment styles. I would say that, for the most part, but not always, they're GARP-y, undervalued, sub-billion-dollar, let's call it.

Then there are people alongside the edge of it who might go much smaller and much less liquid, or might have more of an event-driven angle to what they focus on. But I'd say that if you're talking about the core, everybody reads their letters, everybody talks to each other, and all of us catch up with each other once every 3, 6, or 9 months. Maybe we see each other at the Berkshire meeting or something. But that's what I would say.

Okay, I got a question. The United States of America has about 330 million people living in it—

Roughly.

Do you think that the entire nation of 330 million Americans has only about 50 people—50—who do small-cap investing in the—

No, no, no, no. This was small-cap fund managers running similar styles who are chatting with each other. I think there are many more.

Much more.

What fascinates me is this, and I don't know the answer. Maybe there aren't 50 people like this. Maybe there are another few hundred people with a somewhat similar style, but they're just not talking very much with the 50 that you and I know. I don't know the answer. I'm genuinely curious.

By the way, let me do a plug for myself. If you're a small manager—not small in terms of your size of AUM necessarily, but in terms of the way you invest, like small-cap, micro-cap, or even small- to mid-cap—and we don't know each other, reach out to me on LinkedIn. I'd especially like to meet other people in the space.

As you mentioned, there are people in their mid-30s and early 40s, and I think what's interesting is that there's a new generation of investors who are now in their mid-20s or very early 30s. They're doing what you and I started doing roughly 10 years ago, plus or minus a couple of years.

I know some of them, but I don't know many of them. I would love to get to know them. They're younger, they're viewing the world differently, and they have their own competitive strengths. Their competitive strengths are different from ours.

I'm always very happy to meet those younger investors. Older investors, too, but with older investors, I'm more likely to know them because they've probably been in the business longer, and I probably met them one way or another. But the younger guys and women, I don't necessarily know all of them. I'd like to meet them, learn what they're doing, talk ideas, and so on.

How dare you?

That's what my podcast is for. Reach out to Andrew, then reach out to me too, because I would like to meet other people in the space, especially younger investors.

Ditto. Ditto. I'm always looking for young podcast guests, too. So, cool. Artem, let's wrap it up here. You touched on expert calls a little bit, and we touched on AI a decent bit, but we're running super long.

Expert calls have been a big focus of mine recently, but actually, both have. I'm spending so much time using different AI tools, and I've got so many things I want to talk with you about there, but we're running super long.

Artem Fokin, maybe my favorite guest overall. I can't think of anyone off the top of my head—one of the most popular guests. Thank you so much for coming on. I'm looking forward to the webinar, including a link.

Artem's got one more thing. There's always one more thing with Artem. Go ahead.

Oh, yeah. As you know with me, I think it was the Burford, our 2nd or 3rd podcast, when we were desperately trying to wrap up the podcast. I was like, “Andrew, one more point. One more point.”

Okay, so one more point on this one. I think it's the last one. When you say, “Oh—”

“You're spending more time on expert calls and thinking about the best way of doing them.” You also think about AI a lot.

Yes. Oh—

Okay.

Don't spoil the webinar.

But it's interesting. Because you mentioned that—it's your fault you mentioned that, right? It's large language models. The language is the key word here.

Where else, other than SEC filings, is there so much density in words?

Well, what else? Again, it's a perfect match, in my opinion.

You were the one who told me this. I mean, this alone revolutionized my research project a little bit. I hate saying “revolutionized” because I don't think it did, but it sped it up a lot.

You know what I do now when I want to read an expert call? AlphaSense does have a summary of it, but what I've started doing is, if I follow a company and it's got 5 expert calls, I toss all 5 into an LLM and say, “Summarize them for me.” Then, if I'm still interested in the company and want to learn a little bit more, I go to the individual expert calls and read the summary there. If I'm still interested and really want to dive in, then I go read the expert call itself.

By doing that—again, I give credit to you; you were the one who told me this—when I'm reading it, instead of, you know, when you read new information for the first time, it's actually very hard to process. But if I've already read a summary—in that case, almost twice—I know what's coming. I can really parse the language, and I really understand it. I found my retention and absorption are much better.

So, that is just one way it is such a perfect match. It's been an incredible tool for marrying expert calls and AI together.

I appreciate you giving me the credit. That's true. I did tell you that in terms of how I was using AI in expert calls.

Now, what I would say—and let's conclude there because I don't want to spoil too much about our second conversation—is this: this is the framework, which is very simple, and it's not mine. I learned about this framework from Paul Hilal, who used to be a portfolio manager at Viking, and he did 2 fantastic podcasts: one with Ted Seides on Capital Allocators, and another one, Invest Like the Best, with Patrick O'Shaughnessy.

I believe the Patrick O'Shaughnessy podcast came out first. Both are fantastic. I probably listened to them 3 times each over the last couple of years. Paul Hilal offered a simple but, in my opinion, incredibly effective framework: digging, analyzing, deciding.

What I found when I think about my own uses of AI is where AI fits in this 3-stage framework—digging, analyzing, deciding—and where I can use more of it, where I can use very little of it, or maybe even 0 of it. Let's pause here because that's how I will be kind of training.

Okay, great. This is a great pause. Artem, this is awesome. I just wanted to come on and improve the process with you, and hopefully I've improved my process a little bit, helped you improve your process a little bit, and hopefully our listeners are going to improve their process a little bit from listening to this.

This was awesome. We'll talk soon, buddy, and we'll go from there.

Excellent.

A quick disclaimer: Nothing on this podcast should be considered investment advice. Guests or the hosts may have positions in any of the stocks mentioned during this podcast. Please do your own work and consult a financial adviser. Thanks.