VALUE: After Hours (S07 E02): Jim O’Shaughnessy on AI, ventures, publishing, markets and his new book Two Thoughts

Johnny HopkinsStock ScreenerLeave a Comment

In their latest episode of the VALUE: After Hours Podcast, Tobias Carlisle, Jake Taylor, and Jim O’Shaughnessy discuss:

  • The Science of Simplicity: How Limited Choices Drive Better Decisions
  • Two Thoughts: The Inspiration Behind Jim O’Shaughnessy’s Quote Collection and Its Unexpected Journey to Print
  • Embracing Paradox: Navigating Convergent and Divergent Problems in Investing and Beyond
  • Moore’s Law, Exponential Knowledge Growth, and Solving Tomorrow’s Better Problems
  • Creating Optimistic AI
  • AlphaGo and the Game-Changing Power of AI
  • Optimism in Innovation: Lessons from the Simon-Ehrlich Bet
  • Innovating Across Industries in the Era of Mass Customization
  • From Netfolio to Canvas: How Tax Optimization Became the Unexpected Key to Custom Indexation Success
  • Revolutionizing Publishing: How O’Shaughnessy Ventures’ Infinite Books is Empowering Authors with Technology

You can find out more about the VALUE: After Hours Podcast here – VALUE: After Hours Podcast. You can also listen to the podcast on your favorite podcast platforms here:

Apple Podcasts Logo Apple Podcasts

Breaker Logo Breaker

PodBean Logo PodBean

Overcast Logo Overcast

 Youtube

Pocket Casts Logo Pocket Casts

RadioPublic Logo RadioPublic

Anchor Logo Anchor

Spotify Logo Spotify

Stitcher Logo Stitcher

Google Podcasts Logo Google Podcasts

Transcript

Tobias: This meeting is being livestreamed. That means it’s Value: After Hours. I’m Tobias Carlisle, joined as always by my cohost, Jake Taylor. Our very special guest today is Jim O’Shaughnessy. He’s an investor, author, entrepreneur. Exited entrepreneur? Current entrepreneur?

Jake: The Godfather himself.

Jim: [laughs]

Tobias: [crosstalk] Jim is there anything that you don’t do?

Jim: [laughs] Listen, Toby, most of what– I have a very limited skill stack, but I’m just very lucky that it happens to be really good ones.

Tobias: A very specific set of skills.

Jim: I’m really, really bad at most things. So, yes, there’s an endless list of things that I do very, very poorly. But anyway, yeah, I tend to jump down a lot of rabbit holes.

[crosstalk]

Jake: Pretty sure that the family’s cataloged all those weaknesses for you.

Jim: Oh, in detail, in detail. Now, I’m worried about my grandchildren. My oldest grandson just turned 11 and he started to make fun of me.

Jake: He’s getting wise to it. [laughter]

===

AlphaGo and the Game-Changing Power of AI

Tobias: We were having a very interesting conversation just before we came on, and I almost didn’t want to interrupt that conversation to start the podcast because Jim was discussing something very interesting, but we were– I need to sort of get everybody up to speed to where we were. We were discussing O’Shaughnessy Ventures, OSV. You have your own AI, and your AI is trained on perhaps a different– A similar dataset to the others, but you’ve added some additional stuff in. And we were talking about the way that AI has revolutionized games like Go. And so, do you want to sort of– Maybe we’ll just start from there?

Jim: Sure, sure.

Tobias: What is Go? Just for folks who don’t know it.

Jim: So, Go is one of the oldest games in human history. It is deceptively simple looking. You have black stones, and you have white stones. One player plays black, the other one plays white. It was usually mostly popular in Asia, but became popular worldwide as people discovered the game. If you just look at the numbers, the amount of potential moves in Go dwarfs the potential moves in chess. Because when– Remember when IBM beat Kasparov and everyone was saying, “My God, this is the most complicated game. And now the computer is beating the grand master”? And everyone was like– Go was kind of like, “Hold my beer.” Because the amount, by a staggering level, numbers that are so crazy that you have to write them in notation form, and literally the number of potential moves essentially exceeds the amount of time we’ve been alive in the universe.

Jake: Or electrons in the universe.

Jim: Yeah, exactly. So, it’s a whole different ball game trying to play Go. And so, I remember reading at the time because I was fascinated by that Go players saying, “Yeah, well, chess is kind of for kids.”

Jake: That’s like checkers.

Jim: “Yeah, that’s like checkers. You. It’ll never, ever, no matter how much time you give it, be able to beat a human player.” Well, famous last words. And what we were talking about was the idea that the way that they trained AlphaGo was quite different than they trained Deep Blue, or whatever IBM called it. It was Deep Blue, right?

Jake: I think so, yeah.

Jim: -that beat Kasparov. IBM used brute force, essentially. It fed into Deep Blue every recorded game of chess. And fun fact, there are a lot of really good recorded games of chess because of the aficionados of that game. There are also recorded games going back much longer of Go. But they had an insight. They’re like– So, they originally trained it that way and it did really well. And then, they thought, “What if we just gave it the rules? What if we just gave it the rules of Go? And then, let’s see what happens.” And so, what happened was it played 100 million, 200 million, because numbers like that are meaningless to AI. Games with itself, and got so good that it beat the reigning champion of Go.

Toby, you mentioned that there’s a documentary about that I haven’t seen yet that I’m going to watch. But it was all emergent. It was, what are complex adaptive systems? Emergence comes from below. And so, they had that little bit of insight. I wonder if it’ll be even better if we don’t force human reasoning and way of gameplay on it. And I was telling you guys about this book that I’d finished called Maniac, which is mostly a fictional account of Von Neumann, the genius. They used to call all those guys the Martians because everyone said they’re so smart, they’re aliens-

Jake: Different species.

Jim: -not from here. Right, there are different species, but it’s told in a fictional format. And then this bit about the Go is tagged on at the end of the book, which I found really fascinating.

Jake: I heard an account of– and I’m not 1000% sure it’s true, but that when Kasparov was playing Deep Blue, that at one point it actually timed out and just made a random move, because that was what it was programmed to do. And it was such a counterintuitive move that Kasparov was like, “Wait a second, what’s going on here?” And it threw him for a loop completely. And then, he like mentally couldn’t recover and get his– Because he was like, “Well, why would it do that? It’s making the optimal move. What would that mean for me?” And then now all of a sudden, he just cascaded out and that was what ended up leading to him losing, was actually an accident.

Jim: Well, that’s interesting because in the story of Go– And you know, my own way of looking at investing, you remember the old Pogo cartoon? We’ve met the enemy and it’s us? AlphaGo doesn’t have, or AI doesn’t have emotions, it doesn’t have senses of time. We are so time oriented as a domesticated primate, that like, “Oh, the meeting starts in 10 minutes, we’ve got to get there,” or, “I don’t have time for that.” We talk time all the time. Time is meaningless to an AI. And so, I always enjoy when people go on and on about “But AI hallucinates.” The original hallucinators are we humans.

So, the idea though that you’re dealing kind of with an alien intelligence, that’s what I find fascinating about this because the same thing happened with Go, but it didn’t do it because of a timeout. It did it because it had calculated that there was a 1 in 10,000 chance that this move, if he played the game the way that AI anticipated he would, the human, that he would win. And so, it makes the move and at first, the grandmaster Go player looks at it and smirks and kind of laughs. And then, you can see his entire disposition change. He looks at it again and then you can kind of watch him crumpling as he realizes, “Oh, my God. That actually– This program is going to beat me.” And so, I just think, what a time to be alive, guys.

Tobias: The documentary that I’m thinking of, I forget what it’s called, but Google itself has put it out on YouTube and it shows that the system records two things and records the likelihood that this move is moving you closer to a victory, and it also includes the likelihood that a human player would make this move, and so it progresses just relentlessly. The computer systems just relentlessly drive towards the highest probability result. But then, it says that the chance that a human would make this move was like 1 in 8,000. It was vastly– It was a very unusual move which gave everybody a little bit of pause. And then, the moment that you’re talking about, he gets up and he walks around outside smoking a cigarette. He’s clearly had some very damaging psychological–

Jim: Yeah, but I think that if we can learn to turn off our fear systems, I think fear is still the ruling emotion in our human OS. And if we can stop and think for a minute that there is a way that we can collaborate with this very, very different type of intelligence, that is a win-win for both of us. I think the idea that AI thinks and proceeds in a very different fashion, ability to look into liminal spaces, for example, no human–

I mean it kind of delights me that my first book was Invest like the Best where I said you can simply take a manager, take their portfolios, put them on a data set like Compustat or Value Line and then suck out how they differ from the average stock as a whole and then use that to create in Silico that manager. Well, AI is this on steroids. It doesn’t think the way we think. And I look at that as a really beneficial thing because if you’re going to interact and co-create and collaborate with the AI, you’re going to be able to see a lot of blind spots that we just can’t see normally.

There’s a great quote that goes along the lines, I can’t remember who said it, but the quote is “No matter how smart somebody is, no matter how creative, no matter how insightful, you can never ask them to make a list of things that would never occur to them.” And one of the things that I’m trying to build out with our AI platform at O’Shaughnessy Ventures is that. I want the AI to suggest things to us that we would have never conceived of. And then, we have the option of looking and saying, “Wow, we better actually incorporate this,” or as is not equally likely, but is as is likely, “Eh, I get where it’s coming from, but I don’t think that’s going to actually add to the efficacy of anything we’re trying to do.”

And rather than be afraid, be delighted that– What did Jobs say? Computers were bicycles for the mind. AI is a rocket ship for the mind. And I literally believe that when retrospectively people are thinking and talking about this period 20, 30 or 40 years from now that they’re going to, with a straight face, say that AI was as impactful as Gutenberg.

Tobias: Hmm. I think it’s interesting to look at when Deep Blue played Kasparov, there was some controversy because he said he was evidently– Computer programs had played in what they describe as a closed fashion, which I don’t really know what that means, but he said he was planning to beat this game in a closed fashion. And then, it made a move that wasn’t the move that a typical computer would make. And he says that he thought they had been given– The people who were programming Deep Blue had been given a list of moves and the top move, and they had some very smart chess players standing there watching it. And they had just said, “Don’t make that top move, make the second move.” And then, Kasparov said that sort of some human had interfered, and they hadn’t played this closed system. And it seems funny now to look back on that because that’s I think the late 1990s.

Jake: Yeah. It’s was like in 1998 or something.

Tobias: [crosstalk] The chess systems are so vastly superior to humans now. And there’s an online thing called Stockfish that people can use too. And it just gives you the perfect answer that computationally– What did you describe it before the computational exhaustion? What’s that–? It just plays every single move from this position to the very end.

Jim: Right. That’s brute force.

Tobias: Brute force. Right.

Jim: Yeah, well, but we would underline that the really killer model that AlphaGo did was not a brute– It was not a Brute force model. Essentially, it learned organically and came up with solutions from the bottom up. And what I find interesting about that is it came up with moves that humans would not conceive of, would not make or explicitly we’re told not to. “Never make this move.” And that was what they were trained from when they were kids playing Go. And I think that–

Tobias: I’m just going to say I think it’s because it gets– You don’t want to leave something stranded because it’s a game of territory and you can surround and then capture that little bit of territory. And so, you don’t want to leave anything out by itself. And that’s what it did explicitly. But that ultimately was what led it to beat the pro.

Jim: Yeah. And I just think in every– I think, listen, AI is an omnidirectional technology that will affect every single industry, it will affect society, it will affect relationships. I mean, it’s going to be a game changer in so many, many ways. But rather than get all freaked out about it, I always try to sort of invert the argument and think, but look at all the new things that we are going to be able to do that we couldn’t even dream of doing on our own, even within a large organization. I’m thinking, and some of my advisors at OSV are trying to talk me out of it but I would love to make a long bet– Your guys are familiar with longbets.org.

Jake: Mm-hmm.

Tobias: Yeah.

===

Optimism in Innovation: Lessons from the Simon-Ehrlich Bet

Jim: Yeah. I’d love to make a long bet that essentially the innovation used correctly. And we’d have to figure out how to word this bet that kind of models the Simon-Ehrlich bet. Are you guys familiar with that one?

Tobias: No.

Jim: So, Julian Simon was a rational optimist. And he wrote a couple of books called The Ultimate Resource 1 and 2. He could have maybe had jazzier titles, but the takeaway from his book The Ultimate Resource, which he wrote 40 years ago, I think, was that we human beings were the ultimate resource. And the bet that he had was with one of the gloom and doomers, members of the cult of Malthus-

Jake: [crosstalk]

Jim: where– Club of Rome. Ehrlich. And what Simon did was really interesting. He said to Ehrlich, “You pick the rare minerals. I don’t even care. You get to pick them. And I’m going to bet with you that on an inflation-adjusted basis, all ten of them are lower in price ten years from now than they are today because I believe in human innovation and ability and imagination, and you don’t.” And Simon won. And it’s so funny to me that bet, that whole thing is not widely known by more people. It’s just like the– I always call it the power of pessimist because I look for things to root for because it’s just so easy to find things to root against. It’s so easy to be against. You could be a nitpicker. You can just say, “Oh, that’ll never work.” It’s just too easy to be a pessimist.

That’s why I think when you look for things to root for you get much more committed to steelmanning the argument, to steelmanning the opposition’s argument, to making certain that the case is one where this really is a great thing, but it requires a re-jiggering of the way you think about things.

===

Creating Optimistic AI

Jake: So, just with your AI that you’re building and talking about if it’s basically just all of human thought that’s been written down, which is what it’s training on for the most part, and knowing that humans have a natural proclivity towards pessimism and it sounds smarter, it’s always seems like the smart move. Do you think that AI then has an inherent negative bias in it and that if you might be able to balance that out a little bit by choosing content or maybe taking other content out that is not as conducive to that, like having a more optimistic AI potentially.

Jim: Potentially, I think you have to be really careful because we don’t know what we don’t know and a lot of the negative pessimism that it was trained on. We don’t know much about the liminal spaces that lurk underneath the reason for those pessimistic statements. And AI can look into those liminal spaces. And so, I do think that if you are going to be using an agent architecture with AI, then it probably makes a lot of sense, and we are in fact doing it, to train agents on selected literature or data that we go out of our way to compile.

We were talking before we went live about the literature of the late 1800s through, kind of through maybe the market crash of 1929 when things got a little more glum, but American literature from that era was extraordinarily optimistic and can do. William James kind of kicked it off with his deep look into psychology, etc. And kind of the birth of the philosophy of pragmatism, which is a not uniquely American, but we were the ones who really went all in on that philosophy. So, we’re going to fine tune a model, an agent, that is trained on that literature.

And then, there are other things that you want to do when you are looking for answer that’s quite specific. So, there is a hypnotist, now dead, named Milton Erickson who had an extraordinarily high cure rate. And people were in wonder of this guy. An example, a woman with agoraphobia who’d seen psychiatrists, psychologists, had drug therapies, had every conceivable therapy had not worked on her. And Milton Erickson went to her house, cured her in two sessions. And so, I asked the question, was that something that was just unique that we can’t capture in the person of William Erickson? Or, is that something where we can find a reproducible pattern that we could test? And so that requires training the AI agent on not only all of the books that he ever wrote, all the interviews he ever gave, anything ever written about him, including things attacking him to get to, was this kind of a fluke that can’t be recreated? Or, was there an underlying method that our weak human eyes couldn’t see, but the AI can see?

So, I think that’s an individual example, but you’ll be able to do that for lots of stuff. You’ll be able to synthesize hundreds of thousands of books, scientific research data, etc., etc., and fine tune the model for your specific question. So– [crosstalk]

Jake: Yeah, it’s so interesting to me that all the work you were doing in the 1980s on the quant side of finance is really like totally the same stuff, but just on a more broad scale to what’s happening now. I mean, being able to find things that worked, see if there’s patterns and correlations, and can you replicate it quantitatively in the next population looking forward, I mean, it’s all the same stuff.

Jim: Yeah, it really is. And it’s like—Now, that we can do it in Silico, the applications for it are virtually endless. There’s the medical application where you can recreate me using my DNA to a sufficient degree that you can create a digital twin for me that exists– I mean, what composes me from a medical point of view. And then, you can test things on the digital me that I would never let you test on the real me because they say, “Yeah, in 98 of the 100 instances we gave you this drug, yeah, you died. And ooh, it was a gruesome death.”

Jake: Yeah.

Jim: But in Silico, again, these permutations can be endless. Now you got to apply the same sort of insights to “Well, is that data mined or is that efficacious, truly?” Well, again, because of our ability to do it in Silico, create nearest neighbors to me and you, the answer is, we’ll be able to really robustly test a lot of things that we just simply couldn’t do before. And wow, I just– It’s finally here. I’ve been a journal keeper for 45 years and I have one from when I was 23, and it’s right here. I’ve been dreaming of this stuff since I was 23 years old and I found this and I’m like, “Wow.” I’d kind of forgotten I wrote that. But it was all about Project Microman and the Alpha Machine. And the Alpha Machine. I didn’t use the term AI, but the way I described both Microman, which would be Neuralink, and Alpha Machine, would be AI, it’s like, “Finally.” [laughs]

===

Tobias: Jim, let me give a quick shoutout and then we’ll ask you some questions about Two Thoughts, your new book. Bendigo, Tomball, Texas, what’s up? Jupiter, Florida. Boise. Madeira Island, Portugal, Mississippi. Rockland County. New York. Bellevue, London. Lausanne, Switzerland. Gothenburg, Sweden. Bremerton. Mendocino. Breckenridge. Kennesaw, Georgia, what’s up? Edinburgh. Dubai. Leuven, Belgium, Stuttgart, Germany. Cardiff. Jumped over a few here, sorry. Saskatoon. Helsinki. Riyadh. That’s a good spread, Did I get everybody? Nashville, Mac in Valparaíso. what’s up? Austin, Toronto. Cincinnati. That’s cool, that’s a good spread.

Jim: So, I got to ask, these are all people– This is live, right?

Tobias: It is live. Yeah. That’s folks just dropping where they’re coming in from.

Jim: You just read from where people are coming to us, yeah?

Tobias: That’s it. Bucharest, Romania. Last one in the house. Yeah.

Jim: So, great example of our great reshuffle thesis. Time, space, geography have collapsed and what we are doing right now would have seemed like a magical thing, just somebody from just 20 years ago. And it’s a great example of we now have not only the tools, but we have the ability to build what I call the human colossus. And that’s kind of united minds across the world. And that kind of– talk about– Add that to AI and you’re talking fertile ground for major innovations and breakthroughs coming over the next 10 to 20 years.

Tobias: Are we at the singularity? Have we reached the singularity?

Jim: I don’t think so. I’m not the expert and so I don’t know. My opinion is that a lot of the talk of the singularity of AGI, etc. Is people talk in their book and selling you something. And that’s not to say that it might not ultimately happen that we reach a stage of intelligence that is so advanced that we in a 10-year period create innovations that would have taken us a thousand years in the past. But I do know that if we take all that and set it aside, the tools we have right now that we actually can use are world changing. And so, rather than– I mean it’s kind of like, why would you overhype what is already like an amazing technological innovation? It doesn’t need to be overhyped.

Right now, if all development stuff, all of it, these tools still are incredibly powerful. And Jobs said computers are bicycles for your mind, these are rocket ships for our minds. And so, I don’t worry too much about whether we’ve reached a singularity or not. And then, the question I always ask myself about the singularity is, would we know?

Tobias: [laughs]

Jake: Yeah. I think most especially the version where it was like in 12 hours, it taught itself how to get better, and then it just asymptotically took off and no one knew.

===

Moore’s Law, Exponential Knowledge Growth, and Solving Tomorrow’s Better Problems

Tobias: Well, I think Moore’s Law is kind of interesting because every time we– Not that long ago, Moore’s Law seems to be, which is the shorthand for computing power, basically we find a new technology and then we make that technology better and better and better, we kind of exploit that. And then, it gets to this point where the gains aren’t quite following Moore’s Law. And then somebody finds a new technology.

Jake: That’s cost instead though. Like, that’s what I found interesting is they made it as a cost as well as part of it to kind of keep it on the curve.

Tobias: But I think we’ve sort of– we have managed to find that new technology every single time. And I think that maybe quantum computing is that new technology, I’m not sure. Keep us on Moore’s Law for a very long time into the future. Maybe we go faster than Moore’s Law.

Jim: I think that if you look at the time it took knowledge to double historically, there was a guy, an economist, who did an analysis of this, and he started in 1 A.D., and then he computed how long it took. And I’m sure there are a lot of footnotes that we’d have to look at. But just to give you the general theme, he wanted to see how long does it take knowledge to double. And so, the first one from 1 A.D. took like thousands of years for knowledge to double. And then, far from being linear, it goes exponential. And it’s not just this type of knowledge. It’s all types of knowledge. And the fact that his forecast, and he wrote this, I don’t know, maybe 20 years ago, his forecast is eerily accurate. I’ll have to look for the title, so you can take a look at his work. But we are speeding up exponentially.

And again, I’m a rational optimist, but I’m not Panglossian. I, don’t think that “No. There’s not going to be any trouble ever again.” In fact, quite the opposite. I think that we will continue to face really difficult problems. I just think that there are going to better problems, and they’ll never go away. As we learn more, we’ll learn more about, “Ooh, that’s a problem I didn’t even think about.”

Bucky Fuller used to say that you shouldn’t get mad at people for saying, “I don’t know anything about that,” and he said because they weren’t tuned into it. And he uses, as an example, the guy who invented the microscope. Prior to the invention of the microscope, nobody was tuned in to the microscopic world. We didn’t know about it. We had no way to know about it. But the minute he peered through that little thing, a new big bang and humans were like, “Well, that’s certainly interesting.”

And so, I think we’re going to see similar things happening on this curve as well. We’re going to get tuned in through the ongoing innovation, but with that comes not only benefits, but problems. But they’re going to better problems. And so, it’s one of the reasons I worry a lot about the idea of the so-called precautionary principle. We of course should take precautions. We of course should have plans for if things go wrong, but we should never build them into a place where we actively try to discourage new knowledge, innovation, etc., which leads to, say, stasis. And stasis is death, movement is life.

And so, we definitely will have a bunch of new problems, but there’ll better problems. We’ll solve those, creating even better problems and more intractable-type problems. But you don’t ever want to be Pollyannish about this and just say, “Oh, it’s like everything’s going to be great and we have nothing to worry about.” Not true.

===

Embracing Paradox: Navigating Convergent and Divergent Problems in Investing and Beyond

Tobias: JT, top of the hour. Do you want to do your veggies?

Jake: Yeah. And couldn’t be a better transition, to be honest, as you might find out. So, as I was reading Jim’s book, Two Thoughts, which we’ll talk more about after I get through this, I was struck by how often these two selected quotes, which is what makes up the book. All these really super smart people and kind of the two smartest things that they ever said in their lives. But those two things were often in conflict or paradox with each other. And somehow, yet both were strikingly true, which kind of defies some internal logic for me at least. And I think I have a possible answer that was inspired by a recent read which is one of my friend, Nima, in his 2024 investor letter, he runs this firm called Rumi.

He talked about this and I’m going to see if I can weave this all together for us. So, I think the distinction lies in convergent versus divergent problems. I don’t know if you’ve seen this mental model before, but to step on the punchline a little bit. Convergent problems have a clear single solution that can be reached through logical analysis. While a divergent problem is usually related to human values and complex systems, it doesn’t have a single definitive answer, and it requires a more holistic approach, and you have to consider multiple perspectives and nuance. So, this breakdown of that terminology originally comes from a statistician and economist, E. F. Schumacher. And he actually led a pretty interesting life.

He was born in Bonn, Germany in 1911 and he witnessed the rise of the Nazi regime, and he was pretty disillusioned by all the fascism. He fled for Britain, and it didn’t really go much better for him once he got there. During World War II, he was interned as an enemy alien and it’s an experience that shaped a lot of his views on power and resilience and human dignity. But basically, the British didn’t trust that he wasn’t a German spy. But John Maynard Keynes recognized Schumacher’s acumen in economics, and he broke him out of the internment camp basically and took him on as a protege at Oxford.

And fun fact, Schumacher’s brother-in-law was the renowned physicist, Werner Heisenberg. And yes, that was an actual guy and not just a chemistry teacher dealing drugs in New Mexico.

Jim: [laughs]

Jake: So, after the war, Schumacher, he helped rebuild the UK and Germany and was a leader of the British National Coal Board, which was actually a pretty big responsibility at the time because coal was a very dominant resource still of energy, still is today, but– And his economic thoughts and writing were very influential post war period. In 1973, he published a book called Small Is Beautiful: A Study of Economics as if People Mattered, which is a great title and it was well received, sold well. And one of his main arguments in the book was that we can’t consider the problem of technological production solved if it requires us to recklessly erode our finite natural capital and deprive future generations of its benefits.

So, this is a little bit of the Ehrlich-Simon arguments, but it also actually to me sounds a lot like the position that Charlie Munger took. He was a big fan of actually saving the hydrocarbons for future use because they’re too valuable as chemical feedstocks for our grandkids than to be used to go get groceries, basically. Anyway, Schumacher, he understood this paradox of growth and sustainability long before it was a mainstream concern.

And so, just to summarize a little bit here and drive it home, convergent problems, they’re typically technical issues with well-defined parameters like designing a bicycle or solving a Sudoku puzzle where you can gradually converge towards an optimal outcome. It just takes a lot of elbow grease. And I think AI is probably like really good for that. Think of maybe finding the most efficient way to produce solar panels or something like that, there’s a clear, measurable solution.

Divergent problems are often philosophical, ethical questions. There’s no single correct answer. Maybe think about what’s the best approach to education. Do you instill discipline and kind of cram it into the kid’s head? Or, do you allow free learning? How much? What’s the balance? Or, maybe deciding between economic growth and environmental sustainability. And so, Schumacher’s big insight was that if you show up to a divergent problem with convergent solutions, you’re likely to run into a lot of issues.

And this actually rhymes quite a bit with Jim’s friend, Rory Sutherland’s book in Alchemy, where he talks about engineering solutions versus human or psychology solutions and how different those can be.

So, let’s get this all back to investing and see if we can try to stick this landing here. My hunch is that the world would be a lot easier in investing if it was just purely convergent problems like plug in the right numbers, carry the one, you get this unequivocally correct answer. And because it would be nice if it was easier, there’s a lot of wishful thinking done by people to view it as a convergent problem. If you just work hard enough, there’s always the right answer to distill down. But there are some parts of finance are probably like that, like maybe you know, understanding the payment schedule of a mortgage or something. So, it’s not all divergent problems, but most of the time I think it tends to be divergent problems.

Let’s take the balance of trust versus skepticism. If you trust everything, you’re likely to get taken advantage of. But if you’re skeptical about everything, you’re likely to miss big opportunities. So, as a thought experiment, I’ve often wondered if you took everyone who has ever owned Berkshire Hathaway over the last 50 years, 60 years, and you lined them up in the order of multiple of return on the invested capital and we had just everyone stand in a big line and you force rank them, my hunch is that you would actually probably end up with the inverse row of that population on how they would score on an accounting exam.

So, the people who made life-changing money, they just trusted Buffett to figure it out. And if they knew a bunch of accounting or valuation, they may have talked themselves out and sold Berkshire along the way when it felt like it was too rich. I mean, it’s just human nature. But if you did that exact same exercise with Bernie Madoff or Enron, you’d probably get the inverse where if you could understand a simple cash flow statement, you could plainly see this wasn’t sustainable. So, there’s always these paradoxes, right?

And another paradox might be patience versus urgency. Obviously, patience is crucial for long term success. We’re finding that out in the value world, how much patience you actually need. But you should be waiting for that fat pitch, right? But if you’re too patient, it might look like continually sucking your thumb, not pouncing aggressively, and then you’re letting your few precious trips to the pie counter slip away. Those are all Mungerisms, by the way. So, urgency can drive action, but too much of it leads to impulsive decision making. So, how do you find that right balance?

And even transparency kind of has its own paradox. Investors crave openness from their CEOs, LPs want transparency into their GP’s processes. How is the sausage actually made? But maybe over explaining in volatile times leads to premature decisions. Are you making the right investment or are you making one that you can actually explain to your constituents? And maybe in the CEO case, they don’t want to be open because that’d be giving away some material trade advantage.

So, there’s all these parts of the investment process. Some require intuition that never fits necessarily into words. So, it can be hard to explain to someone else to get their buy-in and so you only choose things that you know that you can explain, which might actually be quite suboptimal. It’s like being in love, like when you know, you just know.

Circle of competence also, I think, has its own paradox. If you only stay firmly within your circle of competence, you’re likely not to expand it very much if you’re not pushing on the edges, which makes you intellectually brittle. I think we all agree the world’s always changing. And any organization, including your collection of synapses in your head, it risks being disrupted if its pace of change internally is slower than the pace of change outside of the world. There’s The Half-Life of Facts, great book by Sam Arbesman of that title. Many mental models have a shelf life to them.

And then conversely, it’s super easy to push against that circle and then end up tumbling outside of it and you think you know what you’re doing, but the market’s going to give you a very expensive reminder that you were fooling yourself and you really didn’t know what you were doing.

So, I think the takeaway of all of this is because of the divergent nature of many problems, you have to be comfortable with some paradox. You have to embrace paradox, and that’s the title of this segment. And at a minimum, when you’re faced with a problem, you should pause to ask yourself, “Is this likely a convergent or divergent situation?” And then, that should dramatically change which toolkit you decide to pull out.

===

Innovating Across Industries in the Era of Mass Customization

Jim: Yeah, definitely. I think being comfortable with paradox is kind of table stakes. If you’re trying to develop models for making better predictions, the truth ought to be predictive in my mind. And what did Fitzgerald say? The rank of a first-rate mind was the ability to hold two diametrically opposed ideas in their mind at the same time.

Jake: Without going insane.

Jim: Yeah, without going insane. And as far as challenging circle of competence, well, I’m a bull in a China shop by that definition because if you look at all of the verticals that we have at O’Shaughnessy Ventures, I have no experience running a publishing company or a movie company or a media company. And yet, I take it as an advantage, not a disadvantage, that I am coming with first fresh eyes to things. Now, I love all of those things and that’s why we are pursuing them. I love books and I think that publishing is stuck in the early 1920s. The unfortunate part here is it’s the 2025s. And so, to be using 100-year-old models doesn’t seem to make a lot of sense.

Filmmaking, Hollywood is run like a medieval guild system. And that really doesn’t make a lot of sense to me. And the idea of moving– We always try to think about what’s the end state. If you can see what the end state of something is, or if you can speculate with a pretty high probability of being right about that end state, well, then it makes it a hell of a lot easier to figure out what kind of path you’re going to take there.

And I think one of the end states that I think has a very high probability of occurring is we are seeing the end of mass production, and we are seeing now mass customization. And in mass customization, power laws aren’t going anywhere. But in the old days, to win a power law, let’s say media game, you had to come up with an I Love Lucy, a M*A*S*H, both of which– Your average I Love lucy show had 60 plus million viewers. The finale of M*A*S*H was more than 100 million viewers. And that’s not really easy. So, big power law, you design for that wonderful 68.5% in the middle of the bell curve, that’s gone. And we’re living under a Mandelbrot distribution, but with a very peaky middle. It’s a chaotic distribution pattern and very long tails.

And I submit that there is gold in our tails and that you can have all sorts of power laws happening, but just mini power laws. We just have created tens of thousands of additional hills to conquer. Will it be doing numbers like I Love Lucy or M*A*S*H? No, not at all. But the economics of those conquering those smaller hills are incredibly compelling and attractive.

===

From Netfolio to Canvas: How Tax Optimization Became the Unexpected Key to Custom Indexation Success

Jake: Was that the insight that led to custom indexation at OSAM?

Jim: Not really. It actually– The insight grew out of my initial attempt at that, which was called Netfolio, which I did in 1999. And at the time, I was thinking, “Well, if everything is going to be digital, every cost is dropping to zero, I should be able to customize everything for everybody.” And the tech wasn’t quite there yet. But when we spun out of Bear Stearns into OSAM, right into the great financial crisis, I suggested to our team that we probably wouldn’t be selling another long-only portfolio for a little while. So, I said, “Let’s build the system. Let’s build the system that trades the way that we uniquely trade,” which just happened to work really well with customization. But then, it was my son who came in and said– And he used a completely different metaphor or analogy. He was like, “You know AWS? We should do what they did.”

Jake: Take a cost center and turn it into a revenue center.

Jim: Yeah. And I’m like, “Netfolio.” And he goes, “Yeah, but 2.0, 3.0, because the tech actually works.” But that got us going on a game that I used to play imagining end states. And so, the game I would play is, I love this book by Peter Drucker, Adventures of a Bystander. It’s a story of all the interesting people he’s met in his life. And a lot of it is set in the 1920s and 1930s and he worked for Merchant Bank in the City of London at that time. And I got the idea—

You know what’s going to happen. Technology is just going to make it so that everyone can be treated like a super rich lord was treated back in 1920s London. Of course, easily, you know, they had a car and driver. You’ve got all the various services that are compounding. You would say Uber originally, but now you’ve got the self-drivers and all of that. They had multiple homes. So, you have Airbnb, but that’s morphing too. And then, Canvas, which was what Franklin Templeton was so interested in acquiring from us was a result of that.

But I definitely think if you believe the end state is mass customization, then you should think within your industry, “What does that look like for me?” And that’s what we did at OSAM. And we’re like, “Well, what it looks like for us is we’ll be able to customize portfolios.” But we were wrong. I want to really underline this because it’s important for people listening to us to understand this.

One of the landmines that we step on repeatedly, and I include myself here, is we think we know everything there is to know about that particular product because we invented, it’s our idea etc., and we don’t at all. That’s why I love markets. Markets are really good at disabusing you. If you think that, “This is right,” and then you put it to the test of the market is very quickly going to say yes or no, you’re right or wrong.

And so, we had all of these features in Canvas that we thought were going to be the major killer feature. “This is why everyone is going to convert to Canvas. This is why they’re going to do it.” We were so wrong. And so, what we did was we limited the amount of advisors we would work with for the first year, and it came back that the number one reason which was up here and all of the other reasons that people cited were down here. What was the number one reason? Tax optimization.

Jake: Hmm. Yeah.

Jim: And we thought like we were– That sounds so boring to us. And yet, that was the conversion. That was the bullseye. If you’re trying to get somebody to understand it, give them that story, and then your conversions for registered advisors just soar.

===

Two Thoughts: The Inspiration Behind Jim O’Shaughnessy’s Quote Collection and Its Unexpected Journey to Print

Tobias: Jim, your new book, which I have a copy of, thanks to you. Thank you very much. I left it wrapped under the tree and then I opened it on Christmas morning, had a look through then.

Jake: Ah. Gift from Jim.

Tobias: It was really nice. Beautiful-looking book called Two Thoughts. And I love the idea, but what’s the–? Maybe you can explain to folks what the process is there.

Jim: Sure. So, I collected quotes all my life because I think they’re really good. Maxims are good and interesting because really what they are, are compressions. By that, I mean maxims exist because time and time again, they’ve been tested, tested, tested, so they get compressed down to this sometimes pithy, sometimes profound idea or quote. And so, since I was a kid, I collected quotes that I liked. And one day, I just was sitting around and thought I should put something up on Twitter. I’d just been reading the list of quotes that I’d kept from Bucky Fuller. And I’m like, I’ll put up two of Bucky Fuller’s quotes and just see what happens. And not much happened. [laughs]

And so, I thought, “You know what? I’m just going to keep doing this.” And so, I started putting up “Two thoughts from,” And I just literally went to my notebooks and found the quotes that I had collected over the years. And then, a really interesting thing happened. People started DM’ing me, or if they knew me, texting me and saying, “Hey, I really love this two thoughts thing you’re doing. Have you ever thought of doing a newsletter around it?” I’m like, “No.” So, I talked to the team and they’re like, “Yeah, we should.” And then, another member of the team is like, “I don’t know, we should do a book of these because it’ll just be a lot of fun. People will really enjoy it.”

And the beautiful part about the book, which I have somewhere around here, here it is, is you don’t have to read it through. I think maybe– Again, listen to the market. What we’re learning is that what people really like doing with the book is just picking it up, sometimes people do it at the beginning of their day and kind of opening it randomly and looking at the quotes and then pondering them and then setting it back down. So, what it really does, I think, I hope at least, is that it just serves as kind of that spark for your kind of first thing in the morning, giving you something that you might not have been thinking about at all or ever thought about. And then, reading the quote and thinking, “Huh, I’d never really considered that,” and then going your own way with it.

Tobias: What two thoughts in particular? Why two thoughts?

Jim: If I give you the honest answer, you’re going to go, “Come on.”

[laughter]

Jim: The honest answer was that just seemed like a good title to me at the time. That’s the honest to God true answer. It was not A/B tested. It wasn’t algorithmically derived like you might suspect I would do. It was literally me. And maybe it was because I had highlighted in my own notebook, I had highlighted two of Bucky’s quotes, the two that I used. And I’m just like, “I’m going to put those two up.” And if you think about it, I could bullshit you and say it’s all about what Jake was talking about earlier-

Tobias: Paradoxes.

Jim: -it’s all about- yeah, paradoxes. And this is so different than this, but I would be lying. The truth is, I had two highlighted, and I’m like, “Okay, two thoughts from our R. Buckminster Fuller.” And also–

Jake: It’s quite hard to say three intelligent things in a lifetime, Toby.

===

The Science of Simplicity: How Limited Choices Drive Better Decisions

Jim: [laughs] But also, there is that science. We learned this early on with my first company, O’Shaughnessy Capital Management, that we used to have brochures, and each one of our portfolio strategies had its own fact sheet. And what we started doing was– We had like ten different strategies from very conservative to very aggressive. And so, what we started thinking was, “We’re going to show all the cool strategies we’ve got,” and no one bought. And then, I started doing the research on it, and what I found was that’s the worst way to try to get anyone to make a decision. The best way is to offer them at most three choices, but two is really optimal. And it’s kind of like, here’s this one or that one. Your conversion rate soars. So, maybe if we want to make up mythology here, that fact was in the back of my mind, Toby, when I decided to make it the Two Thoughts. But honestly, it’s because those were the two I had highlighted. Sorry to disappoint.

Tobias: I heard a story about a shoe salesman who was very successful the other day, and the story went–

Jake: Al Bundy?

Tobias: Someone asked him how he was so successful selling shoes, and he said, “I used to bring out two pairs of shoes, and then—” It was a lady shoe store, and she would try on two pairs of shoes, and she’d say, ‘I’d like to try on a third pair of shoes.’ And he’d say, ‘Which one would you like me to take back?’ And so, then she’d have to say, ‘Oh, I don’t like this one.’” So, there’d only ever be two choices. So even though he’d show them the third pair of shoes, and evidently that made people more likely to make a decision.

Jim: Yeah, and I’m familiar with that. I read that too. And it’s really interesting because it is not intuitive. Intuitive, I think, is you sort of show them how much we can help them, show how many different strategies we have. And then, you’ve got to remind yourself that you and only you are going to be so psyched by the fact that you come up with 12 different strategies. But wait, there’s more. And actually, that isn’t what people are interested in at all. They’re interested in you listening to them closely, understanding their particular problem, and then offering them a solution.

That’s why I love Canvas so much. I love it because that’s what it does. If you really, really feel passionately about some form of social investing, what we– I don’t know if they’re still doing it this way because I’m no longer associated, we sold the company. But we would try to show, “Okay, if you want to take oil and gas out or cigarettes out, your performance is going to decline by this number of basis points over any rolling 10-year period. “But if they really still felt strongly about it, we were able to flexibly address things that you just couldn’t do in any other kind of structured format. And yet, what was the killer app notion here? The killer app was something that we probably, had we spent even a few minutes just kind of thinking about it simply.

Jake: Yeah.

Jim: Tax management.

Jake: Yeah, it’s the cup holders in the car. [laughs]

Jim: Exactly, the cup holders in the car. I love that, I love that metaphor.

Tobias: What are you–

Jake: I have to give you quite a bit of credit because I, self-delusionally fancy myself as a reasonably wide reader and I also am a collector of quotes because I agree, there’s just something amazing about a pithy quote that conveys a lifetime of learning in ten words. But I have to say, probably 75% of the quotes I’d never seen or read before, which is quite the delight.

Jim: That’s great. I’m delighted. I’m happy to hear that.

===

Revolutionizing Publishing: How O’Shaughnessy Ventures’ Infinite Books is Empowering Authors with Technology

Tobias: Jim, what are you most excited about in O’Shaughnessy Ventures right now? We’ve just a few minutes left to go.

Jim: Oh, wow. Well, again, in keeping with my theme of listening to the market, we are getting amazing feedback on our publishing unit, Infinite Books. What we’re trying to do is just take all these tools that we have access to and use them to the benefit of our authors, which ultimately benefits our readers, which ultimately– Just it’s a really wonderful chain. And it just keeps getting better and better. Boy, the market lets you know really fast. So, when we just started talking to people about infinite Books, take a wild guess how many manuscripts we got that were legitimately good? Let’s not say the spammers, but just kind of mentioning it, what would you guess that we got in the first couple of months?

Tobias: Dozen.

Jim: You say a dozen? Jake, what do you think?

Jake: I’d go way over that. I’d say 200.

Jim: You’re much closer. It was like 78 in a month. And of the 78, only a few were not good. Like, of the 78, wow, we were really impressed. And we just think it’s one of those things where just everything’s right about, and this is one of the reasons why I’m so excited about it, it’s a business that’s been around forever and that nevertheless got stuck in old ways of doing things that don’t scale under the new technology that we have today. A lot of it, like your cup holders comment, a lot of it is we’re using technology to address all of those mundane, horrible, long wait times. If you’ve ever written a book and you send in the manuscript, how long does it take them for that manuscript to get back to you as an author? It’s ridiculous the time it takes. We can have it back to you the next day.

And so, the ability to address these things that have nothing to do with the author’s creativity or what the author is writing about, but just getting rid of all of these horrible things that people have to deal with, people just get really excited by that. And so, building it out as an AI first publishing company, but always at the discretion of our author. Does the author want us to turn it all the way on or all the way off in terms of marketing, in terms of press, in terms of all of that things? We have those capabilities built in day one. But if the author’s really old school and doesn’t want any of it, he or she is still going to love the fact that we can get that manuscript back to them in 24 hours. It’s a whole new world for authors. Just that simple thing. And so much, much more to come there.

===

Tobias: Jim, that’s all we’ve got time for today. But if folks want to follow along with what you’re doing or get in contact, what’s the best way of doing that?

Jake: Send you their manuscripts? [laughs]

Tobias: Or send in their manuscript.

Jim: Yeah. The easiest way is to follow me on Twitter, I’m @JPOShaughnessy. Or to check our website out which is OSV, Oscar, Sam, Victor dot LLC, Larry, Larry, Victor. osv.llc. And you’ll get a really a good view of what we’re excited about now, what we’re working on. And you will find a place where if you do want to suggest submit a manuscript, you will easily be able to find that. We are breaking out Infinite Books as its own URL. So, if you go to infinitebooks.com right now, what you’ll find is a pre order form for this book, but ultimately what you’ll find there is the ability to submit a manuscript, ask questions, to do all that. And obviously, we have a lot more in the pipeline than this book. We’re very excited about some optimistic sci-fi that we’re coming out with. The first title coming out is called White Mirror and it is beautifully, beautifully written and so just super psyched by all of it.

Tobias: Well, congratulations on all of it, Jim. Congratulations on a great book too.

Jim: Thank you.

Tobias: JT, any final words?

Jake: No, just a pleasure having Jim on and always appreciate any time we get to spend together.

Jim: Always great to see you both. Thanks for having me on and can’t wait for the next time and you can point out all the things I got wrong.

[laughter]

Tobias: Thanks, Jim. Folks, we’ll be back same time–

For all the latest news and podcasts, join our free newsletter here.

FREE Stock Screener

Don’t forget to check out our FREE Large Cap 1000 – Stock Screener, here at The Acquirer’s Multiple:

unlimited

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.