Invest Like the Best with Patrick O'Shaughnessy - Modest Proposal - AI Commoditization and Capital Dynamics - [Invest Like the Best, EP.380]

Starting point is 00:00:02 Hello and welcome, everyone. I'm Patrick O'Shaughnessy, and this is Invest Like the Best. This show is an open-ended exploration of markets, ideas, stories, and strategies that will help you better invest both your time and your money. Invest Like the Best is part of the Colossus family of podcasts, and you can access all our podcasts, including edited transcripts, show notes, and other resources to keep learning at join colossus.com. Patrick O'Shaughnessy is the CEO of Positive Sum. All opinions expressed by Patrick and podcast guests are solely their own opinions and do not reflect the opinion of positive sum. This podcast is for informational purposes only and should not be relied upon as a basis for investment decisions. Clients of positive sum may maintain positions in the securities discussed in this podcast. To learn more, visit psum.vc. My guest today is Modest

Starting point is 00:00:58 Proposal, joining me for our third conversation and the first in a few years. Modest is anonymous online, but one of the most thoughtful investors that I know overseeing a large pool of capital in public and private markets. He offers insight into many different corners of today's landscape, covering AI's frontier models versus open source models, overcapacity issues in transportation in our post-COVID world, and the potential economic impact of GLP1 drugs and more. I always love hearing what modest thinks about markets. Please enjoy our latest conversation. Maybe a fun place to start would be to teach us about what mid-2000s commodity markets can teach us about today's market. Yeah. So something that's been on my mind throughout COVID and really accentuated by what's

Starting point is 00:01:47 transpiring today with AI is this idea of what happens when surging demand meets an elastic supply. And the mid-2000s commodities just did a wonderful example of that, which is, you know, China had historically been out of the global economy for a long time, joined WTO, joined the global trading environment. And within five, six years, went from consuming very small amount of global traded commodities to a very large amount. And so the way this manifested in markets was one, commodity prices went up a lot. And if you remember the backdrop, which is late 90s tech bubble, emerging markets,

Starting point is 00:02:30 97, 98, total blowup. And so strong dollar, all the capital coming into the U.S., commodity prices crushed. Now all of a sudden, tech bubble collapses, 9-11 in the U.S., big focus on international markets, China's growing fast, and commodity prices are soaring. So money's flowing into emerging market equity indexes, buying miners, buying banks, buying all sorts of this stuff. And the, idea was people were putting multiples on the way I think of it in microeconomic terms was temporary surplus. And when you put a multiple on something, you're implicitly capitalizing a long duration of cash flows. Capitalism in this case did its job, which was people went around the world and said, oh, look, there's some mines or there's some materials. Let's go build mines.

Starting point is 00:03:23 And magically, over the course of three, four, five years, a whole lot of supply came online. We went through the JFC and still in the late 2000s, early 2010s, you still had China consuming an enormous amount of commodities. So prices stayed somewhat stable. But as China's growth slowed and supply continued to increase by 2014, the auto commodity collapse. And if you look at any of the indexes, oil is a little different because it sometimes has some geopolitical issues with it as well. But I mean, if you just look where oil was, it was 140 and now it's 80, you look at commodity indexes still down. And that's not surprising because the supply is permanent. It came online. And the last time I was on, we talked about the capital cycle. And this is

Starting point is 00:04:14 the classic example of it. But the willingness of markets to continue, continually put multiples on this temporary surplus is a constant source of confusion for me if you even go into parts of COVID. Soaring demand for everything digital in 2020 when the real world shut down, that probably wasn't going to persist unless the real world was shut down permanently. And that did not happen. So you're seeing it today in a number of places. I mean, the most obvious place where surging demand is meeting in elastic supply is chips, where as fast as TSM can make chips, they are being sold. The other places you're seeing it are in power generation, where we were talking before we started,

Starting point is 00:05:05 20 years, power consumption in the United States is flat. I was reading a report last week that said they expect something on the order of two and a half percent a year growth over the next seven, eight years. And that summed up to over 30% aggregate growth in power consumption in the U.S. And if you think about it, my gut is over the next two, three years, they're probably right in the growth rate of power demand because these things happen quick. And jokingly say, the idea that we're smart enough to build AI and AGII but not figure out how to power it and not figure out how to power efficiently, it strikes me as a curious combination.

Starting point is 00:05:45 we know that DeepMind in the 2010s was able to help Google reduce power consumption and traditional data centers. And this same report actually showed that data center power consumption from 2015 to 2019 was flat. We've got a step function change here. We've got a new technology introduced. There's a whole lot going on driving power consumption. Betting against human ingenuity that we will figure out a way to be more efficient. And also, candidly, if that growth comes to pass, there's going to be investment in the grid, which is going to raise prices, which is going to expose end users to price increases, which is also going to alter behavior.

Starting point is 00:06:30 It's the idea of putting multiples on temporary surplus generated from surging demand meeting inelastic supply. That's a thing that you just have to accept is reality. but if you're thinking over some duration, it seems unlikely that surplus will persist in the way it first emerges. What do you think the very smartest counterargument would be that maybe it's not temporary or something? What would you have to learn that would get you, I'll pick on an example since it's an obvious one, get you super excited about buying more NVIDIA stock today, which is obviously a prime beneficiary of this entire wave, the prime beneficiary and maybe the best single company example, again, my word's not yours. What would you have to hear to be like, oh, God, I need to buy, put 10% of my portfolio in

Starting point is 00:07:22 NVIDIA? I think you have to differentiate between, are you putting 10% of your portfolio in NVIDIA because you think it's going to be higher in 12 months or because you think it's going to sustainably generate the surplus required to justify whatever the price. And look, ultimately the job of a money manager is to make money. So the first one is probably more important. And candidly, what it is is you read something like the piece that was just put out last week 160 pages on why AGI is coming. Leave aside whether or not you believe the conclusion of that paper.

Starting point is 00:08:04 The point is that person probably has some insight into the spending intentions of the frontier research labs. Those labs seemingly are going to be spending bigger sums than I think most people realize or expect because they fervently believe they're on the cost of building AGI. That means that Nvidia is going to sell a lot of GPUs and probably more than is in consensus today. So I think what you need to believe is that these labs, whether or not they are right, believe that spending more and taking compute intensity up orders of magnitudes as they believe will get them to the promised land that they're going to do that. I think the second question, which is a much more academic question of what do you have

Starting point is 00:08:56 to believe to think that Nvidia generates the cash from now until eternity to justify the price. I think that's a different question, which is, do you think it's viable for a single company to be the choke point for five other of the largest companies in the history of capitalism? Do you think that four guys in a garage in California aren't currently thinking of ways to more efficiently run pieces of the compute that is currently only efficiently done via GPUs? was talking to one of our friends this week who told me about some folks who had figured out a way to strip out functionality and run it on a CPU. That's interesting. And whether or not that scales, who knows? But kind of have to believe that this enormous surplus is not the prime target for

Starting point is 00:09:52 capitalism right now, where the hyperscalers, where every chip designer, where every smart kid in a garage, aiming at this entity generating, it's going to be $200 billion in sales at an 80% gross market, you know, some crazy number, generally that will draw a lot of attention. But that's in the future. If you ask, is the stock going to go up over the next 12 months? Probably what you need to believe is the frontier labs want to spend in a way to grow their compute intensity, 100x, 1,000x, 10,000 X. And that seems to be the trajectory that they are on. How are you processing as a predominantly public market investor where the size of the companies that you're investing in is typically quite large, certainly relative to private markets?

Starting point is 00:10:41 How are you processing the commercialization of AI so far? There's a limited number of, we'll call them apps that have real revenue traction, but it's still on the scale that probably doesn't matter to you. They're not public businesses. But this is really important, obviously. It's something everyone's reckoning with. We've talked about sustaining versus disruptive innovation. before a lot, you and I, how are you processing this whole thing? That's a big question, but

Starting point is 00:11:06 seems like the question. Look, I think there are a number of foundational is a bad word to use in this case, but there are a number of very fundamental questions that I think one has to ask. This one is on the economics of the here and now, which is where is the economic activity coming to support the type of spend that is being undertaken? I think the revenue side is obviously more visible, but I think you also have to tally up the cost side as well, the cost avoidance side, aggregate the economic surplus that would justify the investment. On the revenue side, yeah, you can listen to what Microsoft says about what the uplift is within Azure, and AWS doesn't quantify it, but there are ways to triangulate what you think they might be

Starting point is 00:11:54 generating. And then you can add up all the cats and dogs software applications. And, You probably get to mid-single-digit billion. What you can't really see is the cost avoidance. The Klarna story is probably apocryphal, but it's certainly the case that there are some productivity and efficiency gains on the cost side as well right now. So you add that up, and you probably don't get a number that justifies the $200 billion cap-x spend, but I don't know that's particularly unusual for a infrastructure platform shift. If you go back to the late 90s, someone actually took a deep breath and said, hey, everyone,

Starting point is 00:12:37 we're putting $120 billion in fiber optics into the ground. What is the profit justifying this? Everybody would have laughed. So I don't think this is unusual. I think you're looking right now for examples that can be extrapolated more broadly. So particularly on the cost side, where are the productivity gains? Are these going to broadly bred throughout the economy? What's the potential aggregate sum of that? And then on the revenue side, if it is truly sustaining, then every incumbent slaps a co-pilot into their product, then they charge you 20 bucks more. And this is actually a really boring technological shift. I think what's far more interesting is if there are architectural shifts in products and markets, and it's just so early to know

Starting point is 00:13:27 people I talk to and trust say that is happening in software, but those are going to be tiny companies today, and none of that will be relevant in the public market world yet. Can you describe what that means, either using a current example or a historical example of an architecture shift? is what that feels like and looks like and why it's impactful? I think the question in software is on-prem to SaaS was a pretty disruptive architectural shift. And I think the question today is, are there AI-native companies that are going to utilize a modern architecture where data is queriable real-time by these models?

Starting point is 00:14:12 It's amazing how fast the world moves. SAS is the best business model in the world, three years. years ago, and now everyone's wondering, hey, how come all these things are growing 10, 11, 12 percent, and they don't have any margins? So the fact that we're even discussing whether or not there's an architectural shift coming to its ass is insane, but it just shows how fast that world moves. That's one example. I think on the consumer facing side, there's this real question of, does a horizontal search hype product remain the gatekeeper downstream? Does that get verticalized with specialized agents, things like that. But it's all so early that even if that were to

Starting point is 00:14:52 happen, I just don't think we would have any inkling of it yet. If you think about the incredible dominance of U.S. technology companies now for 15 years, it's been a long, long stretch. Probably when you and I first met, there was still that nice chart that showed like the relative performance of the S&P versus EFA. And it was this nice sine wave that went back and forth. And it was sort of widely accepted. Oh, it's sort of cyclical. It goes one way, then it goes the other way. And then that line has just gone straight up into the right in the U.S.'s favor ever since. Long time now. How do you process that chart today? It's funny. I was looking at something yesterday. I was looking at the rolling 10-year returns of the S&P 500 at market cap weighted versus equal-weighted.

Starting point is 00:15:32 And the equal-weighted was actually ahead on a rolling 10-year basis consistently from basically 2000 to 2018, right around the time we met. That was when the market cap weighted recaptured its glory, and now it's on a 10-year basis running like 2.8% or something a year. Remember, that's per year ahead. If you go to E-Fee, if you go to AQUI, if you go to AQUI-X-U-S, it's just insane. But look, these companies are, yes, their multiples are higher. Apple and Microsoft trade at 30 times. I think in video trades in the low 30s. Facebook and Google trade in the low 20s. They're elevated versus the 75-year history at mid-high teens, but that's not something saying, wow, people have really lost their minds here. It's more a reflection of the fact that their earnings power over that

Starting point is 00:16:27 entire period of time has just grown. If you're a software investor, sometimes you have to ask yourself, if my opportunity cost is Microsoft growing mid-teens, what am I doing here? Because that entity has proven over the last 10 years, the ability to compound its earnings power at mid-teens. Apple is upstream of all economic activity in the digital world for all intents and purposes. We found out how much Google pays to be there so that they can insert themselves one level below Apple into the downstream activity. These are just incredible businesses. So I think it's very tempting as many of us were brought up as mean reversionists to say, well, this can't go on forever. It's going to change. It is true. It will end someday. But people have been calling for this end for years and years.

Starting point is 00:17:18 And I think the companies are in no worse shape today than they were five years ago when people were calling for them to have to revert. If anything, a couple of them are in stronger positions today. So it can't go on forever, but their market cap is not functionally out of line with the cash flows that they generate as a percent of the overall market. Many of them are growing at or above the rates of the market. So for the foreseeable future, you just have to say these are some of the most incredible businesses ever, and they're still performing. So if you're an active manager, yikes. What gets you excited then? You can easily access all those things at zero cost in an index. If you buy an index today, you basically are getting a bunch of exposure to those companies. They're the best companies ever. Everyone agrees. They're probably priced to that truth. What gets you personally excited? And I'm curious about all the frameworks here, like the competitive advantage period that Mobeson will talk about or the benefits and the barriers that Hamilton, Helmer, will talk about in finding new businesses that won't have this problem of pricing a temporary surplus because there's some barrier to entry for the kids in the garage can't come disrupt the thing. Is that your framework

Starting point is 00:18:35 for looking for things that are not just the index? Because that's easy to buy. And as an active manager, you got to find something. I know. I think everybody in this business should be humble about the fact that the index over the last five and 10 years has been undefeated. I know people, I'm sure you probably know more than me, but I know people with all different styles of public market investing. And I would say very few of them would be pleased with their relative performance to the index over the last five years and probably also over the last 10 years. As you said, in that middle 2010 period, there was still a fighting chance against the index. But over the last five years, it's been incredibly difficult. And something that I think

Starting point is 00:19:26 I haven't thought enough about, and I don't hear talk about too much. It's just the stress of COVID and everything that's transpired since really magnified strong management and competitive advantages and really stressed marginal companies and marginal managements. And I think along with all the other great characteristics of these companies that play to their advantage, that on a relative basis, a lot of the market has just been in a very stressed situation at one point or another. And look, the marginal management team is not great. The marginal company is not great. And this overall environment has really stressed and bifurcated those types of companies. As far as what gets you excited, one, I think it's a fascinating time. I'm not a technology expert,

Starting point is 00:20:18 but when you hear people say things like, this is as big as the internet, this is as big as mobile, and you pay attention and you try to learn, and it is fascinating what's going on. But otherwise, I think it's still try to find interesting opportunities with asymmetric risk return, risk reward, and hope that you have a fighting chance against what is a very difficult benchmark to be. And I said public markets only, but let's not explain. exclude the challenge for private markets. You have a lot of venture listeners on your podcast. The NASDAQ 100 has returned 5X MOIC on a pretty consistent basis for the last number of years.

Starting point is 00:21:02 If you just take a 10-year rolling QQ, that's a 5X MOIC for a long time. And so I jokingly ask my friends in the venture world, how many fun families do you know of that have multiple 5X MLICs? It's not many. And there are some one-offs where someone hit the right cycle. But to put up three 5x MLIC funds, very unusual, and yet the public market is doing that consistently year after year. You look at private equity. They claim mid-teens IRAs. Those are 16-17 MLICs. The continuous compounded return only worked if they were returning capital. If money in was the same as money out and the IRA was 15, then you got it continuously compounded. I saw a chart the other day that showed there's about a $300 billion deficit in money back versus money in

Starting point is 00:21:55 over the last three years. That is going to wildly impact the continuously compounded return that the investors are receiving. Again, put that against the 12.5% 10-year Kager of the S&P 500 with a quarter of the leverage and full liquidity. And you'd say, I'm not sure. sure, that was the right move. So the public indexes, I think, have stressed all forms of investing, not just public investing, in saying on a relative basis, why is money going elsewhere other than these indexes? This is a good thing, though, in aggregate, it just raises the bar for capital is not going to allocate itself. We need people, both in primary terms and in secondary market setting prices. We need some amount of active management. Probably we need more of it at the primary

Starting point is 00:22:44 capital level down in private markets when stuff is newer and you can't underwrite it quantitatively like you can in the public markets. So is everything you just said just really good for capitalism because it just raises the bar for capital allocators and biases towards better ones and this is just a good natural thing? The fascinating question, right? If markets were perfectly efficient and did not capitalize temporary surplus, just look through it and basically the price just was capitalism would be less incentivized to do its thing. If commodity prices did, didn't rise in the mid-2000s, many of those minds would not have come online, maybe, in the same way. You probably do need the distortive impact in order to draw in the technological innovation

Starting point is 00:23:28 that furthers society. So yeah, totally agree with that. As to whether or not it raises the bar on who the capital allocators are, yeah, I mean, this is something that I think has been happening for a long time. Mobeson has written about the paradox of skill. And I think this is a big, still to be had debate in the investing world, which is as money flows out of active, the presumption is that better active investors remain. And I think those remaining in the business believe that there are more inefficiencies today than there were as a result of that. If you think about what I said, though, that a lot of people are not happy about their five and 10-year returns streams relative to the index. I'm not sure that the math would actually show

Starting point is 00:24:18 that the remaining investors are taking advantage of these supposed inefficiencies, because if they were, they would be outperforming. And so I think that it's probably the case that you have very good, skilled active managers left. I'm not sure it's the case. That means that they will all outperform. I think it actually probably means it's harder. If you think about everything outside of tech, what has your attention and curiosity most? Like you mentioned the size, I understand it could grow incredibly fast and the same thing in the internet era would have been silly to say. But let's say it's $10 million of impact, real impact from both the cost and the revenue side generated from AI products so far. There's lots of individual biologic drugs that do that much revenue by themselves

Starting point is 00:25:04 in a given year, growing 20% by themselves. I looked up this one drug duPixant that does $14 billion of sales growing constrained in its growth because we can't manufacture enough of this stuff. Some of these other markets are so big and get so little attention relative to tech. And I'm curious what you think about everything X tech. I don't think we've even scratched the surface on the microeconomic and the real existential questions in AI. But that's just AI and tech.

Starting point is 00:25:32 outside of tech, I think the most fascinating thing to me is, again, we last talked on this podcast in the fall of 2020. We were sitting in our houses still at the time. There was no vaccine on the horizon. The market was still pricing permanent zombie apocalypse. In the time since that, what's been fascinating for me to see is how many industries have suffered from a capital cycle as a result of COVID. And at the time, I was only thinking forward, where may a capital cycle occur? And some of them I was right about, some of them less right, but I did not realize how far flung the impact would be. And so I'll give a couple of examples. If you look at the transportation and freight markets, because everyone was stuck at home and

Starting point is 00:26:20 ordering three washing machines and 18 footballs and all the hard goods that we were ordering in 20 and 21, you had this huge surge in demand. In order to move those goods through the network in the United States, lots of trucks were added. A lot of capacity came online. In May of 2022, Target and Walmart came out and said, we are absolutely stuffed to the gills with inventory. We are going to liquidate all of this,

Starting point is 00:26:50 and we are going to stop flowing goods. Two years later, they still have not started flowing goods in, and you have this enormous overcapacity. in the transportation market. And it's not just trucks because truck sets allows marginal pricing for rail for intermodal transport. And so you just have huge areas of the transportation markets where you don't have volume growth, you have overcapacity on the supply side, and pricing sucks. So you're literally two years later, the hangover is maybe starting to wear off in each quarter. It's, oh, I think we're getting close to the bottom. That's one. This one really amused me.

Starting point is 00:27:27 beverage cans grew two to three percent globally a year for long, long time. There are three public companies, two of them, ball and crown. They were spending about $600 and $500 million a year in capital. COVID comes along. People are sitting at home, drinking a lot of sodas and beers, off premise, not on premise. Demand goes from 3% a year to 10% a year. When you're a largest customer, you can think of two very large soda and beer customers come to you in. say, we need more capacity, you either build capacity or someone else builds the capacity. So, lo and behold, capital expenditures go from $600 million a year to a billion and a half a year for each of the company. So they do that for two, three years. Then the zombie apocalypse

Starting point is 00:28:13 ends and people go back to bars and people go back to sporting events and they stop drinking sodas in their home and demand reverts. And now there's a bunch of capacity. And so for two years, it was, oh my God, we have this huge overcapacity in the beverage can, which was a historically very stable, oligopolistic, great market structure, rational capital allocators. So you just go through one by one and you see how distortive COVID was. And look, at a macro level, you see it in the way that prices spike, the way that wages spike. There were level changes in demand that required. a price increase to draw in supply. And so I think I've been super intrigued by a lot of these because if you can catch the trough of those, they overspend, then they pull back, then

Starting point is 00:29:11 capital spending troughs, and then you get a new upcycle. That's a very interesting place to be. Amazon built UPS in two years. That's the quoted stat is they put so much capacity into their distribution network in two years that they rebuilt UPS and they suffered the hangover when demand slowed. So there's all these pockets throughout the economy totally unrelated in some cases to technology where if you can catch inflections in the capital cycle, you probably have an interesting setup. Candidly, freight's not there yet. People have been trying to call this for a year. And so there are areas where I think outside of AI, some of the GLP1, which is the other mega trend, I think you can just do the normal work that you historically did.

Starting point is 00:30:02 But there's this common theme, which is the distortive impact and the capital cycle impact of COVID. There's this ultra-simplifying version of this is like every single shortage is followed by a glut that sends that people talk about. Is there opportunity in exceptions to that rule where there's a spike in demand and the opportunity is that there's some barrier to on-streaming new supply and that it's identifying those areas that create especially interesting opportunities, or is capitalism just so damn good at supply rising to meet demand that it's just always this way? As you described it, the word Kuta jumped into my head. But yes, look, clearly markets are making their bets on where

Starting point is 00:30:45 demand has risen and capitalism will not respond. I think. I think that's always the trick is to understand or to theorize where there are significant barriers. And those are the things you mentioned, whether it's Homer, Porter, or Bobeson, how much of the rapid increase in surplus is a company able to hold on to? Nobody knows right now, but that is the bet that markets make when they put rising multiples on rapidly rising surpluses. To make that real, does a company like ASML come to mind where, yeah, you need what ASML provides, but you can't just on stream a new ASML. It's unbelievably complicated, so it's

Starting point is 00:31:25 process power or something in Helmer's framework. Yeah, sure, but every once in a while, you hear someone be like, oh, this semi-cap provider believes that they have a new process, which would be as efficient as EUV at a tenth of the cost. Do I think that's actually going to happen? I have no idea. That's way too technical for me, but little headlines about that. ASML is an incredible company. they're a choke point for the entire semiconductor process. You have to believe, though, that smart people are aiming to dislodge those choke points at every given moment in time. I think the question in those cases is more the duration there. That's not going to be instantaneous. Look, those are well-recognized IRISAone this year. Both of those were pitched. And I sort of laughed

Starting point is 00:32:11 about it. You're an aficionado, markets, and historian. If you go back 10 years and you looked at the pitches at IRISOne, it would be some distressed debt. Here's a turnaround situation. Here's a short. Here's a middling 12 times PE that we think can rewrite to 15. And this year, we had Daniel Gross interviewing the CEO of Magic. Dev. We had ASML winning the best ideas contest, TSM pitched, just a different world. That's where all the focus is rightly, because that's where the returns have been. And again, that's the point of the job. If you had to pick three companies, you're an upstart kid in a garage and the universe of opportunities to go attack one of these juicy profit streams exists. Which three profit streams would you be most terrified

Starting point is 00:33:04 to try to attack? What would be the hardest castle to storm? That's a great question. Does it have to be in technology or it can be anywhere? Oh, you pick. I think Apple is an enigma because because it's so fundamental at the top of the food chain, and yet it goes four to five years without growing profits, or top line even. But displacing Apple, you almost have to believe in AGI. You have to believe in a model so powerful that it basically abstracts away the need for the integrated hardware software solution. If you think AGI is coming, Apple's probably your target. But in the world that we functionally operate in today, I'm not sure I would want to go attack.

Starting point is 00:33:48 A great comment a CEO once made to me was no independent device has stood up to the mobile, does it the smartphone and want. And I think that holds true to today that it is the unifying device of every ecosystem. And so I think that would be a tough profit pool to go attack, as many have learned.

Starting point is 00:34:09 You would have said five years ago, I think someone on your show pointed this out that if you just watch where VC dollars were being invested, nobody was going after Google. And that if you went on Sand Hill Road and said, I'd like to raise money to attack the search engine business, that you would have been laughed out. And that's generally a sign that there's some enormous power present. I think clearly that's no longer the case, since perplexity is raising money. That was one for a long time. You would struggle to go after their profit pools. But look, there are niche businesses. I guess people out there are probably

Starting point is 00:34:44 trying to figure out, I know they are, trying to figure out a more energy and climate-friendly version of aggregates, which are rocks. But it's pretty hard to go after the profit pool of rocks right now because the way to do that is to have another rock quarry, and they own most of the rock quarries. So capital intensive. The distance, they're local monopolies, effectively. The distance to transport the rocks is the challenge, and they own the rocks, and you don't. So I'm not really sure. You need a local. cost synthetic material to displace them. And if you're smart enough to do that, there's probably some other better use than displacing rock. To say a bit about your reaction to GLP once,

Starting point is 00:35:24 you and I have never talked about this. Yeah, look, I think it's amazing. I have nothing smart to add on the science. I think on the behavioral side, I was intrigued to read, it was a Sanford Bursi and a research report that talked about the occurrence that humans have to drugs they have to be on for a long time. And I think the adherence to cancer drugs was something like 60 or 70 percent, and then it rapidly fell off from there. And one, that's just an insane comment about human behavior. But to the extent today, GLP-1s require this ongoing dosage, I think that strikes me as a pretty big behavioral barrier to getting to the types of penetration that would be very societally beneficial. If you just think about at 35 percent adherence to get to 50 million Americans on the drug,

Starting point is 00:36:23 you can run the math on what the annual number of people having to be on the drug is. I think clearly that the health benefits of them seem extraordinary and we learn about new benefits every day. I think The real question is, does anything emerge longer term side effect wise or anything, but the safety seems to be holding up so far? Then the real question is just, how do you get adherence to these in a way that the benefit sustained? Look, this would be, we talk about the economic benefit of AI. I mean, the economic benefit of taking a huge chunk of the country out of obesity, out of diabetes, out of the long-term, care costs that we absorb via Medicare, Medicaid, via health insurance, that would free up productive resources in an unimaginable way. But again, human behavior is a very difficult thing to change,

Starting point is 00:37:23 particularly if life-saving drugs are only being adhered to three-quarters of the time. What seems craziest to you in the world of capital markets today? What do you just look at and just laugh or shake your head and almost can't believe. I think that Warren Buffett famously talks about the three eyes. I think that David Swenson and what he did was genius at the time. I think that the endowment model and the allocation of assets into privates, I think has gone too far. I think that, and this is not self-serving as a public equity investor. This is as an observer of the landscape. I think that the reasons now for the capital go into private markets is much less about risk reward and gaming the efficient frontier and taking advantage of opportunities.

Starting point is 00:38:19 And it is much more based on institutional smoothing, return smoothing and volatility avoidance. And I think that these pools of capital have, if you want to call it a laffer curve or something, have gotten onto the backside of the curve where, it's probably detrimental the extent to which assets are being moved into illiquid environments. I just don't think that if you actually sat down and did the math and looked at everything, that these pools of capital are best served, having 30, 40, 45 percent of their assets in illiquid privates. Look, we're at a moment in time, clearly where the public benchmarks are extraordinarily difficult to beat. but to some extent, actually to a large extent, the private markets are a reflection of the public

Starting point is 00:39:11 market. So to believe that the forward 10-year returns of public markets are going to come down substantially and not impact the forward market returns of these private asset types is kind of crazy. And I just think that, look, if you were early to venture and you're in five or 10 funds that have demonstrated persistence, that's awesome. The challenge is, for those pools of capital that got in early, those five or 10 funds probably represent an immaterial portion of their overall asset base. And so even if you continue to earn extraordinary MOICs, it's probably becoming less and less of an impact on the overall return stream. And for those later to the asset class, you are getting access to the non-persistent performers, and you are now exposing yourself to probably NASDAQ at best returns with an illiquid profile. Private equity, we talked about private credit.

Starting point is 00:40:15 There's a rush into this asset class. There are a handful of public managers that report their return. So you can look at what private credit is doing. And again, I just question whether the efficient frontier is being properly calculated, taking into account the inherent risks that are there. And so that is the thing that strikes me as craziest is the overall allocation of assets between public and private. If you were put in charge of, I don't know, pick your state pension or something, some massive pool of capital, $100 billion, Canadian pension. What would you do? What would you do? It's funny. When I go to conferences, I love finding people who have. have that job because I think it's very hard. And I'm also intrigued by what they're doing because what they are doing is probably what their peers are doing. And that's where a lot of money is going.

Starting point is 00:41:04 Look, I think you have to embrace the idea that some diversification is good. But I think you also have to be cognizant. Let's say you run $100 billion. If you want to have a venture portfolio, think about the size of that venture portfolio to matter to your overall return stream. Then think about how much that capital is as a percent of top quartile venture funds and think if it's truly accessible. If you look at private equity, you would say, okay, maybe there are some folks who have demonstrated outperformance in the mental markets or something. But for the most part, you're getting four or five, six times leverage to generate, like I said,

Starting point is 00:41:44 a 16, 17 MOIC. I think you have a very hard job, one. Two, today you actually have a bonds are an option for the first time in a long time. So you can earn yield. I mean, think about this job in the mid-2010s where you looked across your spectrum. He said, okay, I can get zero here. So I think bonds represent a real opportunity today. And if the environment we're in offers up one, two percent real returns on the 10-year U.S. Treasury going forward, that is a very different environment than what we had for 15 years. prior. And if you're an allocator, that's something to consider. But look, I think you have to have a

Starting point is 00:42:22 significant weighting in U.S. equities. I think you want to have a significant weighting in international develop market equities. You probably want to have some bonds and maybe you have some privates. But I think you need to be very cognizant of why you are investing in private. I do not have a personal belief behind this question. But why have any weight to international equities? It's a good question. The funny part is when you look at, there are great companies in the rest of the world, and when you look at them, they are usually priced at or above their comparable U.S. peer. So Europe has some wonderful companies, but they are not cheap when you look at them. I think you do have to be cognizant of the fact that over time, unless you believe that

Starting point is 00:43:06 the U.S. just outcrowed the world in perpetuity, which has been a fairly good bet, at least the developed world for a long, long time. I think you want to have some exposure. Episodically, like right now, if you look at the world, I think Japan is a very interesting potential opportunity. And I will say potential because every international investor has gotten rug pulled probably twice in their career already by Japan being this time is different. But what's happening there is very meaningful in terms of the Tokyo Stock Exchange, essentially shaming companies into adopting more forward-looking capital management program. And they're basically requiring companies to put forward a capital plan. And if they don't, they put out a list

Starting point is 00:43:56 periodically of who has and who has not. And in the classic sense, what's happening is those who have not put forward a plan are getting shamed into putting forward a plan. Now, those plans may be underwhelming at the beginning, but every March starts with the first step. And for anyone who's paid attention to Japanese equities over a long period of time, these are wildly overcapitalized for the most part. They have huge cross shareholdings. They don't do buybacks. They have very low payout ratios. They have nonsensical cross holdings of, oh, this is a strategic asset because we founded this company 60 years ago when there was no venture capital, totally unrelated industries. Look, I think there's a chance that could be a very interesting place for capital for a number of

Starting point is 00:44:41 years. So I don't think you want to holistically write off international. But if you're a $100 billion allocator, you need to be dynamic in your process there, identify the opportunity, and then find the manager to do that. And that's hard. I am very sympathetic to that job. I think picking managers is probably as hard or harder than picking individual securities to outperform. If you think about the way markets felt to you 10, 15 years ago as an active participant, buyer and seller, this is like a market structure question. How does that feel most different today? What would you say about market structure, how markets trade, what it feels like to be a person buying and selling in them versus 10, 15 years ago. Has it changed much?

Starting point is 00:45:24 It's a great question. I'm not a frequent enough trader. Other folks might have difference of opinions, but a couple things I've noticed is the definition of what I would call a small or less liquid name has moved from a billion dollar market cap to 15 billion, where the sort of lack of interest in the generic buy 10 billion dollar company that doesn't have news on that day, it's just left for debt. And the attention has moved so much towards whatever the flavor of the moment is, and then the larger stocks, that I just think there's a much larger aperture of names that are left behind. That's one thing. I can't prove this quantitatively, but it does feel to me like objects in motion stay in motion longer, meaning whether it's

Starting point is 00:46:20 algorithmic trading or what have you. But it does feel like when prices are going in one direction, they tend to go in that direction more vigorously and longer. So these trends are quite pronounced until they stop. Those are the two biggest ones. It definitely feels more like a trending market and it definitely feels more like if you don't have a story, the size at which you are irrelevant is much larger than it used to be. If I could go back through your entire history of every buy and every sell you've ever made in public markets and studied those days and those decisions and the things leading up to them. What would I find in aggregate? What would I see as the common reasons why you were a buyer when you were and why you were a seller when you were?

Starting point is 00:47:06 The buyer generally would be, I think if you piece together from the outside, you would be able to put together a story of some sort of controversy, tension, unknown, currently unknown, but probably likely to be resolved in the near-term type event or overhead. And so I think particularly the ones that worked out the best, I think you would be able to recreate, okay, XYZ is the story. People believe this, clearly taking the other side of that with the expectation that it would be resolved in this way, and lo and behold, in some period of time, that happened. The sales would be much, I think, much more across the board in terms of, I'm not sure there

Starting point is 00:47:54 would be a unifying theme. In some cases, the tension would have resolved. In other cases, I was wrong. Some cases, there was a better opportunity. So I think most investors probably have a much more consistent reason for buying than selling. I think circumstances change to to cause you to sell. Sometimes the plan plays out perfectly and you sell, and sometimes they're just better opportunities. So I think that's much more diverse. I think if an investor has a process that the buy should be more similar. Where do you feel pockets of that tension today? That's a great question. I think in some of these capital cycle situations where it's this unknown and if you could find them, I will say the macro right now is probably.

Starting point is 00:48:43 probably driving a lot of tension. And it's interesting today, we had a favorable inflation report. And because the macro environment is still disrupted, I think, by what happened COVID and post-COVID, we've stepped up to a new level of real interest rates. The nominal interest rate is higher than we've seen in a long time. The housing market is in this very strange place where we're selling 4 million existing homes a year versus should be probably five and a half. But within that new home sales are soaring because nobody's selling their existing home. There's just a lot of macro-driven distortions. And so I think there's a lot of tension to be unreleased.

Starting point is 00:49:30 And you're seeing that to some extent today, where if you get better visibility into the path of interest rates, look, there's going to be quick twitch reactions in the residential market and in homes. And then there will probably be a slower but very real reaction in commercial real estate. This is a big relief to private equity. This is a big relief within non-residential construction type activities. So it's not so much tension as there is potentially a rubber band that could snap back in some of these markets that have been very pressured by particularly the front end, the level of front end interest rate. And I think it's hard to believe that with stock market at all time high, the economy doing very well, that there are still very large pockets of the economy that are essentially flat to inter recession.

Starting point is 00:50:30 And those areas that may benefit from this, I think the question is, how much will they rebound? who will capture profits, but the macro is driving a large portion of the non-tech tension right now. Is there any idea or ideas that have your mind on fire lately? Like things you've encountered that you're just blown away by? I'm super fascinated by, we have not talked about this with regards to AI, but I am fascinated by what I perceive to be a coalescing conclusion in the investment world that LLM models are, if not commodities, are going to be relatively undifferentiated and that the frontier models will push forward,

Starting point is 00:51:23 but then whether it's Lama or Nistrales, these guys will fast follow. And that the open source models will, in a classic Christensen sense be good enough. and that the frontier models maybe are relegated to certain types of real leading edge activities, but that they don't make the leap to AGI, or that they don't make the leap so far that they are differentiated. On the other hand, just in the last 10 days, you've had two people, one who's left Open AI and one two nights ago within Open AI, basically say we think AGI is here within three to five years. And to have two diametrically opposed sort of opinions like that, it's really important how that shakes out, not just because who knows what AGI means or any of that. Let's leave even AGI out of it.

Starting point is 00:52:19 If the frontier models are able to establish a significant performance advantage such that you use them and not open source for most things, incredible surplus accrues to Google. to Microsoft and OpenAI. AWS, big loser. Lots of software company, big loser. In a world with relatively undifferentiated models and open source being good enough, enormous surplus accrues to application. Because essentially what you have

Starting point is 00:52:53 is a commoditized model layer. The infrastructure may capture some of it. That's the question we started with. But then this enormous world, is opened up for the application layer. But in the world where the frontier models are that much better, like, why won't a generalizable model be a specialized model? And so I think this, and I've been with people much smarter than me and asked the question. And the reality is, no one knows. There are a few people who very fervently believe in one of those paths. And then the

Starting point is 00:53:27 investment community is just along for the ride. Yeah, but we're all sort of a come to this idea that, look, if meta's going to spend $40 billion fast following, they're going to commoditize their complement, and then you'll just run these models where your data is and data gravity is real. And if your data's in AWS, it's there. If it's at Azure, it's there. That world looks very different five, six, seven years from now than the world where the frontier models build. Again, let's not harp. on building God, let's just say the generalizable model is so good that it usurps everything else. That is just a very different outcome. And it's interesting to me that the investment community

Starting point is 00:54:16 has written that path off. I think that's a great place to close. As always, an incredibly interesting conversation on all things, markets. I wish it hadn't been four years since our last one. Maybe we'll do an every two-year scheduled episode or something like that. Thank you so much for your time. Thanks. Really appreciate it. If you enjoy this episode, check out join colossus.com. There you'll find every episode of this podcast complete with transcripts, show notes, and resources to keep learning. You can also sign up for our newsletter, Colossus Weekly,

Starting point is 00:54:45 where we condense episodes to the big ideas, quotations, and more, as well as share the best content we find on the internet every week.

Invest Like the Best with Patrick O'Shaughnessy - Modest Proposal - AI Commoditization and Capital Dynamics - [Invest Like the Best, EP.380]

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.