Thinking Beyond the Headlines with Jacob Miller

How do you make informed investment decisions when the world feels increasingly unpredictable?

In Episode 41 of the Tao of Chao Podcast, Philip Chao welcomes Jacob Miller, Head of Investments at Opto Investments and former investor at Bridgewater Associates, for a conversation about private markets, global macroeconomics, artificial intelligence, and the interconnected forces shaping today's economy.

Drawing on his experience across institutional investing and private markets, Jacob shares why understanding long-term trends, market structure, and systemic risk can provide a stronger framework for navigating uncertainty than reacting to short-term headlines.

Key insights include:

• How private markets are creating new opportunities for investors
• Lessons from global macro investing and economic cycles
• The growing influence of artificial intelligence on the economy
• Why debt, monetary policy, and systemic risk matter
• The importance of trust and critical thinking in a rapidly changing world

This episode explores how investors can develop a broader perspective, navigate complexity with greater confidence, and focus on the long-term forces shaping the future.

👓 Learn more about our HOST

Philip Chao
Website: https://philipchao.us
Follow Philip on LinkedIn

DISCLOSURE: Views expressed in the Tao of Chao podcasts are individual opinions and they do not represent the employers of each guest or the firm with which each guest is associated. Our podcasts are for educational and informational purposes only and should not be deemed or viewed as investment advice or recommendations. Please consult your personal financial advisor, investment expert, or investment fiduciary before taking any actions about your plan and investments.

Transcript

Copy Transcript Copied

0:00

I think we're on the verge of a true healthcare and biotech renaissance. I think the impact of AI there will be extremely creative and the data we're collecting on how to improve healthcare outcomes is pretty amazing. And technologies like CRISPR are getting more mature. So I'm very interested in how technology gets applied to that space. Welcome to the Tao of Chao podcast, where we will try to find balance and provide a clearer path forward in this uncertain world. Jacob Miller, is co-founder and chief solutions officer of Opto Investments. In his role, he focuses on building technology and designing investment frameworks for RIAs, family offices, and private banks seeking to build or scale private market programs. Miller's career spans both public and private markets. He was part of the investment team at Bridgewater Associates before moving into private markets. developing a long-term discipline approach to portfolio construction. Passionate about making the investment management more transparent and efficient for both providers and users of capital, he co-founded Opto alongside Joe Lonsdale to help the wealth channel and institutional investors better access differentiated opportunities across infrastructure, private credit, private equity, real estate, and venture capital. Beyond Opto, Miller serves as build executive at 8VC, a technology and life sciences investment firm, and as an advisor at Etkium, a data-driven venture capital firm focused on private markets. And here is Jacob. So hello, Jake. Thanks for joining me today. I've been looking forward to this conversation. Jake, before I get started, I know a little bit about Opto, the firm that you co-founded. Tell us a little bit. specifically, what is it that you intend to accomplish? What is it that Opto does so we can get the conversation started? Yeah. So I founded Opto about five years ago now alongside Joe Lonsdale, one of the founders of Palantir, founder of Adapar, runs ABC. We saw a broken private market system. For the bulk of the last two decades, private markets were either inaccessible or the ways that you accessed them were polluted with perverse incentives, with a lack of alignment, and with technology that did not rise to the occasion of meeting modern needs. And I am a firm believer in market economies, and I think they work best when a lot more people can participate. I think the sort of zero-sum thinking that permeates a lot of our society is not all that surprising if most people feel like they only read about success stories after they can access them. And so our goal is to use the best technology process and people to make private markets a core part of how people save and plan for their futures, both in the US and globally, and to hopefully back a lot of the best ideas and accelerate the future in the process. Terrific. I think it's very much aligned with what the Department of Labor has proposed new rules on. I think the word banter around is democratization. Um, which is a beautiful word. Beautiful word with, with some issues we can talk about later. For private assets. So I think you are, you are ahead of the curve, certainly. And I think that the whole issue about benchmarking, which is what, uh, I think is going to be challenging, uh, and, uh, and differentiation and trying to figure out a single term like private credit means so many different versions. Uh, and it's not a, it's not a monolithic item. So I think that with firms like yours should definitely add a lot of value to, if nothing else, conversation and differentiation and benchmarking and thinking about all these different things that people need to know about before they dive into private markets. I know, Jake, you told me little bit of a story, kind of interesting story about a part of your life where you got hired by Bridgewater Associates. And do you mind sharing that story or you rather just go straight to talking about global macro issues? Yeah. So I have always been, as long as I really formed memories, at least very interested in markets and economies. I grew up in the middle of San Francisco and was sort of becoming self-aware during the dot com boom and bust. So even on our block. fortunes created and destroyed pretty quickly. And something just kind of clicked that this was sort of the meta, the meta problem of you got to learn a little bit about a lot. And that's how my brain likes to work. You could call it a lack of focus, or you could call it just gracious curiosity and thinking deeply about these complex systems where, know, everything from crop yields in Africa to what car is most popular in the United States can in their small way, fold into what the price of oil is today. became obsessed with that, studied economics at the University of Chicago, which there's some irony to that because it's the home of efficient markets. Theory and I now, and really for my whole career, I've operated in proving that markets are a little less efficient than that. I was writing about markets, trading markets. Some of the folks at Bridgewater were apparently reading some of what I was writing, which is... I think still a little bit surprising to me. I'm not sure 15 years later if I know what I'm talking about today. I definitely didn't then. But landed there as an intern, spent almost five years there. One of the best crash courses in macro and markets you can get, really intellectually. They call themselves the intellectual Navy SEALs. I think that's both good marketing and to some extent quite true. You basically have paid to get a PhD in macro so spent a while there and saw from that vantage point just how important private markets were becoming to the overall economic picture. When you think about trading something like high yields, you obviously care what's the supply and demand for credit in the economy. But if most of that supply and demand, the marginal buy and seller are meeting on the private side. Now you need to really pay attention to what's happening there to understand the overall economic picture. so became obsessed with the idea of applying that fundamental but quantitative framework to the very messy data problem of private markets. Left Bridgewater in late 2019, met Joe and the rest is history. That's wonderful. So we talked about many of the things that you just talked about thinking about the world in a macro way. and then looking at the different aspects, how it impacts affect markets, both private and public. But interesting, you're talking about University Chicago and efficient market theory and how not everything is efficient and so on, but it is the inefficiency that give us the arbitrage opportunity. If everything is efficient, we should all just fold up our 10 and go home. So to speak. So we want it inefficient in some ways. Inefficient doesn't mean like a transparency. Part of it is how you translate and inform yourself, which is something that I know you spend a lot of time on. Yeah. I think there's two sort of philosophical angles to come at that question at. One is, what are markets really? in computer science, have P versus NP. You have the idea of like ergodic versus chaotic systems. Markets are... self-reinforcing and self, there's a ton of feedback loops and I think there's not to be overly mathematical here, but corollaries to chaotic systems and physics. There's reasons why it takes the largest supercomputers on earth to even model how three quarks interact. And so I think it's a bit hubristic to think that markets, which are a very, very complex system, necessarily have a finite solution, which is what a true efficient. frontier would imply. That doesn't mean that they aren't pretty damn efficient in most cases, and especially in large liquid equity markets. But anyone who's been around long enough has seen clear inefficiencies. The other angle, is where I've spent most of my time in the last five years in private markets, is there's probably fundamentally four ways to outperform. You have to have unique access. So you're able to buy and sell things other people can't. Unique information. You know things other people don't. Unique insight. You're just smarter than everyone else. And lastly, you could be a uniquely lucky person. You probably don't want to bet the farm on luck. And the first two are illegal in public markets. That's insider trading. Do not pass go. Go straight to jail. Do not collect $200. Well, probably collect $200, then have the feds take it back later. And so most of the public market game is that unique insight piece. How do you get smarter than the market and how do you stay smarter with everyone trying to improve every day? I do think it's possible. I saw Bridgewater do it in many markets, but it's damn hard. But in private markets, unique access and information are the lifeblood. are not SEC or registered securities. These are not public recording required companies. But that also, what that really does, it blows out the spectrum. So the difference between top and bottom quartile is 25 % instead of 5 % for public market. So the opportunity goes up, but so does the risk of being on the wrong side of that information and access. your pay payoff is a little potentially higher if you get it right. Also, if you get it right, can be in the twenties by RR consistently. Right, right. So that's fascinating. And one thing I always think about is when we democratize, so to speak, a market that really wasn't democratizable in the past for either by choice or by lack of data, lack of system, whatever it may be. That eventually the private market will also become more like a public market because as you make information available to everyone, it becomes more and more like a public market. So we are arbitraging away some of that secrecy, if you would, of advantages. Not tomorrow, but over a very long period of time. You think that may be true? I think so, but I think it is happening not in a consistent way. And so when you think about things like the Department of Labor action and the early opportunities coming to 401k type accounts, you're generally talking about the very largest fund managers selling their very largest products in semi-liquid formats. At this point for the multi-hundred million dollar loan size, $25 billion plus credit funds for the real estate funds, which are party to, a crazy percentage of real estate transactions in the country. I don't think there's a reasonable way to look at that and say that that could possibly be alpha. Now, I think that private market beta has a place in portfolios. It likely has slight diversification benefits versus... Just from a cap M perspective, if you believe that the economic engine of capitalism will pay you over time, you want to hold that whole engine, not one slice of that engine. That's right. But when we think about... democratization, we're generally talking about the very largest things and I certainly wouldn't walk into those expecting those to be top quartile alpha opportunities. Right. Which gets to like sort of our broader thing. We tend not to talk about democratization or we talk about it, but maybe with a more negative connotation that others do, you know, our belief is that the alpha generating part of private markets will never be democratized. That is what makes it alpha. I mean, there's just no feasible way that the millions of small opportunities that are happening, call it below the $100 million check size level, are ever going to have the transparency that would make me comfortable with individuals investing. So it'll be a fractured market. There'll be more and more private market beta products that are available. Right now, my view is they're still too expensive, but we hope to be part of bringing down the cost of private market beta over time. But that alpha piece, which again, if you can get it right, is so important. is likely immune to being truly democratized. So it's truly manager selection. Manager and opportunity. Yeah. Who do you trust? Absolutely. Go back to the basics. So let's turn to a little bit about your time at Bridgewater. I don't mean the time talking about what happened, but really what came out of that just for the audience of Bridgewater is the $230 billion macro hedge fund. the firm's core strategy, if I understand correctly, is a global macro investing, which operates on a kind of a fundamental cost and effect understanding of global economics and tracking broad factors, like inflation and growth and currency and fluctuation and so on and so forth. What did you find particularly fascinating when you were a younger man sitting in the midst of all this and saying, wow, this is interesting, fantastic? uh, how predictable really it is and how we deal with the longer term of these, some of these trends and how do we take advantage of the shorter term, you know, all that. Can you talk a little bit about that? Yeah. So, uh, you know, exactly right on, on Bridgewater. So, uh, uh, two core products, there's all weather, which is their beta product. It's a risk parity. Uh, but the primary reason most of us investors were there is the pure alpha, uh, fund, which is meant to be a, you know, directionally neutral pure macro, uh, any market that's big enough for them to trade. Tends to focus on call it medium hold. It's not a quant millisecond trading. It's what are those three to 18 month trends that seem mispriced in the market. What stood out to me most early on is A, just how inefficient some markets that seem like they should be totally priced in are. mean, take short term US rates. If you use a simple sort of Taylor Rule-esque model and go back through time and you build your model correctly so you take out all the foreknowledge and only know what you would have known at the time, you can still generate a long short signal that has a, it's not gonna be crazy high, but call it a .4 sharp with no bias to be long or short the underlying. But then you really start to think about that. Why is that the case? Why is there something traded in that liquid market? Well, and then you realize the biggest actor in that is the Fed and the Fed is not trying to profit maximize. They're trying to meet their dual mandate. And so if you understand the player and you try and project what is that player going to have to do in the next three to six months, you can trade that very well and it's not necessarily priced in. Now contrast that to is Google going to be higher or lower tomorrow? I don't know. And I don't think most equity analysts do either. They sound like they do though. sound like they're very, very confident. But we tried to find those scenarios where we believed for a logical reason. There were actors who we could predict and there was a reason that that wouldn't be immediately priced in. And so what was really interesting and what we still try and do here, we're very macro oriented for a private markets focused firm, is get to the player level. It's not enough to... know, predict inflation and growth. That's your starting point. Then you need to say, well, who owns assets and what do our projected changes in those forces due to that actor's behavior? And so, you know, it's things like predicting pension rebalancing flows. If this was more relevant prior to, you know, post-COVID inflation, but, you know, take, you know, Euro bonds trading at deeply negative interest rates. And I think it confused a lot of people. Who's buying these things? Well, if you go to the actors and you say, look, there's hundreds of billions, trillions of dollars of capital that are regulatorily required to buy AAA European bonds at any price, and there are no AAA European bonds that are not negative, well guess what? They're still gonna buy the negative bonds. And so you have to understand your players, their holdings, their reaction functions. And that's very similar to how we think about what creates good exit conditions, good entry conditions on the private market side. It's interesting you brought up that during the negative. negative nominal rate. I'm not even talking about real rates. Where somebody said, well, it's better to know how much you are going to lose rather than put all the cash in the wall, get eaten away by some animal. was interesting thought. So no, I have to agree with that because that's really what happens. So I want to sort of set the stage for a definition of polycrisis. And this is a condition, a polycrisis is a condition in which distinct crisis becomes causally entangled so that they compound rather than they simply add to each other. And in which the tools that are used to relieve one crisis tend to aggravate another. So leaving no action that is unambiguously constructive. I think we are in that situation. I won't name them all, but the ones that obviously come to mind today is obviously tariffs. We never ending tariff change, regardless if we agree with it or not, but it adds uncertainty to the situation. Certainly the war in Iran, which very interesting in some ways that they thought when, I think almost four years ago, that when Russia invaded Ukraine, it was a very short war because it's supposed to be over soon. We're still fighting. And so this one, I think is going to be longer than a month, which obviously is. So those are among others, right? And so why I brought that up is because there are just so many of these issues, shall we say, or conditions currently on a global macro standpoint, let alone impacts all of us in a micro standpoint as well, that I'd like to sort of spend a little time with you today to address. I think maybe we should start with Walsh and the Fed because that's probably their meeting next week, I believe. I looked at CME, the Fed tool site, and they're still expecting an increase this year. And there may be another one in January. So expect in the next six months, a 50 basis point rise, assuming it's 25 basis point each. I'm not so sure the market is right or wrong. I'm not smart enough to bet one way or the other. Tell me a little bit about, before we talk about interest rate, tell me a little bit about what you see and think about as you listen to Walsh or when you read about him, what's your thought about who he is and what is he really trying to accomplish? Yeah, it's, I'll start by saying you always know things are getting interesting when you have to care about the Fed this much. And particularly the individuals. Obviously the Fed is always important, but there were hedge funds in the 80s that traded based on how thick Paul Volcker's briefcase looked when he left the office. We're getting back to that level of Fed scrutiny, which always means interesting waters ahead. Look, I think he has a tough job of, he's smart enough to understand the importance of either perceived or actual Fed independence to markets. Regardless of who put him in the seat and what their expectations are from him, he knows how disastrous it would be if it looked like the presidential independence was faltering. He's also obviously facing continued persistence above target inflation. And with the re-hot war situation in Iran, that looks unlikely to unwind anytime soon. It turns out oil is very important to everything, not just at the pump, but in getting all the goods we buy to the store. So it is the base layer of most inflation. What is interesting, and I always go to, I look at markets before I look at the news. It's my compass for what's mattering. Interesting that gold has fallen 25 % ish since all time highs at the same time as oil has run up. Now, This is not a contradiction. think a lot of people think of gold as a naive inflation hedge. It's a hedge against a very specific kind of inflation, which is monetary inflation. It is not a protection against growth-based inflation when real rates will rise, which is the damper on gold. And it's not a one-to-one capture of commodity-based inflation. What it falling at the same time as that happening tells me is that, you know, I don't know whether they're right or wrong, but markets are fading on the lack of independence. They're hearing Walsh and they're saying, This is not a departure from Fed independence. That's why we don't have to go buy all the gold in case he cuts interest rates from here. I mean, you see that in the expectations of a future hike and one more next year. And so there's that side of the equation. I don't anticipate him particularly rocking the boat on that side. Now, there's another piece which you've written about in some of your writings on him of what does the Fed do about the treasury side of this. And this is two pieces. One, the existing Fed balance sheet, whether or not to roll that over, how much to roll over. We still own a lot of our own government bonds through the Fed and do the hypothecation thing every time we send the interest payments back and forth. But then also continued pressure on long rates from the government deficits and no real sign of reducing that deficit. And look, the Fed can easily control the short end of the curve without something more extreme like QE. It will struggle to change the long end of the curve with bonds being issued there and without rolling over or increasing rollovers. And while the short end, think month to month is more important to most asset pricing, the long end is what's going to drive mortgage rates and long-term lending and a lot of things that matter more to kind of the main street economy. We've had a lot of reduction in the impact of household fixed investment and real estate generally on the economy. It's generally a pretty, it's not large, but it tends to be the marginal mover in a lot of cycles. Not large in absolute terms, obviously the consumer spending is the biggest part of the US economy, but household fixed investment does tend to lead. And that's been based on both obviously the affordability side, but also the interest rate cost side, a dampener over the last 12 months here in the U.S. I'm sure he's staring at that and saying, you know, how do I bring the long rate mechanism down without appearing not independent? I think he's in a bit of a bind there. Well, he is probably doing that by bringing Well, let's use the word coordination, which is what he suggests with treasuries and moving the new issuance, not Hague can move it, but treasury will coordinate by issuing new ones in the short end, which then he controls, he being the FOMC control. So that's a form of coordination is what I'm thinking about. And as you know, we are growing at a clip of six to eight billion dollar more of deficit as we, I think, GOA, GAO, not GOA. predicts. And so we see an ongoing, which is something that It's unsustainable over time. But during his watch, he wants to make sure that, I'm not saying he's objecting, he wants to make sure that he can help where he can in the sense of keeping the interest rate low and can be more affordable for the US government, which ultimately is good for the economy because we already have very little margin for discretionary spending beyond defense. So the country itself is in trouble if we can't finance. I mean that debt problem extends beyond the government. One thing we focused a lot on at Bridgewater and was one of Ray's kind of seminal ideas was on long-term debt cycles. so, anyone in markets thinks a lot about the short-term debt cycle. That's the market cycle. Business starts booming. People are allowed to borrow because they look more credit worthy. That borrowing fuels spending until you get to a point where you overbuild. Some of those ideas go bad. It goes down at cuts rates. And what you see in terms of the debt curve is you get those things from 1955 until recently, every next debt level was higher than the last. And so you did cut debt a little bit at the end of the cycle, but you ended up at a higher point than you started. And what you saw in terms of interest rates was from 55 to 81, the next peak interest rate was always higher than the last. Since 81, until the inflation of 2020 and 2021, each low was then lower than the last. And so you were in the back half of the long-term debt cycle and in need of deleveraging. The problem is we haven't deleveraged. And so across households, businesses, and the government, debt levels stay near all-time highs. Consumers have come down a little bit, but not. Not as much as you would need to really say, hey, we've reset. There is kind of more fuel in the tank for the next one. And Bridgewater, called this pushing on a string of, know, the primary way the Fed makes changes is again, through changing the short-term interest rate. That matters in as much as you actually reduce people's debt burden and as much as you can actually stimulate new credit creation. but if someone had already basically maxed out on their debt, maybe you reduce the debt service costs a little bit, but you're not gonna stimulate that next round of spending. I agree, agree. I just think that if they can extend this, now is no longer trying to solve it, it's trying to buy time and then try to figure it out, right? So it's either one through taxation to reduce, which is not politically convenient, or through growing the economy. In my view, has to be the latter one because... If I know anything about the government's reaction function is they don't use increased taxation to lower the deficit, they use it to spend more. Exactly. And so you have to eventually grow your way out of it. And that becomes harder, you know, back to your starting point here, the more uncertainty there is because growth does not like uncertainty. If I don't know If the lumber I need to import will be 20 % more expensive tomorrow, if I don't know what the price of oil is going to be within some reason, it's harder for me to get a loan. It's harder for me to get comfortable investing in my own business, let alone the fact that I'm coming into that already at an elevated debt level. That's right. Now, fortunately for both of us who live here, America's growth is still the most robust of similar economies. We've managed to avoid a lot of the sort of deflationary true low growth scenarios of Europe, Japan. Um, but, you know, as these problems continue to mount, we get closer and closer to that reality. um, it's interesting when I listened to Walsh, uh, he talks about independence, whereas though I would say half of the audience at least is thinking about his independence from the white house. He is very mummed on that point. He never talks about it. He talks about independence, but he doesn't say specifically independence from what. so it's up to the listener to. to interpret it's independent from the White House or independent from the treasury. And then of course, then he invokes coordination and he still says it's independent. So I try to figure out what does independent really means for him. And of course the opacity that he promotes, it's also suggesting that we're not going to have as much transparency as we would have since the end of Greenspan era. sort of the more transparent era of the last three rounds of Fed. So what do you make of that? What do you make of the opacity? we started our conversation talking about private market being less public information, good and bad. In the public market, it's just purely beta. So how do you juxtapose that with the Fed messaging and transparency? So it's an interesting question, me, I'll use a hypothetical. Imagine we said, hey, you know, all that really matters is something like the Taylor rule. And so you use your difference to unemployment, difference to inflation, you weight those things. We don't need a Fed staff by humans anymore. We're going to hand it to a computer that just runs the Taylor rule and tells us what the interest rate is supposed to be. In some ways it's conceptually correct. What ends up happening though, is that markets move when things are different than expectations. And so if you know exactly what the Fed computer is going to do next month, then it should be fully priced in and there should be no impact on markets. Where the Fed is able to really change things is when they do something that is outside of expectations. And so I don't think we have enough data on Warshett to know if he's playing three-dimensional chess or not, but a generous read would be a more opaque Fed reserves the right to have more market action. And if you're in a period where you think you need to bring that back, you want to be able to surprise markets in one way the other. And so, you know, that's sort of how I'm approaching it of like how much of this is a opacity is for opacity sake and how much of it is a understanding that if we're in a period where the Fed really needs to move markets does surprising the market more give the Fed a bit more power to do so. So when I read it, I read it a little differently, but same idea, what you were talking about is who is the dog and who is the tail, right? So he, as I read it from his various speeches and also the press conference, it's really, he discouraged the idea of giving too much information to the market because then market become the dog and we become the tail. Kind of, I mean, I become the tail and we are the dog, you know, that kind of thing. And we as a Federal Reserve shouldn't be doing that. We should be the one to make that decision and market will follow us. I think that's a 180 from what it used to be. Used to be saying, look, we don't want to surprise the market. We want market to know and not be surprised so that they can have more certainty, which is to your point earlier. And look, in a period of relative stability and high growth, that's probably the right approach. In a period of potential stagflation, I understand the rationale of not letting the market lead you and also wanting to be able to surprise it if needed. But there's a balance there. So we'll see how that work out because I think that market probably tend to prefer a little bit more information. then of course, then he went into intent to, or is in the process of creating five groups to help him make decisions. And he claims it's going to be, I'm sure it is, experts and so on. But my question is, are these experts that who lean towards his way of thinking because he is the one who's going to pick them. And so, so it's really deflecting what he wants to do by putting into these groups. So he owns no decision-making. He said, well, the group decided this and I agree, I picked them. Or what do you think about having these groups and rethinking? communication being one and inflation being, you I won't repeat all five, but what's your thought about something like that? Is that something that you find it really helpful, is encouraged, is a good idea or a mixed bag? think the, what will matter is the execution of that. There's a real problem, which is that a lot of the data the Fed relies on is becoming less and less relevant. Take CPI, even their preferred, know, sticky measure. I think if you talk to If you survey 12 people on the street, their experience of inflation does not equal what that measure says. And with growth, know, aggregates hide just how much there are clear winners and losers in our economy. you know, this is much more clear in places like Europe, like the economy of Italy and the economy of Germany are so clearly different. You can really clearly see. the fact that the ECB has to manage to a rate that works for everyone and probably a little bit more for Germany. Like if Italy could devalue its currency by 50 % and cut interest rates, they could stimulate a lot of growth. They can't because they also have to. Of course. We, I think, look at that less because I think there's reasons why labor is more flexible here. People can move from West Virginia to New York to California. Sure. There's a longer history of us all being on the dollar and intrastate wealth transfers essentially are larger than they are in Europe. But it's a real problem. We have two economies and aggregates tell a very imprecise story. Talk to anyone in San Francisco today and they'll say this is the biggest boom in 20 years and talk to someone in Chicago and they'll say the city's collapsing. And so all that's to say The idea of rethinking how information gets to a Fed chair and to a Fed committee and how decisions are made is probably necessary. Now, I don't like that power. Eventually, we need to figure out what are the new data points we actually trust. If what these committees do is within five years, there's a new set of statistics that we're looking at and we all have access to. If it's of all finger in the wind and vibes based, I think that's more dangerous. But if it's an acknowledgement that the macro indicators that you know, worked well enough for a more homogenous country from 1970 to 1995, really are starting to fray at the seams, then I think that's probably reasonable. So it's interesting because he actually talks about what you just talked about in a slightly different way, that the data they've been using are stale data, this yesterday's data. They really want today's data. Obviously, there's no tomorrow data until tomorrow happens. But you know, more current, shall we say. So for example, I... I've spoken with State Street and they have a real-time inflation tracker that scours the entire internet to figure out the price differential from platforms and so on and so forth. That had been quite accurate in a real-time basis. I think he is very much interested in, if I understand him correctly, is employing more of that kind of data or maybe giving a little bit more weight. than just looking at backward-looking data, which I don't think is wrong. And I know that only because I've spoken with some of the ex-fed officials. When I say exes in last term, that they do use some of that data. Now, it's just a matter of weighting the data, how much emphasis. I believe that's true. I don't know if you have a different understanding. Yeah, I mean, with all this stuff, you just have to continually be testing how accurately it predicts what the next... systematic print is. But I've seen things like geospatial data on cars on the road can be a very useful indicator of employment. We rely on lot of laggy and somewhat imperfect survey data right now. There's probably cleaner ways to capture that from real-time economic data. And again, I think there's both the which data points you trust problem, then I'm also thinking a lot about what is stated of this like, we're not going to track every individual in the United States, but also the aggregate of GDP masks, a very complex underlying story. Um, can't leave our conversation without touching on AI because I think that's something that, um, you are, you are insider. So, And, before I get there, um, the, this economy, uh, the resilience of this comedy and the growth, and we talked about all that. A large portion of it is the amount of investment that we're making in AI. The CapEx, I mean, if you look at CapEx, mean, almost 90 % is from AI related. I don't want to just say pure AI, but AI related spending and projecting forward even in the next couple of years, this trend doesn't look like it's going to end. And so that's one thing. Then we look at what the Chinese have done. is apparently, well, number one, their energy need is a little different than ours. They have a lot more renewable and all that. But let's not talk about that for a moment, but really thinking about the, according to them, if we believe it's true, is that they needed less investment to create somewhat comparable outcome. Now I'm going into an area that I'm not an expert. You are. Do you see that we are investing the amount that we're investing forward? because we are still looking backwards to go forward that that's what we needed. Those are the amount of money we need to spend. We need to build this. We need this, this, and this is based on a set of assumption that has proven to have worked so far. But do you see the future could be different, more like a lower energy requirement to do some of the same type of output than have we not potentially overbuilt, over invest? Just curious. Yes, a lot to discuss here. Where to even begin? know, what I'll take, I'll take the latest Chinese Kimi release. All I've talked to a lot of people in security and kind of forensic computer science. All indications are that they've essentially used Claude Fable to train that. And so if you already have the Frontier AI model, you do need a lot less compute. to build one that responds similarly to it. That is true. But we saw this with the DeepSeq release last year. Eventually people realize what those limitations are, and then the frontier models come out with the next generation, and it takes time for people to catch up. I think there is... leaving aside IP law and what's really going on there, some healthy competitive pressure to that. And we should always be thinking, you are there more efficient ways to be training these, to be stacking GPUs, to do all that stuff. But what I'd say is, compute is not getting less important. And even if you could, you know, given that it exists, train a Fable level model on less infrastructure, you know, then the question should be, well, what could the Kimmy people have done if they had all the compute resources they want? Like. This is what they're doing with their hands tied behind their back. And most people are going to want the maximum thing available. Maybe the new models are good enough to do 40 % of software engineering stuff, but there's, what could we do with more? But zooming out to your point on CapEx, I think this is a pretty common story that is certainly not limited to AI of like, we will overbuild compute probably. not every infrastructure project being launched today will end up having a positive ROI. I also don't think they'll all end up having a negative ROI, but like certainly they'll be overbuilt. What does that mean though? That means that compute becomes cheaper for everyone. And so when we go through these phases and this was true of the railroad boom, it was true of biotech build out. It was true of the initial original IT wave. It was true of cloud. There are positive externalities to a lot of parts of the economy, even if those specific. projects don't have positive ROI. We will be left with more compute than we had before. again, sorry to those people who invested in those projects if it doesn't work out, but someone will use that computational power and it will now be cheaper for everyone. so we're all looking at how expensive token costs can be today. Well, if we had 10x the server capacity, that cost will come down. Where I get most excited about AI, and this goes back to our earlier point on how do we grow ourself out of the debt is not on necessarily the frontier side, but it's on real world applications of AI in pretty boring spaces. And so this is, you know, we've worked with a company that's automating healthcare billing procedures. This is almost a trillion dollars a year is spent just figuring out between the insurance company and the healthcare system, how much people are supposed to pay for each procedure, people faxing stuff back and forth. crazy. They've been able to cut it's only in a pilot and about a dozen healthcare systems. They're not cut that cost by 80 % and cut the time to payment by 90%. Yeah, that's real economic savings. And if that gets cheaper for that company to run because computers overbuilt, like we all benefit. Healthcare gets cheaper, more efficient, better for everyone. In things like construction and even just permitting, but also we invested in a company that's doing autonomous construction equipment and in logistics and but also in like more exciting applications like cancer diagnostic and stuff. This is not someone cranking out to a not very good novel using AI. is people's lives being saved. This is healthcare getting cheaper. And so we get really excited about what we call boring AI, of how can we apply this to legacy industries that have been pretty immune to innovation for 30 plus years. And I think that if one thing is gonna help us grow out of the debt, it is revitalizing the core industrial base and the boring parts of the economy using a tool that really nothing like it has existed before. So you are basically talking about how we bring inefficiency to become more efficient as a result, saving money, be more productive from that step, productive not in an economic sense, but more productive as in, you know. and you this in the core macro data, right? If you look at inflation, even with the most recent inflationary spikes, by and large the story since 1980 has been services inflation is above target and goods inflation is well below target. I mean, the computer either of us can buy today, if you do it in raw compute space, just quality space, the size of the television you can buy for $300, like all those things are dramatically cheaper. What we haven't been able to fix is how expensive services economy things are. And, you know, especially in areas like education and healthcare, that has a real impact on people's affordability of life. If this helps us bring that down, You solve the inflation problem and the growth problem in one swing. So, and when you look at that, you, think of in two ways. Um, um, I'm totally with you because I, I expect to be overbuilt. Yeah. I cannot tell you which one it is and which state or which program, uh, that's just human nature. That's not, that's nothing just specific to AI. That's the way we are. We tend to get too excited going one way and then get too excited going the other way. And the boom and bust goes on. So yes, I think that the net effect are the users are going to benefit, but investors may not, depending on which one you pick. investors are saying, look, we are taking risk and we know they're going to be downside. Well, the rational ones understand that and they're to be upside. And sometimes we don't win every single thing. But so, there's a natural check to it over time. But the expectation continues to be forward. Part of why we are very clear with our clients that you don't want to play this just on the public side or just on the private side. Just buying NVIDIA is not a capture of AI's impact. It's a specific part of the AI stack. We think about it as a base layer of basically plant, property, equipment, and energy. Then there's the hardware. Then there's the core models. Then there's the software layer. And then there's the consumer layer. And we spent a lot of our time figuring out kind of what's the picks and shovels approach that's going to work regardless of how any one of those buckets works in the next year. And then a lot of that software and then consumer layer of thinking like who actually benefits if we overbuild and things get cheaper, what parts of the economy can be transformed in next five years that are largely flying under the radar of a lot of VCs. They're not big, sexy consumer products. It's, no one wants to say, do I do? you know, better deed management for properties. But if you get that right, it is a very large industry and it helps a of people with housing. So, you know, we're looking for those kind of slightly under the radar, less overpriced opportunities today. Yeah. Or they are not being focused on. Yes. find it, you know, that's what the alpha is, right? I mean, to some extent. Yeah. It's one form of alpha. And it's again why we're a little less worried about the overbuild because a lot of where we're investing You're not in that side of business. you're not in that kind of business. And those people, their margins will go up as compute gets cheaper. So one other thing that you just immediately talked about, which I like to sort of just touch on and see what you're thinking. I heard recently that some people compared the US and China's approach to AI in the following way. China is more about human replacement robots, humanoids, and all that. I'm generalizing, obviously, that's not complete. US is still more supplemental, augmenting human work and so on. You believe that's true. That doesn't matter if China's right or wrong. I'm saying, do you believe that that's the direction that most of the US AI industry is really looking at supplementing rather than replacing? I think that is, yeah, as you said, obviously with exceptions, a general heuristic. And I think, you know, obviously there are exceptions like what Tesla's trying to do with robots. There's a couple other robotics companies. I think in, again, in general, a lot of the value we can extract from AI in the next five years is going to be from empowering a human in the loop, not replacing them. I'm not a believer that AGI is coming in the next three to five years. We don't know what consciousness is in humans. I don't think we know how to replicate it yet. But if I can make that smart human 10 times as efficient tomorrow, that's where a lot of the benefit will come from. I we do this internal opto, like we're not reducing our, we have a very seasoned and top Silicon Valley engineering team. What's amazing is that our top engineers can now ship 10 times as much code. They're still doing the reviews. It's still at their code quality bar. We're not saying we don't need developers. We're saying we actually need better developers because the best developer with the right tool is something that is really unique now and is getting more and more powerful. So let's say I have three teenage kids. I don't, but let's say I have three teenage kids and I introduced them to you, Jake. I said, look, you know, talk to uncle Jake because uncle Jake knows a lot about the present and projecting forward. I don't, I'm not in that business. And so they come to uncle Jake and say, uncle Jake, what kind of major should I be in? What kind of courses should I be taking? What is the future for me look like? What is my career? Just think it's true, not right or wrong, but you how should I think about it? What would you say to these young, young folks? Yeah, I would say a few things and I would first say, you know, I hope that I'm right 51 % of the time. take a rest. Me too. But, um, I think that we will see a revitalization of the importance of the humanities, um, and like writing well and communicating well. because, I mean, we see this internally, we have some great, uh, content and writing people and staff like Kristen here, like taste is pretty irreplaceable still. And what that person who is really good at communication can do, building a couple skills files for Claude, using that to create a lot more content than they did historically, but not removing their human judgment and taste from that equation. Also say like, I don't think a lot of people are good enough at writing today. I think more people should focus on that. Two, I think that really learning to think in systems. And this is part of my view, what becomes less important is being a junior front end developer. Cloud is already really good at that. What's still irreplaceable is someone who really deeply thinks about architecture, data management. How do we make all data within a firm accessible? How do we build a rag so that you can access it without burning through tokens? How should security work? Security is getting so much more important as attacks get more intricate and more robust. And so, you know, kind of a double thing there and maybe I'm just... Stem and English so far. Yeah, and then, you know, I don't know, I'm not sure what this means for that group of teenagers, but a thing I'm thinking a lot about is how much trust matters in this world. You know, before, even before, let's go roll all the way back to before computers. For someone to, you know, type up a 300 page document, you know, on a typewriter or by hand, that's a lot of effort. There's so amount of trust that goes with, well, you put all this effort in, you know, maybe I take a look. And so when you saw a magazine on the shelf, you knew how much work that went in, you know, how hard it was to get ads in that. You know, the internet was the first way of removing that trust. It was very easy to create a listicle website and get some shady advertisers. And, know, those of us who grew up in it, know not to trust anything on the internet, but, you know, maybe actually my grandpa was very savvy, but I've met other grandparents who it's like, oh, you you're coming from a world where if someone was in the printed word, you could trust it. That's not the world anymore. Now it's even more extreme. You could have a farm of bots create an entire website and 95 % of the users can be bots. And you've sort of flooded the world with content. And I mean that generally, not just like written content, but all forms of everything is just flooded. And what does that do? I think it really increases the value of curation and trust. but I don't think we have what the next set of trust looks like from an organizational perspective. think that's gonna be a huge area of growth in the economy in the next 10 years. I just don't know what it looks like yet. But like I find most of the internet unusable now. And I'm waiting for whoever comes next and comes up with like, how do I trust that what I'm seeing, and I'm not saying it has to not be AI generated. AI is really good at some things, but how do I trust that I'm doing this for the right reasons? Yeah, yeah. Fascinating. I take the same direction that you. mentioned, STEM obviously, data and math and all that is foundational. English and math are the two things that you always should have. At the same time, you look at how many liberal art colleges in Pennsylvania and all across the countries are having real financial problems and closing down. At a time, think we've done ourselves a disservice of there is a big difference between an easy English degree and a difficult Classics or History degree. Not all, this is not just, hey, liberal arts are good. It's, you know, that process can be made really robust or it can be the easy path. And if you're willing to put in the work to really getting good at that craft. And so I thought of one more thing I'd tell those groups. I think we're on the verge of a true healthcare and biotech renaissance. I think the impacts of AI there will be extremely creative and the data we're collecting on how to improve healthcare outcomes is pretty amazing. and technologies like CRISPR are getting more mature. So I'm very interested in how technology gets applied to that space. So I'm thinking about courses that teach how to think, because that's the part that I always say that AI, the A part we got it pretty good is the I part that we still need humans. And how do we create more I genetically? I don't mean genetically like having more children, you know, as humans. And I don't know what courses that is. And I don't know how to pick a major or pay a concentration that really focus on analysis and thinking and so on and so forth. Well, to your point, it's not that it's a discipline, it's a frame of mind within any discipline. And that comes so much down to the quality of the teacher in that context. Are they pushing people or just saying, believe me? Believe me is a particularly dangerous sentiment. Trust me, trust me, Koucher. We're back to the trust thing. Yeah. But that is, I mean, I... I went to a small high school in San Francisco that had, didn't have any grades. We all still got into good colleges, but the point was like, you're not here to prove you checked the box. You're here to learn how to think and like, it will be distracting if you're just like, am I right? And that's really, you know, I learned more in high school than I did in college for sure. Because it was not about, I think about this as exercises versus problems. I studied economics and classics at UChicago. I loved my classics major. I honestly found the economic side. underwhelming because it focused on exercises and I think exercises are things where there is a defined solution. It can be wickedly complex. were doing some, you know, we'd ported over Hamiltonians from physics and it was, you know, 20 pages of problems at work, but there was a right answer and you just had to find it. But most of the world is problems. Problems are things where you have incomplete information and there's no, you won't know if you're right for three to five years. Well, it's testing how to, I mean, it's about judgment. You know, how do you, how do you do that? But Jake, I mean, I can go on for another hour. I know you are probably going to run out of time and I really enjoyed it. think this is not only timely conversation, but really we talked about polycrisis when we first started is that they add to each, each layer adds to it. And the outcome is somewhat unpredictable. And it's very hard to navigate either as an advisor or as an individual who's making decisions. And that at the end is based on our analytical skills and our ability to understand the world. So thank you for sharing your insight and I have really enjoyed hearing you and listening to the way you think about it. You're at the forefront of a lot of these types of stuff. So thank you for being my guest. Thank you. It's great. You can always find more episodes by visiting philipchao.us slash podcast or find us on your favorite podcast app. You can always leave us feedback, ask questions, or request a topic for us to discuss by sending an email to pc at philipchao.us. Views expressed in the Tao of Chao podcast are individual opinions and they do not represent the employers of each guest or the firm each guest is associated. Our podcasts are for educational and informational purposes only and should not be deemed or viewed as investment advice or recommendations. Please consult your personal financial advisor. investment expert or investment fiduciary before taking any actions about your plan and investments.