The Real Ticking Time Bomb with Michael Cembalest

I feel like there is deflation in summer. It's getting shorter. Hold on. It's the middle of July already. John. What's wrong? Oh, no, you know what? I think you might be backwards. He's backwards. He's backwards. No, that's right. That was right. Is it? It's a little loose. No, it's not right. It's the cord is on the left side. Come on, Nicole. Oh, okay. I know I just did this. Yeah. Sorry. There we go. Well, we lost May. Mike and I were talking on his birthday. He used to be able to fish in shorts. May is gone. Yeah, May is gone. May is very cold now. Oh, in New York, the thing that people don't realize is September, October is better than April and May. Weather-wise. Yeah. And people forget. Like people get excited. Oh, it's spring. Spring is not great in New York. No, it's not. Fall is wonderful in New York. Yep. What is the Latin term that you use to describe the eagle octopus? The Aquila ceph. So Aquila is Latin for eagle? Yes. Okay. And cephalopod is in. Okay. All right. which is the prefix used in Latin and in biology. It's a little bit of a disturbing image, but it catches your attention for sure. Yeah. The GPT is amazing now in terms of you can give it very specific guidance in terms of what you want it to draw. I was going to ask if this was AI generated or you had an artist inside of JPM. We used to use external artists, but it's just so much easier now. And even compared to a year ago, the degree to which you can continue to layer on additional instructions and have it create whatever image you want. How much does that frustrate you when people are talking about or complaining about what AI can't do when it's evolving so, so fast? Like we're looking at this in a vacuum as if it's a finished product sometimes when people are talking. Yeah, look, I think I like the fact that when people do empirical analyses, they track how it's doing. you know it's still it's getting better at certain things but you know think about it this way when the cost of being wrong is very high i think the hurdle should be very high uh you know there were there was a time a few years ago when people felt like radiologists wouldn't be needed anymore if you if you are at risk of breast cancer you still want a doctor plus ai instead of ai by itself um and so it's it's still not ready for prime time on certain interpretive things in medicine but it is getting better at things that have low value like making drawings of i actually eagle octopi i actually like that it's not perfect because i like to yell at it i get a lot of satisfaction i can't say the things that i can say to my uh claude to other people so i like to just put it all there and i find it to be uh or ai well i'm i'm definitely kicked off of this this claude i'm not easy i'm not easily kind of surprised but i am very surprised about the number of people that use it as a psychiatrist or a counselor or for for marriage advice and things like that my wife my wife is talking to it about nutrition all day and used to pay a nutritionist like go every week here's what i ate here's what i should have ate here's the choices that i made at this like italian restaurant that makes sense that's mostly She much prefers talking to Chad GPT as an – it's not that it's her nutritionist. It's tracking the minutiae of what she eats and doesn't eat each day and she just likes that experience better because it's whenever she feels like it. It's not I have an appointment. It's like anytime I feel like talking about what I'm doing, I could talk to it. I mean these things are slavishly trained to be so obsequious to you that you have to be careful exactly what you do rely on. Although I do have this – like a lot of us, I have this fancy Manhattan internist. And whenever I call him and I'm describing my symptoms, I can hear him typing them into the computer and giving me a readout from Google overview. Well, I ask you for relationship advice. That seems a bit – I know people are doing it. It seems a bit strange. Yeah. Yeah. Yeah. I mean, maybe that's one of the reasons that you're having a problem in your relationship to begin with is that you kind of are trusting in a computer. But I think it's the convenience of always on more so than it's I trust the result from this more than I trust the person. And so professional services organizations are going to have to find a way to take their own internal knowledge base and put them in a setting where somebody who wants to access the way they would have answered a question can be 24-7 always on. And it's not really replacements and extension. One of the things that's amazing, I mean, I'm sure we'll get into this when we get into the meat of the discussion, is free open weight models trained on a company's own data can now outperform the frontier labs, which are much more expensive. And so that's kind of a huge deal. Bridgewater published something recently showing that an off-the-shelf completely free model had better results at maybe 5% of the cost of the front of OpenAI. I read what they said. So they are feeding all of these economic data releases and news articles and research into this thing. And they're trying to see how valuable how valuable each different item is. And they're trying to like weight the response of the A.I. to these things. It's designed to collate all the information that exists overnight and create a reading package for the portfolio managers. Yeah, pay attention to this. Don't pay attention to that. And so, you know, they're using some qualitative judgments. But what they're finding is that the free model trained on their own data does a better job than some of the frontier labs, which are much more expensive. I mean, this is something that we're going to have to watch as we start thinking about where we are. Just over the last 48 hours or so, some data is coming out of OpenRouter showing that almost 50 percent of all of the tokens from U.S. companies are being directed to Chinese models rather than the U.S. frontier. More efficient. Less token per use. Yes. Yeah. Like if you're trying to figure out a cost per performance, those models are cheaper. I don't think 12 months ago or even six months ago, that was going to happen quite so quickly. What is the reason for the superior efficiency of the Chinese model? They're just distilling the US model but using less inputs? They have no MSG. No. So why would somebody route something to a Chinese model and why would it be more efficient to do that? Well, I mean the bigger question is why is China spending billions of dollars creating models that they're giving away for free? I don't think there's any question that the performance of those models every three to six months closes the gap with the US ones. There's all sorts of data from artificial intelligence and ELO scores and all of the different benchmark GPQA diamond scores. I mean there's a whole cottage industry of assessing reasoning, math, coding, language capabilities of these models. And the GLM 5.2, which is the new Chinese one, is a hair's breadth away from GPT 5.5. Not noticeable for the average user. For the average user, it's perfectly fine. It's perfectly fine. Yeah. And so the question is why are the Chinese companies putting so much money into things they're giving away for free? Thoughts? Well, there's a lot of debate about this. People inputting data there. There's a lot of debate about this. I can't escape the fact that the most valuable companies in the U.S. universe are either the frontier labs themselves or the companies that sell hardware and equipment to those frontier labs. And if you're China, if you can stick a pin in that balloon, that achieves probably more of your long-term objectives than a lot of other things you might be doing like industrial-listenage. Oh, so you think it's a – they're rendering the advantage that we have less valuable. Oh, my god. But our market cap keeps growing, so I don't know if it's working. If China could wake up one day and figure out a way to make the mega cap companies worth a third of what they are right now, they'd do it in a second and they'd spend a lot of money doing it. Now, you can also make some arguments that they're trying to create an ecosystem within China that relies on their own compute and things like that. It's probably a little bit of a mix of everything. But it is remarkable to see these companies creating and providing these free tools that you can now run on your own devices. You don't necessarily – none of your data has to be routed to Chinese servers. You can run the whole thing in-house. You can even run it offline if you've got a powerful enough NVIDIA chip on a laptop. So it's – there's – I can imagine at some point that the Department of Commerce is going to have to start thinking about the impact of those Chinese models and do they want them proliferating in the same way that Europe is trying to figure out whether or not they should put a cap on the importation of Chinese EVs, which are decimating the European auto industry. As I say, the Europeans are still trying to agree on a definition for AI. I remember from the social media age, the question was, why are they giving away Facebook? Why are all these new products being created and given away for free? And the answer is, if you're not paying, you are not the customer. You might be the product. They're not giving away anything for free. They're serving you to advertisers. There's probably something to that where it's more valuable. It's more of a strategic priority in China to have U.S. and other international users putting information into these programs. It is an answer to one of the other things you mentioned. Look, I talk to a lot of people. This is way above my pay grade. I have to say of all the technological revolutions that have ever happened, this one is the hardest to kind of understand the bits and pieces of. I mean compare this to the fiber build out and the internet and B2B software. I mean those are pretty simple. compared to this but the people that i talk to that live in this world say that you know a small component of chinese success has come from training their models on u.s models yeah um but there's a lot of independent engineering that they're doing on our end as well and we shouldn't we shouldn't underestimate the degree to which the controls that have been put in place by the u.s have forged a an ecosystem in China that is determined to figure out its own way of building these things out on its own. So they now have an alternative to the ASML lithography machines. They're not quite as good, but they're getting there. Yeah, I think I think it's pretty obvious that you're right about that. And, you know, Apple's got a Apple's got a massive business in China. Their U.S. partner for AI is going to be Gemini and be Google. That's not going to work in China. So they have a deal with Alibaba and they're going to use models that are coming, I guess, from Alibaba. And it will be an entirely separate AI ecosystem that Apple utilizes there versus here because Google can't operate. All right. Let's go. 250. All right. Michael, it's the 250th anniversary. It's a piece. Our episode. This is our 200th anniversary. Hold on. Yeah, we lined it up. Come on, don't move. Hold that up. Good job, Mr. Cole. Whoa, whoa, whoa. Stop the clock. Here's a word from our sponsor. This episode is brought to you by Federated Hermes. Active ETFs are changing the way portfolios are built, giving advisors more flexibility for their clients. But not all ETFs are built the same. Federated Hermes puts the investments in their active ETFs through a ruthless vetting process, giving out a wide range of market scenarios so only the strongest survive. The result, a suite of 12 active ETFs spanning the full stock and bond market. Whether you use them as core building blocks or tactical allocations, you'll get the strategies you want in a convenient ETF wrapper. Simply put, Federated Hermes has the active ETFs to help you build portfolios designed to last because they've been vetted for it. Explore the full lineup at federatedhermes.com. ETFs are subject to risk and may lose value. Federated Securities Corp. Distributor. Before investing, carefully consider the fund's investment objectives. risks, charges, and expenses. Read this and more information in the prospectus or summary prospectus available at federatedhermes.com slash us. Welcome to the compound and friends, all opinions expressed by Josh Brown, Michael Batnick, and their castmates are solely their own opinions and do not reflect the opinion of Ritholtz Wealth Management. This podcast is for informational purposes only and should not be relied upon for any investment decisions. Clients of Ritholtz Wealth Management may maintain positions in the securities discussed in this podcast. It's the 250th edition of the Compounded Friends. We are still celebrating the 250th birthday in the United States of America. And we are going to cover a very special piece of research that has come out over the last week from our special guest, returning champion, fan favorite, the legendary Michael Semblis, ladies and gentlemen. A little more. The crowd is going wild. This man walked through the rain from 47th Street to be here with us today. All right. Michael is the chairman. It's four blocks. Yeah. Michael is the chairman of Market and Investment Strategy for J.P. Morgan Asset Management where he leads strategic market and investment insights across the firm's institutional funds and private banking business. He is the author of Eye on the Market, I Never Miss It, a widely read commentary covering markets, economics, energy, and policy. Previously, he served for eight years as chief investment officer at J.P. Morgan's Global Private Bank after joining the firm in 1987. Michael, welcome back. So happy to have you here. How are you feeling today? I'm okay. Are you a big patriot, medium patriot? Where would you put yourself on the scale? Like are you like Mr. America or like where do you stand? It's not a trap. I'm genuinely curious. It sounds like a trap. No, no, no. No, I thought you were a patriot fan. No, no, no. The answer to that was going to be absolutely not. No, no, no, no, no. I know you're not. But yes, I am a patriot. Okay. I am a patriot. Yeah. I ask that knowing the answer. The tone of what you write over years is very much debunking people that talk about South America. It's very much about the innovation in our economy, the dynamism, et cetera. And obviously I know that's JP Morgan's brand new building. They've got an American flag glowing in the breeze in the lobby. I've heard Diamond's stump speech in person many times on TV about why this is the greatest country in the world so I just wanted to kind of get you on the record as a patriot and someone who loves this country but I'm you know I'm analytically I'm a patriot yes this is a country with the United States has the highest ratio of people immigrating here versus the people that leave still still than any other country in the world and the data that I update every year with world bank data so you know the people are voting with their with their feet so would you say you're like an 8.7 or so like are you a 6.9 yeah look like a lot of people like my relatives came here built alive for themselves and uh so it still works all right so let's put this out warts and all and look and and in that in that piece um and you know i i i included a couple sections at the end that my biggest long-term concerns in addition to the to the unsustainable level of federal debt are issues related to the rule of law and what's going on with federal funding of medical research. Let's put this image up to set the scene. You titled your special edition of Eye on the Market, Semi-Quincententacles. You got it. So it's the semi-quincentennial. Right. But the point that you're making here is despite all of the doubts and concerns over the years about when is it not going to be America's moment anymore or who's going to replace them or what's going to replace the dollar. Despite all that, our tentacles are still very much wrapped around the globe. Yeah. I mean, our clients have benefited in innumerable ways from an asset allocation approach that more or less since 2008 and March 2009 has been overweight. The dollar, U.S. credit, U.S. equities, U.S. real estate, U.S. infrastructure, just across the board. And there have been blips in between. But I think there's – people always seem to be looking for opportunities to say take profits and so far that hasn't been the thing to do. When you run the numbers, being long, the U.S. has been enormously profitable. The only thing that I've ever seen that matches the profitability of being long the U.S. over the last 15 to 18 years was being underweight Japan when it crumbled from being 65 percent of global equities in 1989. That's the only thing that compares. So what makes it hard to be long the U.S., long the dollar, long real estate, long U.S. stocks? what are what why do we have so many people who are like knee jerk up that's it that's the top for america or well look because because on a cocktail napkin the the p e ratios look very high and they are very high but there's a bunch of other things that you have to look at which is okay but what's earnings growth doing and if you actually look at the sectors that have the highest valuations they have the highest earnings growth so i mean a lot of your listeners are familiar with a concept called the peg ratio, which is just a PE that they all love to look at and earnings growth, which they love to look at separately. Why not merge them? And if you look at the peg ratios for the US, they're not that out of line compared to the rest of the world, even in the tech space. Who said the quote I don need to pay an analyst to tell me that a stock at eight times earnings is cheap I need somebody to tell me when a 40 times earnings stock is cheap i love that no it expensive no the opposite the opposite when eight times is expensive and 40 times is cheap yeah i don't know who said that quote it's but it's it's a wonderful one right michael my favorite part about i'm trying the dollar too i mean we'll get into that but like the dollar is everybody's punch it's the dollar it's the dollar the deficit the unsustainable nature of of all of this in fiscal policy and monetary policy whenever people talk about debt to gdp and all the problems that they have with what's going on. They never mentioned an alternative. Okay, fine. So let's just humor them, right? Let's just say that the US is on the ropes and it can't maintain its reserve currency status forever. What's going to replace it? And you did a very good job not only lining out why the US dollar is still the dominant currency, but all the reasons why maybe the Chinese alternative is completely absurd. Yeah. Yeah. Well, look, a day will come in our lifetimes when the piper has to be paid on some of these things. And we can talk about what that might look like. But my job as an analyst in managing the firm's money, and we oversee $4 trillion on behalf of institutional individual investors, is to make sure that we don't run for the hills too soon and leave enormous profits on the table. That's what's driving this. So the first discipline you have to have is everybody thinks that the dollar is due for a collapse. Are there any signs it's happening? Are there any signs happening right now? And so we have all of these metrics we look at in terms of the dollar share of world trade, the dollar share of foreign exchange allocations, the dollar share of SWIFT payments, the dollar share of international current denominations. And they're all stable. So it's the table on the top there, the dollar share of cross-border loans. What you're showing is year by year, there's really no change. But not just stable, dominant. So yes, maybe one day – yes, dominant because the US is only 20 percent of GDP and most of these numbers are much higher than that, roughly double. So maybe one day that will change. But I'm always struck by the lack of the discipline for the dollar bashers to even acknowledge that none of the pillars of the dollar as the world reserve currency are shifting. And it's kind of easy to come out and give like a Rubini-esque trashing of the dollar. But you should have some discipline to say, are there any signs that it's happening? Who's getting share on the bottom left? Nobody. Well, it's not the yen, the pound or the yuan. And the ones on the right are currencies that are – I mean these are very small. They are micro currencies. None of these – so none of those could be reserved. So Japanese yen, great British pound, Chinese yuan. There's a lot of stability in this chart at least over the last five years. nobody's really gaining anything right is that okay yeah and and you've seen small allocations into the sing dollar norwegian krona stuff like that but the bottom line is that the the things that make the dollar the world's reserve currency aren't really changing and that's that's the important takeaway and like maybe one day that will change but so far there's no sign of it so the last thing we're going to do at this point is make major structural changes to the currency overlay in the portfolios that we manage yeah so one of the things that the dollar bears harp on is the buying power of the dollar they act as though things don't cost more in other currencies also is point one and then point two they'll say oh the dollar has lost 96 percent of its purchasing power over the last 50 years whatever they say um which it's a tautology yeah the price of everything has gone up and okay but then the problem is so you say to them okay well those are the same people though that also would tell you to stay out of the stock market You can't be a dollar bear and a bear on stocks because stocks have more than helped you keep your purchasing power and it's the only answer. Yeah, your dollar does not belong under your mattress. Real estate and stocks are the only answer if you're worried about the purchasing power of the dollar. Now, if you're worried about US fundamentals over the last few years, the places that people have gone is gold and crypto. And so one of the things that we watch is that I think is a key arbiter is when is the GIC, one of the big institutional investors to Masek and the big sovereign wealth funds going to make strategic, not just tactical, but strategic allocations to digital currencies in their foreign exchange reserves. That would be a seismic shift. And so far, for the most part, that has not happened. I also was fascinated to see Warsh's initial public statements because – I mean think about this process. Trump has been out saying that he wants the funds rate below 1 percent by the end of this year and other similar things. And you can kind of see a confrontation brewing that rhymes a little bit with the confrontation that Nixon had with Arthur Burns. and but Warsh has kind of come out with this with some statements that reinforce historically that he's a hard money guy that's not going to get pushed around. Cal, she is at 53 percent chance of a rate hike this year, which is, I think, roughly unchanged from before Warsh actually got in the seat. And so we started the year with the markets pricing in cuts. So, you know, so far he and I don't think it's a coincidence that some of those comments have coincided. with some of the steam coming out of crypto and gold. One of my favorite charts, speaking of gold, you see this all over the place. The amount of money that central banks have in gold. Let's do chart four, Daniel. All right, well, maybe this is a price issue. So let's say you have a guy and he's got $100,000 in cash and a house worth $100,000. And all of a sudden, Mark Andreessen comes along and wants to buy his house for a million dollars. His asset allocation now looked like it's 90 percent real estate, but that's simply because of a price change, not because he bought more real estate. That is a very simple metaphor for what's happened to the gold market. Gold prices have gone up. So the value of gold and foreign exchanges has gone up. The actual allocated troy ounces of gold that these – the central banks own has fallen in valuation terms and that's the gold line in this chart. So gold has gone from like 1,500 to 4,000-something an ounce. And that is 95% of the reason why gold as a share of foreign exchange reserves has gone up as much as that. It's not because the central banks are kind of chasing it and hoarding it. But they write these articles as though the central banks are loading up on more gold. No, all that's happening is the gold that they currently own. I mean is it any surprise that the kind of gold bugs are misrepresenting the data? No. It goes with the territory. I can't believe it. So this is amazing. You wrote that. All right, so China's maybe another potential, right? The Chinese economy is huge. If it's not us, perhaps it's them. You said if China fully opened its capital account, possible outflows could crush the R&B and trigger collapse in Chinese equity real estate markets, even with the controls. China reportedly experienced a record $800 billion to a trillion in capital outflows in 2025. Yeah. I mean China does a lot of amazing things, right? There's a lot of innovation in China. They're catching up to the US in terms of patents and all sorts of experiments. And so when you say bad things about – when you say critical things about certain aspects of the Chinese economy, there's a lot of – like every country, there's good things and bad things. One of the undeniable aspects of the Chinese economy is in addition to personal surveillance, they have monetary surveillance. You can't get money out of the country. and the M2 to GDP ratio in China is off the charts compared to other countries. And that was one of the key charts. M2, the money supply. Right. So you're looking at – Because it can't leave. Yes, not this one. Okay, so maybe we – But let's look at this one. Let's play – for many years, my wife was on the board of Sesame Street. For 20 years, she loved the show and she kind of used – Fun fact. Yeah. Well, she used it to kind of teach me spelling. And so let's – in Sesame Street, they used to have – one of these things is not like the other, right? So that's what this chart is. So you don't have to be a macroeconomist to understand this. This looks at the money supply, central bank assets, and foreign exchange reserves for three different kinds of places. The first one are pegged currencies like the Hong Kong dollar and the Singapore dollar and things like that. They peg the value to the US dollar. Right, and they manage the money supply to keep that in line. And then the middle one is – and by the way, the first one also includes some of the Gulf currencies that are pegged. The middle one is the Mexican peso and the one on the right is China. So one of these things is not like the other. And so in China, the growth in the money supply has massively outstripped the accumulation of central bank assets and foreign exchange reserves, which tells you if they ever loosened the foreign exchange controls. That's all excess money that would immediately leave. Immediately leave, which could crush – and they would never do it. It goes into what it's already secretly going into now, Vancouver real estate, things like that or what does it do? All of the above, right? I mean there's no – I mean I don't think it's really clear where exactly it would go. We just know that it would go someplace. OK. And so – and the reason I put this in there is any – the definition of the world's reserve currency is that you don't have that kind of sword of Damocles hanging over your head. And so China is not a candidate to be the world's reserve currency if they have this kind of structure. It would go into dollar- How do they not have – with that explosion in the money supply and all of that money captive in the Chinese economy, how do they not have significantly worse inflation? Like don't we think that – Well, it's the real estate collapse. People are – So they have enough deflation in real estate that it's stopping that money supply growth from becoming inflationary. And remember, there's massive amounts of Chinese saving in fixed income instruments and things like that. OK. Yeah, like insurance company contracts. China is like the United States 100 years ago. There's little to no safety net in terms of Medicaid, Medicare and things like that. So people have to precautionarily save for their own time. Hence the high savings rate. And they're having increasing life expectancies and no safety net. So people have to precautionarily save to do that. OK. Here's the other one. they're dumping our bonds slash they're going to start they're going to stop buying our bonds the global bond vigilantes we went from China owns us was the thing that late night comedians would say on state like we're debtors to China that was the problem now it's the opposite China's selling all our bonds they're dumping our debt what do they know that we don't so let's tackle this concept so Again, like the premise that global investors would respond negatively to this administration and its policies is certainly plausible. There's just not a lot of evidence that that's actually happening. So I don't start out with a kind of a value judgment or a bias against some of these statements. All I want to do is see whether or not there's any reality to them. Yeah. And from what we can tell from the chart on the upper right and on the bottom, foreigners are still accumulating treasuries and T-bills, corporate bonds and equities at roughly the same pace they have been. But the share is going down. Why? Because we're buying more of them? No, because we're issuing that we're issuing debt so much more quickly. So that's so that's more of some. Yeah, it is. It is. It's a lot of debt. So the share has gone down. But Russia so far is really the only country in the world that said we're going on a bio strike and we're going to sell our treasuries and agencies. There's no evidence that that's happened. People always get hoed-winked. The Chinese numbers that are officially reported have gone down. But a guy named Brad Setzer at the Council of Foreign Relations has done all the work on this. They're buying a ton of them through this Belgian Euroclear account. And when you adjust for the two of them together, they're treasury- They have to buy them because they need – They don't have much of a choice. They need – first of all, you run out of places for the money. Where are you going to put it? Yeah. But second of all, they conduct so much trade in dollars, they need to have access to it. Well, essentially they run a large capital account surplus and so that capital has to get invested someplace. Even if politically they wanted to do that, there's a limit to what they can do. That's right. Okay. Should we care whether or not China is holding our bonds, buying our bonds? Is this a thing that people – I think so. We should – like investors should care. Yeah, because it's unclear. I would rather have that than China decides to bail out of $500 billion and to prevent yields from going up. Besant has to strong arm the banks to buy them, which is how the European banks function. Like I'd rather not have the US banking system acting as a backdoor placement agent for US trade. Is there a danger of that if the tariff rhetoric gets worse or like is it – is there a legitimate danger of that do you think? China is buying – as you mentioned, China is buying all those treasuries and agencies in its own self-interest. Nothing to do with us. And so – and if that self-interest ever changed, they would allow them to amortize down and or sell them. And then we would have to see how much other demand there is for that long duration. Have you looked at like the sovereign wealth funds from the oil producing countries and whether or not they're potentially taking the place of the Chinese buyer? The pace of buying is roughly the same. There are no major changes there. OK. All right. So we still got it. I mean we are – as the world's reserve currency, we still have what is referred to as this exorbitant privilege of being able to run large deficits and have people buy the bonds without worrying about what happens to yields. So do you think there would be a reckoning that would happen as a result of an individual event? Would it be a slow degradation of the transactions? I think that – and I mentioned this in the piece. My concern is in 2031 or 2032, you reach this crossover point where 100 percent of federal tax revenue has to use to get paid – to pay interest on the debt and entitlements with nothing left or anything else. That is a scary day. And two or three years before then, I would expect the rating agencies to threaten to downgrade the United States unless there's some serious restructuring of the entitlements. That's a ticking time bomb in real life. That's how far away. Yeah, I think it's three years away. So three years from now, the ratings agencies are going to look to two years later. That's right. They have to give the US a chance to say, look, you've got a couple of legislative sessions coming up. And we all know what that blueprint looks like. Paul Simpson published it 20 years ago. We know there's only four or five big policies that can move the needle and we all know what they are. Tax hikes. Well, the fairy tale is we just grow our way out of it. But is that possible? That did happen under Eisenhower. Yeah. And under Eisenhower, he inherited a debt to GDP ratio that wasn't that different than what we have now. It's about 100 percent of GDP and it fell to 60 by the end of his eight-year term. They didn't raise taxes. They didn't cut spending. They grew like crazy because he had pro-growth policies. that's what the difference is he inherited that as a consequence of we defeated the empire of japan and the nazis right no but the point was that the decline like you said the decline in that debt to gdp ratio happened because nominal and real growth was high right so if you if you if you want to replay that playbook um that's kind of what trump was running on before he went into this nativist territory and why the business roundtable and a lot of business people and people in Silicon Valley were initially enamored with the policy mix they thought they were getting because they thought they were getting a version of that. Let's pro-growth, let's build everything and it just kind of shifted from there. But I also don't have a lot of confidence that a democratic administration would prioritize growth. And I think a lot of other things would get prioritized instead. What if we take a 5% stake in open AI, like some people are suggesting, sell it to CO2, had a trillion-dollar valuation, boom, 50 billion bucks, dump it. Something tells me that would be a negative market indicator rather than a positive. Are you surprised that given all the deregulation, which they promised and they actually delivered on, I would say – They have been – For sure. Are you surprised that with all of that and the Fed having eased in 22 and 23 and just generally getting extension on the 2017 TCJA last year, like all of the things that they've actually done, we're still talking about 2.8% GDP growth? Yeah. Are you surprised? Well, deregulation is a slow-moving thing. It takes a while. So when you look at deregulation of the airlines and telecom in the 1990s, Clinton was the beneficiary of that. But it took three to five years for that really to mobilize the private sector to do things differently than they had been. And those things take time. And then, you know, at the same time that the administration is pushing this deregulation, they're also shrinking the labor supply. They're also increasing the cost of doing business to tariffs. One of the interesting things about what they're offsetting it with the other things. Oh, yeah. There's a lot of offsetting policies there, and it's not always clear to me that they understand those offsetting forces. Around 80 percent of the kind of AI trade are subject to tariff carve-outs. So by hook or by crook, around 80 percent of semiconductors and related things are not subject to these tariffs. But only 20 percent of the kind of power build out is subject to those exemptions. So – and that's where the bottlenecks are. So the cost of transformers is going up. I mean like the deficit clock that keeps going. I have a version of a deficit clock in my head except it's the price per megawatt of a combined cycle gas turbine, which doubled over the last three years and is still rising. And you can't have one anyway. It will take you. You need three years before you're on top of the list. I mean, yeah, people have actually talked about invoking the Defense Production Act to kind of compel GE Vernova to produce more of them. They've announced a partial expansion but not kind of a wholesale increase in production. I don't blame them. it really just them Mitsubishi and Siemens that make these things which is why there so much demand for repurposed shipping engines from Wartzilla and airplane engines that people are stringing together in daisy chains to try to get this And why people all of a sudden have bid up Bloom Energy because they make these solid oxide fuel cells that can convert natural gas into power. And so, you know, but that's the administration. Some of the policies have hurt and gone against the. I guess that's the answer for how we have these American companies reporting earnings that are up 70, 80, 100 percent year over year. And then you just look at like overall economic growth and you just don't see the same sort of momentum. Right. Well, remember, around 70 percent of GDP is consumption, right? This is we've all this. We all started studying this as markets people in the early 90s. And that's not going to grow significantly faster. So 70 percent of GDP is consumption. But roughly 60 to 70 percent of S&P profits is production. So the S&P is a production index and the economy is a consumption index, which is why when people say, oh, my God, look how well the stock market is doing. It doesn't make sense relative to the economy. Those kind of gaps can persist for quite some time. And the reverse can happen too. Look what happened in 2001. The economy was doing fine. The NBER wasn't even sure until years later that there was a recession at all, and we had a 40% decline in the stock market. So these things can disconnect. Michael, you have access to the smartest analysts in the world, and I'm sure your clients are asking this all the time. What's going on with the member stocks? So Daniel, chart 16, please. This is from Joey Politano. We are now importing more from Taiwan than we are from China, which was – that would have been completely unthinkable just a couple of years ago. In dollar terms, yeah. Just completely unthinkable. And the question is, as semiconductors are now 18% of the S&P 500 or whatever it is, are these stocks, these companies that were historically absolutely cyclical, have we removed the cyclicality? Because to me, this is like the whole thing over the next couple of years where the stock market goes. I was fine with this until about three months ago. You're fine with what? The semis being as big as they became? Yeah. Proportionally to other stocks? Like Micron is now bigger than Meta or right there. Yeah. I mean to me, my concern is not the performance of these stocks but the performance relative to what's going on around it. So let me explain what I mean. From 95 to 99, the comm equipment stocks were tracking the comm services stocks. And so the market caps and the earnings are growing up together. And you started getting this signal in early 99 where Cisco and the other infrastructure stocks were still zooming. But the ISP providers and those kind of companies started to roll over. And the smart people, Gary Brinson, I don't know if you remember him, but some of the smart market analysts at the time that saw this coming said, wait a minute. Who's paying for all the infrastructure? Companies whose market caps and earnings are rolling over. The stock prices of the buyers were plunging. There's no way they're going to buy as much of this equipment next year because the stock market investors are telling them not to by selling their stock. And you saw – and if you wanted to look even more closely, you would compare Cisco and Verizon. And like Verizon stopped going up and Cisco kept zooming and like it was clear who is Cisco going to sell this stuff to. And the last three months, the hyperscalers are starting to roll over even though the semiconductor index is zooming. At the end of the day, if the big four hyperscalers plus Oracle can't convince the market that they're earning a good return on a trillion five of capital spending outlays, those infrastructure stocks are at risk at some point. So I'm starting to feel a little nervous about where we are right now. I remember the optical stocks were the last to get the memo. Yep. Because they're – Oh, their order book looked great. Right. They're in the back of the train. They don't know that the locomotive in the front of the train just hit something or went off a cliff. They will be the last car off the cliff. They won't see it coming. All they can see is the car in front of them. Right. So if you're a component supplier to this, I don't give a shit what your guidance is. Right. Because you're not – And chances are. You're not going to be the first to know. So chances are before this is done, they'll all announce a production capacity expansion. Right. Did Micron do that today that they're investing in America to produce more? Well, that's kind of – that's a – Micron just signed a deal, a massive upstate New York facility. Yeah, that kind of rhymes with the TMC project, semiconductor projects in Arizona. It's done for geopolitical reasons. Okay. I give the administration credit for strong arming foreign countries to finally commit FDI in the United States. That's what other countries have been doing forever. And so I think I like the way they're doing. Do you think Apple's lobbying efforts at getting the White House to approve them buying memory from Chinese suppliers has a chance of working? And even if it does, could it potentially be meaningful or are we just going to be looking at this memory supply constraint for as far as we can see? You know, these, I mean, sub two nanometer semiconductors and the really good high bandwidth memory stuff, the leading companies have so, you know, have such economies of scale. I think it's hard to imagine a world where they – It's not going to be this sudden bounty of supply. Michael, I'm curious about where this warning label came from. So when you started to write about the AI stuff, you wrote, this section includes technical AI jargon, conjecture on AI products and services that are rapidly changing, views on non-public companies with disclosures. I mean, you go on, you say, if you have messianic opinions on these topics and cannot stand to see other points of view, or if you are shy at the topic X or Y did not get mentioned, and whatever. Who are you talking to? I get, you know, a lot of people who email you. I get I get fan mail. This is a preemptive shut the up. No, I get a lot of fan mail and on different topics and some of it's not so fanny. And, you know, look, at the end of the day, I expect our clients to judge us based on our portfolio performance, you know, not on the things that I say on podcasts. But I find that the people that exist in this universe have more passionate, unbridled, one-sided opinions as any sector that I cover and that includes energy. You're saying the AI, the people that are focused on AI. Right, right. Debates about token prices, Jeevan's paradox. I mean these people can really get themselves into a tizzy and I think that's because a lot of them. spend a good chunk of their time on Twitter and other quasi-anonymous platforms where they can scream and rant at people anonymously and aren't used to actually having discussions with people in the room. And so people have really, really, really strong opinions about this kind of stuff. Michael Barry is getting people very excited. And I think his research is really interesting, right? How long are some of these products being depreciated? And does that match up against their actual useful lives? He's asking a lot of good and interesting questions. But I think people sometimes are getting very, very carried away. So you called this the most technically complex market catalyst in the last 40 years? Without question. So say more. Well, I mean, if we think about the market catalysts of the last, you know, in the 30 years that I've been doing this, you know, B2B software, the fiber build out. I mean, looks quaint. Yeah. I mean, you middle school level education could equip you to understand what was going on. Now, all of a sudden you have to kind of start with Google's foundational transformer paper in 2018 and, and go on from there in terms of how these things are actually functioning. And, you know, what does it take to produce some of these DRAM and GPUs and how do They compare with TPUs and CPUs, their energy intensity. I mean these are some pretty complex technical things. And so sometimes you'll have a company that comes out and announces that they were able to do something. And it takes quite a bit of work to unravel it and see whether or not that story is real. The problem is the market is reacting immediately regardless. Right. Right. The market is making up its mind really quickly. Right. The market is also getting a lot wrong because of the rapidity of the changes that are happening. But like Market was very convinced that Google was about to either cannibalize its ad business or just in general have the floor wiped with it because of OpenAI. And then the market did a total 180 on that story. And they said, oh, actually, Gemini is superior to a lot of these other LLMs. And Google's user base is going to enable it to monetize faster than anyone else. and then you just have like this complete 180 in that name and that's one of like many examples i can come up with do you ever a couple years ago people were panicked about the fact that um apple had this default agreement that google was going to be the default browser and in europe all of a sudden they said no you know people have to choose the default browser And they ended up picking Google anyway, right? I mean, when's the last time somebody said, hold on, I'm going to Bing that? Like nobody does that. I'm going to use DuckDuckGo to search for something. Right, of course. Okay. I just – I feel like the more rigid you are about how this is all going to develop, the less able you'll be to make money because of how quickly everything is moving. I am concerned, though, that for the last few years, the big hyperscalers have been financing their capital spending out of internally generated cash flow. And just towards the end of last year, they started borrowing to do it. Now, so far, the borrowing numbers are small, and you have to have the discipline to look at the debt-to-EBITAR ratios. And so far, they look OK. But the trend, the first derivative of the trend is shifting away from internally generated cash flow. and some of the charts that we had in the deck in the piece showed that the free cash flow projections are for the hyperscalers are heading towards are heading very low but don't you think the market's being sober about how they're treating the equity of these companies the market is saying yeah we don't really love this in fact yeah and that's the reason why the hypercales haven't performing well have not been performing well look at the chart the chart at the upper left is essentially the capital spending in r&d as a share of revenues and the chart at the upper right is the free cash low margins. You know, this is this is these are some pretty steep declines. And so within 18 months, the hyperscalers are going to have to demonstrate that JP Morgan and the and the big corporate entities in the United States are going to be rapidly increasing their AI adoption and paying for it and not scrambling to use free Chinese models. And I can tell you within JP Morgan, there are lots of CFO meetings about escalating token costs and what should be done about it. And so my understanding is that we're getting more efficient with token use. The problem is the aggregate demand is ramping so fast that it doesn't matter that the cost of a token is falling because once you give people these tools, the only thing they can think about is what else can this do? Yeah. And now you have corporations imposing hard caps on employee use of AI. I'm more worried about that than about – I don't think any of this goes back in Pandora's box. It's not 3D printing where people play with it and they say, you know what? We don't really need that. Right. Like nobody is turning the other way. The agentic AI stuff is extremely powerful. And, you know, I think that's going to continue to grow. And those are massive token consumers. But, again, like there's going to be a lot of pressure in the same way. And the same thing is happening in software. Right. Every CTO is being asked, why can't we use some of these tools to displace some of the vendor software relationships we have? And even if we can't, can't we call our vendors and threaten to do it unless they cut our subscription fees in half? Yeah. So there's a lot of pressure. This is the first time that these contract renewals aren't layups at a built-in 3% rate increase or whatever it is. Automatic escalators as far as the ass can say. I write an alternatives review about private equity and private credit and hedge funds, infrastructure, things like that every two years. and at the end of and you know as a courtesy i i send the sections to some of the people that we partner with in terms of money management and i sent the private credit section to somebody at blackstone uh and um this was last november and one of his many criticisms of my private credit section which was the longest section in the piece because of all the concerns we have is that i was underestimating the value proposition of their software exposures. Right. Underestimating. Underestimating. You and the rest of the world is now underestimating. That's how quickly the narrative changed. Last November, they saw that as a selling point. Oh, my God. And they also had probably close to the highest software exposure. That's right. In November, nobody was really talking about the death of sales force. That's how quickly it changed because by February, that wasn't a discussion he wanted to have. Michael, you wrote – here's the important point. According to JP Morgan Equity Research, the unit economics of ASICs – so we're talking about Google, TPUs, AWS, Tranium, Microsoft, Maya, Meta, NTIA. They're improving versus NVIDIA. You wrote that the hyperscaler is using their own in-house chips report total cost of ownership reductions of 30% to 40%. Right. Okay. Is this why you think NVIDIA is stuck in the mud? Because the forward P of NVIDIA is as low as it's been. I think it's at 18 times forward. It's 20 right now. I don't think that's bullish. That's not a good thing. I don't think NVIDIA getting cheaper is like awesome. So first of all, like the thing you just read is a great advertisement for the other thing of me saying this is more complicated than other technology revolutions. Because for the last few years, we were all told that NVIDIA GPUs were untouchable and that the efforts that the other hyperscalers were making to develop their own alternatives were not ready for prime time and that they were stuck contributing forever to NVIDIA's 70% gross margin. That was wrong. And all those companies that you just mentioned have been able to build their own accelerators. These are all fall under the category of accelerators that migrate compute away from the CPU. And they've and none of them have successfully yet figured out how to or even if they want to sell them broadly to third parties because you'd need a sales force and you need customer support. And NVIDIA has all that. You need all the software. But they were huge customers of NVIDIA. So even if all they do is build their own tools to replace the NVIDIA stuff they used to buy, that's big news. And yes, NVIDIA has a trillion-dollar backlog, but that's not the point. I listened to Andy Jassy on a call maybe last quarter or the quarter before, explain that if Amazon's training in business internal would be valued the way that we value other semiconductors' business, like the annual revenue that that would be equivalent to it would be one of the largest companies in the world yes but but again right now it's it's it's mostly internal to amazon but couldn't that change on a dime it it could but again the cloud doesn't even the cloud was internal to amazon until one day it wasn't but it doesn't even need to to be a problem for nvidia which is i think what you're mentioning which is like these are some of nvidia's biggest customers saying, you know what, we're going to buy some GPUs, but less than we used to because we're building our own and saving a lot of money. And the market is sussing that out because the FOP is shrinking and it has been for a long time. And I think the market is being quite sober in a lot of different areas, despite the fact that there was a lot of insane behavior, despite that there is a lot of leverage. Josh and I were talking last week, the amount of money in these double levered ETFs, there's like 500 billion dollars of notional exposure and this is why you see the giant whipsaws day to day i spent some time with our derivative guys are you worried at all well i mean i'm i spent a lot of time with our derivatives guys understand i had i had a weird chart in the piece that looked at the the contribution of these levered etfs to changes in daily market cap swings yeah and And the way that they're hedged, right? So the bottom right, right? The bottom right chart. Unbelievable. So what are we looking at here? Oh, boy. So the way that these providers hedge these exposures is they have to buy on the way up and they have to sell on the way down. Essentially, imagine a market neutral fund that's got to go home flat at the end of every day. So if they've offered these leveraged ETFs, they have to constantly be chasing the market up and down. And as you can see, just this year, the amount of money that's been pouring into this stuff has increased a lot. And that just means, like you said, you're going to get a lot of 2.30 p.m. to 4 p.m. noise. and it's just – I think it's going to be kind of spooky for the average investor. If these stocks peak and then trade flat or trade in like a 10 or 15 percent drawdown from their highs – These are all dead. For extended period, this activity though will go away because it's not fun anymore. And also I mean if they flat it also a problem for the construction of the product because it not embedded leverage You have negative compounding Right right And so you got the cost got the cost But I saying the dollar value will come out of this game They go on Kalshi They'll do something different. My favorite indicator on this page, for better or for worse, is the chart on the KOSPI. So the margin loans on the KOSPI. And all I can tell you is, and I'm remembering back to 2008 and 1997, Korean retail is basically the scariest market indicator because by the time those guys start to pile into stuff, you're really, really, really late. This is piling in. Yeah. But yes, and in fairness, there's a reason why they're piling in. Samsung and SK, the growth there is otherworldly. It is. But again, they're the last train on the caboose and we'll need over the next six months a pretty sharp rebound in the hyperscaler projected cash flows and AI adoption in order for this all to settle out. So when you see Amazon sell $30 billion in bonds or you see – That doesn't worry me yet. That still puts – The numbers are not big enough. That still puts their debt-to-cash-flow ratios. How about Alphabet selling stock after 10 years of buybacks? That's different. Yeah, that was different. Equity fundraising for CapEx is not – I don't think anybody – Well, I mean it tells you. We looked at the chart already. They're not generating any kind of internally generated cash flow where they could just write some internal checks to do that. Right. Okay. So I wanted to get your opinion on SpaceX. I thought the lockup, what they were doing – why not? I think the way that it's done right now is so gamified and so predictably awful. It doesn't work. The typical six-month lockup, you see shares fall in. They sort of settle out. I don't know that I love that they only released 3% of the company or sold 3% of the company. I think they did that very intentionally. I don't really love that part of it. But to have this staggered lockup, some price points where more shares can be locked up, I didn't hate it. Yeah, the question is like direct placements as an alternative to IPOs. Let's see how it works. I mean at the end of the day, by the end of this year, we're going to go from something like 3% to something like 50% of the shares being part of the free float. I don't know if there is a lot of analyst support that has come in this week, predictably, obviously. I don't know if you saw the FT Alphaville buy recommendation. Stop. FT Alphaville initiated coverage today on SpaceX with a 12-month price objective of infinity. They don't do that. Stop it. Tell them you saw it. Tell them you saw it. You don't have to comment on it. I saw it. I don't want to comment. Okay. There would be no comment on that. If you look at though the way that SpaceX placed and priced and the activity since, I know it's been only a few days. There were a lot of really dire predictions out there about how it was going to screw up the indices. People are going to – maybe people dump stocks but we can't really see it on the chart. It's in the queues. There's no evidence yet that anything super funky happened as a result of SpaceX's coming public and it being included in the index, at least not yet. I'm going to say so far, the worst predictions. So far it's fine. And I like the fact that S&P stuck to its guns and did not make an exception about companies having to be profitable before they get – and put it to the index. You like that they did? Yeah. Somebody has to have some profitability standard or else you end up with the Russell 2000, which is a pile of dog shit companies that don't make money. Right. The NASDAQ put it in because it trades on the NASDAQ. So if you're going to let this thing trade, you have to have a good enough opinion about it. I think that I wouldn't be surprised to find out. The S&P and NASDAQ and a bunch of people all kind of discussed like how can we split the pie here in a way that's reasonable? OK, you guys put it in your benchmark. We're not until the – like it wasn't an all or nothing thing. And so I think it's a reasonable compromise the way the indices chop this thing up. How do you feel about the health of the bull market right now? Let me lead you to water. So Josh and I were talking last week about if you are just looking at the market inside the market of things that you would want to see participating in a bull market to give you confidence that things are going pretty okay. Regional banks, transports, industrials, small caps, micro caps. Wall Street banks. Wall Street banks, Citigroup on fire. There is a lot to really like about the current state of the market. I'm not saying it's going to continue forever, but just today, diagnosing the health of the market. It looks pretty good. The internals look okay. Obviously, we've got some strange valuation shifting going on in terms of what's going on with the hyperscalers and the semiconductor companies. But the market breadth numbers are not great. Transports look great though with the Dow. And then away from the stocks, retail sales were way better than expected. Bone growth at the banks. These are all signs of a healthy environment. The thing that we look at the most is in the PMI indices. It's the ratio of new orders to inventories. Of all the things that we follow, we follow hundreds of them, that one has the closest relationship with the performance of the stock market. What sort of orders are they measuring? Because I remember you spoke about that once before. New manufacturing orders. You could also look at new orders and services, but I don't like it. It's unclear what exactly they're doing. But new manufacturing orders is very clear and remember we talked about the S&P as a production index, not a consumption index. So new manufacturing orders relative to inventory accumulation, and we have this link that has this Trump tracker that has all the charts on the economy that we follow. People can look at it and that looks okay. And a lot of the contemporaneous coincident leading indicators and forward leading indicators look okay. Consumer part, a little shaky. rising credit card delinquencies, rising auto delinquencies. But again, you know, those aren't huge drivers of corporate profits. Do you think the election is going to have an impact on the stock market in the second half of the year or the fourth quarter of the year? It usually doesn't. It's very noisy, but I can't remember a midterms that made me feel differently about what I was invested in. Yeah. You know, it's kind of like, you know, you know how there's certain things you just don't spend time on because it's a waste of time and you either need to do other work or rest. That's how I feel about the. You're not getting questions yet about. Well, that's how I feel about the State of the Union address. Like it's meaningless in the scheme of things. And I kind of the midterm elections, you know, rarely have have substantial policy implications for investors. Now, you know, Democrats take the House and the Senate. What happens to the Trump agenda? Do they impeach him again? I don't know, but we'll see. If you want, we can spend a couple of minutes talking about redistricting and things like that. OK, do you worry about whether or not Anthropic and OpenAI will be able to actually go public this year? And if they have to postpone it to next year, could that potentially be looked at as a threat to the earnings expectations of all of the suppliers to the AI ecosystem? Is that a thing that you're concerned about or not really? A lot of the CapEx is supposed to come from these companies. Right. And ostensibly, they're going to need liquidity to hit these commitments that everyone's got baked into their earnings expectations. Yeah. Yeah. Well, you know, S&P buybacks are still running at a pretty decent clip and there's a lot of cash M&A taking place. So cash has been around 70 percent of M&A compensation this year. So I'm not worried that there's like, quote unquote, not enough cash to buy these new IPOs. I don't think that's an issue, particularly when they're only selling three percent of the float. Right. So it makes it very easy. Yeah, there's no stock for sale. So – but I'm – as a general principle, like let's take away some of the valuations and the noise since the IPO. I think if I had to rank the business, the underlying business uncertainty, I would say OpenAI the most, then Anthropic, then SpaceX. I think the market is – I don't think the market is going to like OpenAI. I think it has more business uncertainty than the other two, right? I mean, and, you know, everybody, everybody, again, everybody's going to have their own ranking, but that's mine. Why do they have more business uncertainty? I think when you look at the concentration of how OpenAI makes money, there's not a lot of, there's not a lot of advertising revenue. It's a subscription service. Not as much enterprise. Yeah. It's much more consumer free. And those are the kind of things where sentiment can shift. I just – I don't – it doesn't look to me to be – the dynamics don't look quite the same as the other two companies. Yeah. They sort of have a flavor of like AOL in the early internet to them where it was like the first thing that everyone used but nobody stopped there. Yeah. I mean the network effect is pretty powerful but Anthropic showed how quickly those kind of things can change. By the way, one of my summer projects is our COO has asked me – so J.P. Morgan is a member of Project Glasswing, which is the – The cybersecurity. Yeah, the 20 or 30 companies that were given Mythos. And so the COO asked me to partner with our cybersecurity team to write a piece on what we've learned so far from Project Glasswing and what's coming. and so I'm in the midst of working on probably publish that by the end of the month. That's an eye-opener. The name of the piece is called Patchmageddon because you know the story. Bruce Willis went to space, of course. Well, no. Whatever that is, I didn't see it. They said before we release this to the public, we should let these companies know how many things need patching in their existing. And the pace of confirmed vulnerabilities is going up like this and the pace of patching is down here. And there's also physical infrastructure issue with programmable logic controllers. This is going to be a carnival for bad actors. So state-sponsored and non-state-sponsored bad actors are going to have a field day for some period of time using these models to discover bizarre hidden vulnerabilities that nobody knew existed. And so when you look at the vulnerabilities that these programs find, it's like, well, you know, if you log in this way and then you do this other strange thing and then you hit the following three keystrokes, all of a sudden you can execute, you know, a corporate takeover. And like, and they find these things that you would, that no human would ever have tried to do. That's why they're so good at drug development because they can work through all these permutations that nobody would – combinations of molecules that nobody else would ever think to do. And then all of a sudden they find like they did recently an antibiotic – a molecule that can combat this antibiotic-resistant MRSA. It really does feel like it's an alien in our midst. It does not think like humans think. It arrives at answers that are palatable to the human that the answer lands on. You almost don't know how it got there. As a matter of fact, a lot of banks started using neural networks two, three years ago to approve or deny debit card and credit card expenditures on the spot. And because they weren't able to describe how those neural networks actually worked, they had to work – people complained to regulators. And the banks have had to overlay these new tools that can try to take what the neural network did and translate it into English, which it doesn't always do such a great job at. Right, because you might get the best outcome 100 times, but on the 101st time, you get a very bad outcome. And if you can't explain where it came from or why, it's problematic. I don't think it's going to amount to like a market-wide thing. But buckle up because individual small mid-cap and large companies are going to end up in the headlines from time to time because they got breached and attacked and they'll have to kind of sort out what they do about it. Can I make an observation? Yeah. Tell me if you agree with this. I feel like you're in all your glory right now with all these cross currents and all these like new technological things. I feel like you're having a lot of fun trying to solve these puzzles. Do I have that right? Do you agree? No. You're not enjoying it? No, I'm 64 years old. It's hard. My brain is having to work so much harder per unit of output. I feel like you're perfectly positioned amidst all of these things that you've been studying your whole life and where they overlap and all the lessons that you've done. You might not feel this way, but we do. We're getting a lot from you. Anytime you publish something, immediately New Semblers dropped. We read it. That's how we feel when you publish. I remember my kids used to talk that way when there was a new Drake album. They would say, oh, no, wait, not Drake, the other guy. You see, now I'm not even doing it. Well, that's you. You're the Drake of finance. Frank Ocean? Yeah, new Frank Ocean drops. Anyway, they used to say something dropped and they would be all excited about it. Yes. Well, we feel the same way. We feel that way about your research. And then you may – the one thing is you may get an email every once in a while that's like not a fan. That's okay. It goes with the territory. By the way, Jamie is great at like – who cares? You know? You got – everyone has their own opinion. There's a million people who are reading you that don't send an email but they love what you're doing. You just won't hear from them as often as you'll hear from the one person that's got an issue with one thing that you said. Right. So don't overindex to that. would be as a veteran of getting feedback for things I say in public. Yeah. Well, can I tell you a lesson that I learned? Yeah, please. You learn things at different points in your career. I write the energy paper. I probably spent three or four months on it. It's the thing I work the most on each year. And this one was called Fighting Words because of all the debates and things like that that people have about energy. And the cover was a three-way shootout from the movie The Good, The Bad, and The Ugly. And it was a three-way shootout between advocates of nuclear, advocates of renewables, and advocates of fossil fuels. And you find all three of those represented in society and within the administration itself. And I had a couple of occasions where a certain client was very vociferous with me about their concerns about the way that I was evaluating the energy markets and confronting me about it. And so I wrote about that in the introduction to the energy paper and anonymously so that nobody would know who this person was. But they knew who it was and they were very, very upset. They were upset. Why? Because you used their argument. I used that interaction, which wasn't the easiest one to deal with. I use that interaction as a metaphor for the for the fierce debates people are having about energy. And I probably use certain words or phrases that to describe those interactions, which which weren't really super well received. So what's the lesson you learned? Don't do that. Oh, I learned lessons. OK. What have you learned? That there's there's a line not to cross there. And and I got I got some phone. I got some calls from Mary, uh, from Mary Erdos, my boss about, about that particular one. Um, cause you don't know, you know, exactly who people are. And, um, but it's, it's ending up in a good place because what I've decided to do is I'm going to, well, this is a first for me in my entire career in next year's energy paper. I'm going to allow that person to select the policy energy expert of their choice to write an unedited, unexpurgated section in the energy paper on why they disagree with me and on what. So that point of view will make its way into the research. It will. And you won't have to anonymize the person. You could just use the argument. And so it wasn't kind of planned that way. But, you know. We all do what we got to do. Yeah. So, you know, you learn different lessons at different points in your career. And I learned one of them this year. Well, fair enough. Well, listen, we really appreciate you coming by. walking us through some of these charts. I want to tell people where they can find more from Michael Semblist. I generally post almost everything on LinkedIn. That's right. Okay. So you want to follow Michael on LinkedIn, of course. And then you drop videos when these reports come out. I tell Michael I love the video. It's easy to digest. Each eye on the market is accompanied by a video podcast that I record from my home. so okay um all right so if you remember you can also follow my fishing adventures i do on instagram oh boy what do we have here wait what are we doing that better not be a firecrackers no firecrackers wait wait wait is she serious rob you have responsibility for this Happy 250, everybody! USA! USA! Very nice. Well done. This is a unique program. Nicole, wait till you see your footnote. Hold the call field. Here we come. All right. Thank you to Michael Semelis, ladies and gentlemen. Thank you so much for watching. Thank you for listening. Happy 250th America. And we'll talk to you soon. We'll see you again. Thanks again. Nice talk. Very nice.