Value Stock Geek on the Magnificent 7, tech, $AZO, the Weird Portfolio | S07 E44

Summary

Value: After Hours is a podcast about value investing, Fintwit, and all things finance and investment by investors Tobias Carlisle, …

Transcript

I think we're live. This is Value After Hours. I'm Tobias Carile, joined as always by my co-host Jake Taylor. Our special guest today, uh, no video, but he is here in [clears throat] Voice and Spirit. Value Stock Geek. What's up? >> Hey guys, thanks for having me on. >> Welcome back. >> Um, for the folks who don't know, Value Stock Geek, give us your investment philosophy. those who have been living under a rock. >> Sure. So, I um I run a Substack where I analyze a company every week and um I'm basically looking for quality companies and then I'm trying to identify the cheapest names in that universe. It's a lot like the uh magic formula, but there's more, I think, qualitative work that goes into that. It's just not systematically selecting the wonderful companies. And then uh the quantitative component is more about evaluation. And so what sort of things have you looked at? What sort of names have you looked at? >> Yeah. What's what looks good right now? >> Well, over the years, I mean, I've been able to pick up some really good names at cheap prices. Like I was able to pick up a lot of good tech companies actually in 2022. Um, that was like Taiwan Semiconductor and Meta. Um, I was able to get those at like 10p and that kind of thing. Um, nowadays those are obviously not cheap anymore. Nowadays, you're seeing a lot of um liquor companies actually are popping up a lot in in the uh in in those screens that I've been running. Um I'm about to do a rebalance in in early January. Um and it looks like a lot of those names are going to enter that universe like uh Brown Foreman and Dagio. Um where there's been a lot of talk about how Zoomers aren't really drinking as much as previous generations. Um, and you also saw sales decline after uh COVID. There was a big spike during COVID and now it's kind of returned to normal and I think people are just extrapolating that trend. So, I think that's what you'll see a lot of these days. Um, but yeah, overall I think um there's a lot of interesting little opportunities out there. >> A drunk pig in the python. Is that is [laughter] that a thing? [clears throat] >> Maybe. Maybe. Um, >> I saw a note that alcohol consumption is back to like 1930s levels or something like that. >> What they prohibition? >> Well, prohibition's the isn't it the 20s? I know that's that's the roaring 20s. I don't know. >> Yeah. Yeah. The 20s and then they repealed it in 32. I think it was one of the or 33. It was one of the first things FDR did after he got elected was repeal prohibition. >> Um, >> so we drinking as little as we did in prohibition. I the charts I've seen go back to the 60s and we are at a low but if you kind of like zoom out you know it's more of like we've gone from we're down 3% or something from total alcohol consumption >> yeah I I think people read into it like they're saying zoomers don't drink and I don't know if that's necessarily true. There is a lot of concern that marijuana is crowding out alcohol spending. Um, that's that's a big concern as marijuana becomes more legal. But overall, I mean, I think I think things are pretty stable. I think people are still going to be drinking alcohol 10 years from now. Um, maybe a little bit less, but I I don't think there's been this mammoth societal change. I think it's more I think it's more just about alcohol consumption got really hot during COVID because people were locked up and sales went through the roof and now they're kind of normalizing and maybe people are adjusting their habits a little bit but I think I think it'll go back to normal. >> Used to be this idea that booze was pretty recession resistant because when people go through bad times they tend to drink more. >> Yeah. Let's get the unemployment numbers up and then that'll uh then the booze will come back. Is it crowding out by by weed or is it a lack of money in Zoomers or is it like a everybody's online and they've seen all the studies and they've decided that they don't want to drink as much. What do you what do you think's the >> causing that little trend, that little downd? >> I really think it's just about um things getting to things um reversing after COVID. I I think it's really just more about that. I do think those longer term concerns are valid and are something to keep an eye on, but I don't think there's enough evidence to say that people have really significantly changed their habits. Um, I I think more of like people are looking for an explanation for the price action and they're creating these narratives like zoomers don't drink. They're just looking at their screens all day and they're not going to bars anymore. And uh, everybody's replacing marijuana with alcohol. And I think it's just like they're just crafting a narrative that I don't really think there's some truth to it, but I think it's it's being exaggerated. >> Is it impacting any part of the market more than any other? Like how does how do hard hard liquors spirits hold up against >> beer and things like that? Yeah, I mean a lot there was um a ton of interest over this over the um Seltzers like I I know in the early 2020s like if you look at Sam Adams stock that went through the roof because of their hard selter brand. Um yeah and I mean I I think I think that's just more of an alternative to traditional beer. I think the alcohol companies are adapting and and adjusting to that. And then there's the traditional spirits which is like more of Diagio and Brown Foreman's wheelhouse. >> It's funny how many charts went crazy through that 2020 2122 period and have come back to earth and how that's hard to it's hard to tease out from that like what is secular and what is part of that I don't know whether 22 pulled a whole lot of demand forward or it was just a whole lot of stimulus through that period. >> Yeah, you're right. There's so many charts that look like that where it is really difficult to discern what's a secular trend, what was just a COVID thing, what how things are things just returning to normal. And I mean, you can see it in all kinds of things. You can see it in uh I mean, we saw it in lumber. That was a pretty good example. Remember that? >> Yeah. Um where there was this big spike and then again there's a narrative that goes along with that. Oh, we're in a housing shortage and lumber prices need to go up and blah blah blah. I feel like anytime you see something do well or do poorly, it's almost like people make up some secular trend to go along with it and it's usually not. It's usually just a cyclical thing. >> The narrative follows the price every single time [snorts] >> pretty much. Yeah. Um and today, I mean, you're probably seeing with AI, there's probably I don't know if AI is necessarily the kind of secular trend that everybody seems to be convinced it is. And then I'm sure if it AI spending goes down just a little bit it'll be what a waste of time this was nonsense and you know the truth is probably somewhere in the middle [clears throat] right there's actually there's quite a few companies that I would imagine you're there's some names falling into your quality and cheap bucket that had been quality and expensive uh for a long time and now all of a sudden are starting to look a little bit more attractive. Is that is that accurate? >> Well, that's the thing. So every week I look at a company and I would say 90% of the time it's too too expensive. But it's all about just constructing this big universe of quality companies and then trying to identify them when they're a little bit beaten up. There were actually a lot of um last year at the end of the year there were a lot of good defense names that that were good. Um >> my screen picked up um Huntington Engles which you know builds the aircraft carriers and that kind of thing. >> Yeah. Um, and there was concern in Q4 that there were there were all the issues that were happening with the budget that, you know, they would that these contracts wouldn't be fulfilled and that all turned out to be nonsense and now the stock is up 70%. Just by reverting to the mean. Um, and there's there's a lot of things you see like that that happen. I mean, it's the same thing with tech. Like I looked at all the tech companies um in the beginning of the substack in 2021 and at that point I said these are all too expensive and then lo and behold a year later Meta is at a 10p. Taiwan semiconductors at a 10p. Google was like a 15p back then uh which cheap for Google but you you'll see a lot of that a lot of that stuff happens. People get convinced it's a secular compounder. I'll add it to the universe because there is some truth that the business is good. Um, and then you just wait a little bit and it usually doesn't take that long [laughter] and all of a sudden um, this is cheap again. And then it's at a moment when everybody thinks that, oh, it used to be a compounder but now it's garbage. And it's usually not. Sometimes it is, but usually when you're looking at a lot of these really high quality companies with Moes, usually they just go through cycles and it kind of baffles people's minds when they do. to something like um I'm just going to use Copart as an example, but you know, [clears throat] historically a pretty amazing business, >> but was always very expensive. Now, let's call it a 23 or something PE. Is are things starting to are they interesting at that level for you or are you really waiting until it's like they're down in like a 10? >> No. So, I look at one of the big things I look at is um price to sale is trending in price to sales. So, what price to sales ratio does this normally trade at and what does it trade at now? That's like a key metric I look at. And so, that keeps me from saying like I need to get super deep value where like some companies will never trade at a 10p, but if you have some like Copart is a good example of a company that may not get that low, but if you look at trending and price to sales, you can identify these inflection points when they might be a little bit cheap. Um, so yeah, usually I look for I rank everything by discount to its typical price to sales ratio and I rank all the companies by that criteria. Um, and then I'll look at absolute earnings yield. So I'll say like is the earnings yield at least maybe um in excess of the current 10-year Treasury yield. So you know basically today is like the peak is it is the um enterprise multiple below 20 you know and is the price to sales and then rank that by the price to sales uh discount to where normally >> are you using price to sales because it's such a clean metric like it's there's not a lot of counting shenanigans between revenue and revenue. >> Yeah and I've noticed for a lot of these companies price to sales like you're right sales is more stable. Um, but what you can really pick up there are cycles in margins. So you look at margins and price to sales for a lot of companies and often it looks like a wave and you can identify points where okay this is probably just a cyclical phenomena and it's trading at a super low price sales multiple because margins are temporarily depressed. um as opposed to and I think that's better than what the traditional magic formula does where it's only looking at trailing 12 month ROIC >> because I think what that does is it tends to identify companies at cyclical peaks which is why as Toby's research has proven you know that that that tends to underperform pure value and I think it's because it identifies a lot of those companies that are trading at cyclical peaks like I remember looking through magic formula screens and it would often things would pop up like that like Cal Main 2015 the egg producer, you know, that was popping up on the magic formula screen and it was mainly because there was I think back then there was another bird flu and prices were high and well what happens you know markets work people you know egg producers will create more eggs and uh prices will eventually come down. Um, and then it identifies a lot of fads like Fitbit. Like that was another one that was in there, >> right? Yeah, that's it's it is prone for hula hoop uh behavior. >> Have you seen those those charts on margins? I post them occasionally. S&P 500, 400, and 600 over like the last 20 something years >> just up and to the right or >> it's amazing. Yeah, the the margins like margins are 13 or 14% for the S&P 500 on average, which is just extraordinary. Even even mid and even mid and small are elevated well beyond their long-term averages. >> Yeah. And I guess the question is is that going to continue or is that going to mean revert? I don't I don't really know. Um I guess you could argue that there's less antirust that these tech companies are different and nobody is doing anything to capture their market share. On the other hand, it's a mean reverting. You could say it's a mean reverting series and all these companies top are in for some trouble. I I'm not sure if I know the answer. >> It has been a mean reverting series. I don't know if it is. Like that's it's one of those things that's Buffett and Husman and Jeremy Grant. >> It's a sweet time to mean revert, doesn't it? >> Yeah. Like [laughter] I think someone said 6% I think Buffett says 6% is >> and Grant both 6%. >> It's like an unusual number and here we are. We're talking 13 or 14%. That's crazy. So, I've tried to um play the devil's advocate because my own natural gut instinct is also like this is not sustainable. Like everyone's going to get cooked. Uh these are crazy. So, I'm trying to like temper that a little bit with running a counterfactual, but like so what would I need to believe about the world to think that not only is like 13% the new normal, but maybe we go higher. Um, you know, I heard or um I read about Michael Dell talking about like uh AI potentially adding 10% productivity. I don't know exactly what that means like in an income statement, but let's just assume that it's taking 10% of cost out. And let's say that you like what if let's just imagine this crazy scenario like what if public companies all companies were able to run at like 20% margins. Uh then what like is this actually a pretty is the market relatively cheap today in a 20% margin 2027 world uh and what would be a reasonable multiple for a 20% uh basket of companies like 30 times earnings 40 times earnings I don't know like that maybe uh and therefore >> keeping this up in 20 years time [laughter] >> right uh yeah these bold bold macro predictions by uh by area economist But I don't know. I'm just trying to like temper my own biases which I know are more towards the like this is all and is going to come crashing down. >> Does it come out of labor? >> Is that the is that the other side of the coin? >> Well, and I and of course like the counter to all this is like okay um yeah so now we fire half the workforce. Is that what we're saying? Like where's the demand for these products now if you half the people are unemployed? Uh, also this is assumes that like no one is going to compete ever again in business, which >> like wait, I'm I got 20% margins and I see that you have 20. Well, maybe I'll take 18 and uh and I'll take some of your market share. And then you go, well, hold on. I'll take uh 14 and then I'll take some of your market share. >> Uh, that's how it's supposed to work. >> That's how it's supposed to work, right? >> Yeah. That's how capitalism is supposed to operate. No doubt about it. But yeah, I'm I'm I mean I have the same biases. like my gut says this is a mean reverting series but like you I tried I also thought that seven years ago, eight years ago and it doesn't seem to be reverting. So I I don't know maybe it's just that these tech companies aren't being reigned in by any antirust anymore and they're eating up a lot of other businesses. They're destroying a lot of businesses along the way. You could argue that's a good thing, a bad thing, I don't know. But it seems like they're dominating a lot of they're going in and they're taking away a lot of like advertising businesses for instance are basically getting destroyed by Google and Facebook and that's been going on for a while and you know those were businesses with certain margins and they're consuming that and I don't really see what what is the competition that's going to change that. Um it's hard to it's hard to imagine at this stage of the game. Well, aren't they all coming to compete though on AI in a and they had their cozy monopolies before where they >> sort of didn't try that hard against each other in in their space and then uh it feels like they're all kind of heading towards one goal which would be to win the AI world which to me seems like okay now there's now the elbows get a little sharper and maybe there's actually competition. >> Yeah. I mean maybe this is the catalyst. I feel like AI is like protein, you know? It's just it's like it's in everything these days. Like you try to buy any kind of thing, it's got additional protein protein in there. >> It's got some AI in it. You're like, well, what am I going to do with that? [laughter] >> Get swole, Toby. What do you mean? >> Gmail keeps on trying to get me to AI my email account. I'm like, dismiss, dismiss, dismiss. Go away. [laughter] [clears throat] >> Yeah. And I don't even know if what we have as a what we're calling AI is really AI. I'm not an expert on like by any means. >> It's the marketing, that's the label, the advertising. >> It kind of seems like maybe this is just like searching the internet and plagiarizing Like >> I know a lot [clears throat] of it does feel that [laughter] way, isn't it? >> It's brondo. It's got electrolytes in it. It's what plants crave. >> Like is this really Howal 9000 or data from Star Trek? I don't I don't know. I'm not sure if it's that level yet. >> Little bit of a non-seced. Do have you have you followed Autozone? AO. A ZO is the ticker. >> Yeah, great company. Yeah. Um, AutoZone and O'Reilly, they're both really good companies. They >> have a great little niche. They buy back tons of stock. >> Yeah. I mean, that's their business, isn't it? Buying back stock. >> Yeah. And business is good. >> I I saw that AutoZone whack today. Uh, or has over the last few days, I thought, because >> perfect cheaper buybacks. Yeah, that was my initial that was my initial thought. But the point of the tweet that I saw was that um they've used they've managed their payables and their working capital to create a lot of the money that they use to buy back stock and with first brands and those other companies falling over like they that they may not be able to play that game anymore and I think that's the reason why the the stock has been hammered today. >> Do you have a view? Did you ever dig into that enough to sort of >> I did. I I wrote them. I've I've written both of them up. Autozone and O'Reilly. Um I I didn't know AutoZone had a big decline today, but if that's the case, I wouldn't be surprised if um my screen picks it up at the end of the year because it's in that universe of wonderful companies. Um and I don't really once it shows up there, I don't really question it. I like I'm going to buy the 15 chiefest companies in that universe. I do trim the universe, but um I'm not going to exclude something because I think it's secular or that the trend has reversed or whatever. Um, I'm going to trust the the process and and it'll probably >> argument is they're using working capital as a source for buybacks >> that they've been freeing up cash flow in every single way that they can. And one of them is by, you know, you pay your suppliers layer and layer. >> I mean, you can do that one time, but that's not >> you can extend it. You can Yeah, that's right. You're right. But at some point Yeah. And then it unwinds when you >> But the collapse of first brands is going to impact them. It changes their financing. >> Interesting. >> I'm not over it enough to know. I just, you know, I like that's the I I like companies that buy back a whole lot of stock. And so I'm just wondering if that impacts anything else. >> Yeah, I I'll have to take a look at it and see if that's true. Um, but yeah, I also like companies that buy back tons of stock. That's definitely I always look at one of the things I'm always focused on is shareholder yield. I want to see a really decent and growing shareholder yield. Um and if you do really look at any company that has bought back a significant amount of stock over the last 20 years. >> Um almost always you get a pretty really good result for for investors for any kind of company that does it. Sometimes it's like a delayed reaction like you see that with some like Dillards like they they bought back stock for 10 years >> nonlinear way >> and then it boom it just all happens at once you know so like a coiled spring there >> they got a little bit lucky with co I think there was some stimulus that flowed through and that was a big part of that >> yeah so that was the catalyst >> good things happen to cheap stocks Toby >> or not sometimes [laughter] I've said that. I don't know that that's true, but I hope so. If you say it enough times, it might come true. [sighs and gasps] >> Um, before we came on, we were talking about Palunteer and Nvidia. Uh, their reactions to Barry's uh, so Barry's got a new Substack. You guys have taken a look at the Substack. >> Read one of them so far. >> Did you? >> I read one article. Yeah. >> What was his What was his Palanteer Nvidia? What is this Palanteer Nvidia thesis? >> It's around stockbased comp and that they're aggressive with that. And if you do an actual kind of owner's earnings look at it, like Nvidia's trading it more like 60 times and not 30 times if I remember what he said correctly. >> That's a that's a sin around the here these days is it like paying a whole lot of SPC. >> I know >> that's just business, isn't it? 15% a year goes out doing SPC. >> That's exactly right. Like it hasn't mattered for at least the last 10 years. Yeah. I mean it's just it's like okay well cool you could be technically correct but no one who cares like it's going up. It doesn't matter. >> You like accounting or you like to make money? >> Yeah. Exactly. >> Do you like [laughter] >> he really seems to think that this is a lot like the late 90s. Um I I he talked a lot in the article about basically comparing the current era to when he started his blog. I think around 97 or something when everything was really getting going and at the time he was living in, you know, the epicenter of the tech bubble. So he was surrounded by it all. So he seems to think that the two environments are very similar. >> 97. Nobody's got the balls to say 99. Nobody's got the balls to say March 2000. [laughter] [clears throat] >> Yeah. >> Everybody's like it's four years away. But that was one of the interesting things that he talked about was that um the spending the capex didn't peak in March of 2000 like the market did. It peaked like a year or two after that. >> Yeah. >> The market had already like given up and fallen off a cliff while the spending was still happening. So even if maybe you know we have this big runway in front of us that everyone can see of AI capex maybe the market doesn't you know it looks past that and it it looks down the slope from there. I mean who knows but that's I think that was one of his points. >> Yeah and I I remember Cisco traded at like I think it was 50 earnings 50 times earnings at the peak or something and um the company did very well like the business did very well from 2000 to 2010. It grew. I think it tripled or something. Um, but stock was dead money because >> I think it took out its all-time high uh like in November >> when said it was an AI. >> 25 years. 25 years. It took >> 25 years for a business that probably >> grown I don't know over 10% a year for 25 years. Um, >> but that's prices stock prices can get a little ahead of themselves. >> They can. >> Ain't that the truth? I mean, I [clears throat] think that was the uh the thesis that the uh the Nifty50 have ultimately delivered like nearly a market perform since the Nifty50 were founded in you know whatever the 70s sometime >> even though they got ahead of themselves in the 70s like they fell back there was a you were you were dead money for 50 years or something in the interim. >> That was um Jeremy Seagull wrote a paper about that. >> Yeah. and and he I think he did it in 92 and he said if you bought the if you bought them they all recovered but at the same time it's like okay you had like an 80% draw down and you just delivered a market return I don't know if that's a great result I'd rather buy the N but to his credit it is true that those companies continue to compound in value but my takeaway from that was well then I'd rather buy the Nifty50 in 1975 versus 1971 [clears throat] >> and you'll get a much result >> wasn't like Philip Morris drove like 50% of the results or something of that too. So like if you took that out if you didn't hold that through the whole thing like I don't know if you ever made it. >> I could be wrong about that but >> No, that might be true. Yeah. I mean there were a lot of really good companies in there like McDonald's was in there. Um Xerox, >> Kodak, bunch of really [laughter] >> Yeah. Then there's the duds like Polaroid and Kodak. That's true. I mean, they weren't duds. They weren't duds until they were until they were crushed. They were pretty strong for a long time. >> Not for >> Yeah. >> They were the modern equivalents. >> Yeah. I read once when um Buffett brought um Bill Gates to one of the meetings of, you know, of he would have like Walter Schllo there and all his friends. Um, >> yeah. >> And they they I I remember he wrote that they asked um Bill Gates about Kodak and Kodak Bill Gates said in 1991 or so Kodak is toast. Like he saw the writing on the wall at that point, which no one else really did yet >> because I I think it was trading at book value or something in the early 90s. I guess it worked for a little bit until till it didn't. I want to know who picked that cameras were toast because it seems like as much as that destroyed film like it re it was really the phone that destroyed cameras because there's just no reason like the cameras on a phone are so good now and that >> can you imagine buying a regular camera nowadays and bringing it with you on vacation. >> I mean it's like a it's like a steampunk thing or it's like you're doing it for the look, [laughter] you know? >> Yeah. Yeah. >> It's fashion. It's not like it's functional. >> Yeah, absolutely. I don't know anyone who does that. Um, but yeah, I remember that like when I was a kid, you know, my dad had the huge camera with the strap and everything and you go on the family vacation and half pictures are like eyes are closed or like it's just out of focus. >> A month after you a month after you take the photo, you get the film developed and then you figure out what happened at in the moment. Yeah. >> Yeah. Remember those little um there would be like these little like closet stands in mall parking lots where you'd go and get the film developed. You remember that? >> Yeah. like that. [laughter] >> Yeah. >> Yeah. And [clears throat] then you >> Now they're all Dutch Bros coffee shops and stuff. [laughter] >> Yeah. [snorts and clears throat] >> Is Dutch Bros still going? >> Oh, yeah. >> Is Dutch Bros still a Is Dutch Bros still a popular stock? Do people still talk? >> Not as popular as it was on the IPO, but >> yeah, >> I think I mean it's still probably legit. I don't know. I remember when it first came out, I looked at it and I was uh trying to underwrite like what growth rate was implied in the current price and it was like basically you had to know for sure that it could go east of the Mississippi successfully and you know brands just it's hard to tell if brands are going to can go national or not. I mean Seas Candy couldn't push that direction uh but it was killer in California. You know Dr. pepper is like 20% of the soda market in Texas, but like >> 1% everywhere else. Like there's no guarantee that brands can travel. And so that that seemed like kind of a dangerous bet to me when I looked at it. >> There's a psychological component to taste >> that in chocolate and soda and stuff like that. That's just insurmountable. If you if you didn't like get exposed to it as a kid in a happy scenario, then it just doesn't >> it didn't. >> There's nothing special about it. >> I talked about >> Oh, go ahead. No, no, go ahead. >> I was just saying Buffer talked about that with Coke. Like you go you went to baseball games as a kid and you remember the taste of Coke. Same thing. >> I was going to say it's easier though when it's like uh enough caffeine and sugar and fat to like, you know, kill a horse than >> Yeah, [laughter] >> those are keys. >> Those help. Yeah, >> some addictive component. Um, let me give a shout out and then uh JT will take it away. Uh, Brown, Marabu, Marabos. Is that Toronto? Tombble, Texas. Tyler, Goththingberg, Sweden, [clears throat] Peditva, Lausan, Switzerland, Tallahassee, Jupiter, Florida, Serbaton, London, the main streets of Jared's Cross. I love it. Yorkshire, Englean, Nashville, Tennessee, Toronto, >> Havtown, Pennsylvania, >> Belleview. I saw a good one down here. Valareerezo, Snomish, Forzy. Am I saying that correctly? >> Stay dry there. It's It's quite quite wet in uh Washington right now. Good luck. >> Is Forsome Is that where the Is there a racetrack in Forzome? Am I forgetting? Let me know. That's a good spread. That's fun. Thanks everybody. Uh JT, take it away. >> All right. Today's veggies are a synthesis of actually two fascinating recent reads that I had. And um the first was called collective hallucinations and inefficient markets. And it was a deep dive into the neuroscience and psychology of hallucination and how the brain is constantly predicting and reconstructing the world and how perception itself is basically like this controlled hallucination and and really they arise uh from these predictive signals that kind of drift away from grounded sensory feedback. And then the second was uh an interesting piece that was about AI as the new shadow banks. And you know these obviously couldn't be more different on the surface. The second article uh explored the financial ecosystem growing up around the AI infrastructure, chip supply chains, data centers, uh credit structures around them, all these speculative expectations around compute demand and the way that capital markets are behaving almost like a new shadow banking system that we we kind of had in uh you know the derivative world of of mortgage back securities in 2008 for instance. Uh, and so as I'm reading them, like something kind of clicked together about them, and I thought I might share that. Um, and really one is about brains hallucinating and and the other is about markets hallucinating and and they're actually kind of describing the same phenomena but at different scales. So both the mind and markets are predictive engines. They're both constantly guessing about what comes next. And and they both hallucinate when prediction outruns reality. So today we're, you know, we're going to explore this kind of extraordinary parallel between these systems. And they're really five core ideas that I want to use to tie everything together. And those are overbelief, illusions that become real, narrative gravity, sensory poverty, and why hallucinations sometimes can drive progress. So we'll start with the first one, which is overbelief. Uh, and in neuroscience, hallucinations don't appear because the brain is malfunctioning. they appear because this predictive machinery it becomes too confident uh and and it's always filling in these missing information pieces and constantly smooing smoothing out the gaps in sensory input. So for instance like you've got a literal blind spot where the optic nerve exits your eye and connects to your brain and there's no sensors there at all and yet your visual cortex just paints over that whole area with so cleanly that you never even notice it like it just fills in the gaps. Um, and under normal circumstances, reality is constrained to those predictions. But when when those constraints get weakened, the brain leans harder on on its prior uh estimations. And so it trusts its own internal model more than the external world. And that's that's what overbelief is. So now let's like think about AI like uh you know an LLM works in a lot of the same way like it's trying to predict which is the next token and you know whether it's grounding data is weak or absent retrieval isn't available the model leans harder on its statistical expectations and it fills that gap in with patterns that kind of quote unquote make sense internally even if they're maybe not true and that's how you end up with hallucinations. And then in markets, perhaps the most powerful overbelief machine, uh investors predict future revenues and earnings and the and companies predict future demand for things and governments predict future technological competition and um and when when real cash flows or so sober feedback is thin, markets become trusting of their own narratives at that point. So AI infrastructure with chips and data centers, power buildouts, it's getting financed today, not because the demand is is known and verified. It's because the expectations of demand are so strong. Uh and when these internal predictions outweigh the external constraints, all three systems, you know, the mind, models, and markets all hallucinate about it. So the second [clears throat] idea here is one uh is illusions that generate real consequences. So in humans, hallucinations don't just stay in the mind. They trigger bodily reactions. If you hallucinate a threat, your nervous system is going to respond and queue up as if that threat was real. You know, your body really enacts the hallucination. Uh the financial system takes this phenomena to the next level, which is investors believe in this unstoppable demand for compute. Billions of dollars are flowing into GPU factories, etc., uh cloud providers, and this ecosystem reinforces everyone else's belief that this is a real thing. Uh and and then something really astonishing happens like that belief becomes infrastructure. We have it now like it gets built. Uh and and that imagined future becomes a physical one. Um and even you know AI kind of participates in this loop as well like a hallucination maybe but with a persuasive explanation gets posted online and then later future models pick up you know they scrape that and they turn that into a factual source. So you're kind of like bootstrapping yourself uh into a into a belief or what is quote unquote true. Um so in each case whether it's a brain or a model or the markets these illusions escape the confines of prediction and alter the world itself in a reflexive way. Uh and so [clears throat] let's let's jump to the next one. Um this is the what you might call narrative gravity. And this is the idea that you know there's this pull towards stories that feel coherent. We all want that. And it shows up everywhere. You know, the mind is instinctively looking for patterns and explanations. Even, you know, with incomplete sensory input, it stitches together a narrative that will make sense internally. AI does the exact same thing with trying to hit optimization targets. Uh, and then in markets, narratives are even more powerful than numbers often. And the story is often the asset, right? Like this compelling narrative, AI will transform everything. Compute is the new currency. Uh, GPUs are the new oil that attracts capital. long before the revenues ever appear. And that narrative gravity explains why hallucinations become so believable. Uh and then the fourth idea is this idea of sensory poverty. So in environment uh where there isn't like let's say darkness or monotony or sensory deprivation, you will then uh that sensory grounding in the external real world is weakened and the brain will then generate its own version of what reality is. Right? That's why you see the monster in the dark. uh AI systems also do the same thing uh when they're you know when feedback is scarce they you know they'll skip citations and retrieval and just come up with something uh and the models running kind of in free mode uh and its output you know is kind of cranes its own input and it just like goes off on its own financial markets they do the same thing you know when fundamentals are weak if in in early technological cycles like AI like uh the actual revenue today is is tiny customer demand is totally uncertain unit economics are immature like we have no idea what it actually costs to run all these these tokens. Uh we must extrapolate then out from this very sparse data and in the absence of a of a sensory feedback you know the narrative becomes the only thing that really matters. Eyeballs during.com is a good example of this. All right and [clears throat] then now the fifth idea um and this is the maybe more surprising and maybe optimistic version of things is that these hallucinations often are how progress happens. Creativity itself is, you know, a structured hallucination. The brain's ability to imagine what does not exist uh is what produces hallucinations under stress, but it also leads to creativity. AI does the same thing. Um, you know, it's it it it can be a source of novelty, which is actually pretty impressive. The models can make these leaps that maybe humans hadn't thought about yet. Uh, and in finance, hallucinations are this kind of hidden engine maybe behind a lot of the major infrastructure revolutions. You know, people were speculating to fund railroads that are still in use today. Uh speculating to build fiber optic cables that we're still using today. Uh speculating on, you know, electricity grid decades before there was any profitability. And the same thing with cable systems, you know, in the ' 60s, '7s, ' 80s. Uh and all these things happened and maybe it's happening today with AI and compute. Uh and even if the demand collapses, maybe the infrastructure remains and you know, we we're all going to be better off for it. Uh now of course we'll leave this this uh we'll ignore the depreciation schedule controversies around all this that you know maybe throws a wrench in that but let's let's zoom out a little bit and kind of wrap this thing up. Um what we've covered is a shared architecture across systems that seem unrelated. Brains hallucinate perceptions. AI models hallucinate language and markets hallucinate future cash flows. And all three are hallucinate for the same structural reasons. They're predictive engines operating under uncertainty. And [clears throat] when predictions outrun their grounding, when feedback is weak, when narratives are strong, these hallucinations appear, not maybe necessarily as malfunctions, but but as the natural extension of inference. And the deeper insight is really this, that hallucinations aren't errors maybe to be stamped out necessarily. They're really how complex systems feel their way into the future. And even as they distort and mispric, and sometimes people get hurt. Um, the job of an investor in this AI cycle, I think, is to stand inside these hallucinating systems, minds, models, markets, and keep asking, is this the kind of illusion that that leaves useful infrastructure behind, or is it just another story that ends up on the shelf of uh the library of mistakes, which we talked about last week? >> Good one, JT. It's a so if it's a it's a hallucination if it turns out not to be true, but it's a vision if it turns out to be. [laughter] >> That's right. And if you can figure out the difference a priority, then uh you'll you'll do well. >> You're the Messiah. >> Yeah, exactly. >> Are there other animals that experience hallucination or is it a uniquely human thing? >> I would imagine that other animals I don't know for sure because I think it's kind of hard to tell subjectively, but I would imagine that they're there's hallucinations in other animals. I mean, they're making predictions the same as same as we do about their environment. Um, I mean, other animals dream, so that's probably another data point that that might be uh point in the direction of hallucination. >> Gotcha. Yeah. I just was wondering if maybe there's a unique human thing towards trying to visualize and predict the future that's unique to us. >> Um, >> no, I mean, every every animal has a working model of their environment and kind of how they fit into it. And there's that sensory that expectation is always crashing against the sensory input to see like what surprise is which is like oh my model I need to update it. >> Gotcha. Okay. Interesting. Well, a lot of it sounds like an explanation for the momentum factor if you think about it. >> I posted about momentum factor today. It's been very very strong. It's been, as you'd imagine, it's been strong for a very long period of time, but it really took off in 2000. I contrasted it with the equal weight versus S&P 500 which has collapsed you know relative to the market capitalization weight index which is up a lot which is why it looks like it's collapsed but just the average stock is not doing very well while the the big end of town is doing very well. >> If you had to bet if you were a betting man over the next 10 years would you bet S&P 500 market cap or or equal weight? I'd bet equal weight because if you can look back over a hundred years, it's clear that small beats large and equal weight is just one expression of that. You can you can create the same chart with like S&P 100 versus 500 where in that instance 100 is the large like 500 is the large relative to the equal weight. Like there are a lot of charts. I think we've discussed this before where you've got to decide whether you want to see >> do you want the thing that's worked over 100 years or do you want the thing that's worked over the last 10. >> And I think you can I think you can see like I posted another chart today which was the Russell 2000 versus the S&P [snorts] 600. So the 600 is the small cap index and it is the 600 that's outside the 900 that includes the 500 that's large and 400 that's mid. And so there is like 500 stock potential overlap with the 2,00. So they're not directly comparable, but there's some overlap. The big difference is that the S&P 600 has these profitability requirements for inclusion in the index, whereas the 2000 is just all comers. And the 600 as a result in sort of a little bit of a defiance of that large over small because this is the profitability factor is so much more powerful here. the 600 has massively outperformed the 2000 since like 1990 I think that chart goes back to so it's 35 something years but you can clearly see into in it there are these like distinct little heart attacks in 2000 where there's a massive spike to the 2000 over the 600 and then in 2020 there was another big spike to the two to the 2,000 over the 600 and we're currently in another big spike of the 2000 over the 600. And if you pull up the 2000 and you look at the largest names in the 2000 or representative names in the 2000, like it's a very >> inspiring list of uh >> junky index. Yeah. There's a lot of science experiments in there. >> There's a lot of stuff where they're they're pre-revenue, you know, they're still trying to figure out they got a vibe rather. They got some got a logo and some branding. [laughter] >> They're telling people what they're planning to do sometime in the future and people are like that's a $25 billion enterprise. But yeah, the contrast is pretty amazing. Like I was looking at the returns on um like the S&P 400 and the S&P 600 and I think they're both up like 7% year to date versus the S&P 500 which is up 17% which is 18% now which is pretty insane. Pretty insane. >> That's a pretty big delta for things that are not that far apart from each other otherwise. >> Yeah, >> it's very much a sized driven market. The bigger the company, the better it's done. But I don't know if that's because when you get these like technological revolutions, you get new companies going to the forefront and that's what drags the whole index up or if this this is like a 2020 style, you know, just a we call we call it a meme stock market, but it's like every every boom is a meme stock market because there's some narrative that drives >> interest over, you know, the underlying fundamentals delivering or generating the returns. And so we're back into that kind of market right now. >> I mean, >> if you look back at the last major bubbles, like say the late 90s, Japan in the 80s or in the nifty50, pretty much all of them were large cap phenomena. >> So it makes sense that that's what would be happening now. >> But is the is the has the meme moved into the large caps in this instance in a sense of like the AI >> Nvidia? >> Well, yeah, maybe. Like clearly there's a there's a fundamental story there as well. It's very very impressive on a fundamental basis. >> 80% gross margins. That's a >> plus the growth >> and the seemingly like infinite demand like just endless demand for these things. Like they cannot produce enough. >> Yeah. >> But it does seem like there's a little bit of a defection from that ecosystem. Like >> Google's training making its own chips and training them. I think Tesla is too, but maybe it's doing it with Nvidia. It's there's so many announcements it's hard to kind of figure out from the headlines what what is real and what is not. Apparently, we're selling to China now, too. But, uh, with a 25% rake [laughter] >> for the big guy. >> Yeah, sure. >> 25% for the big guy. >> It does. It's hard to imagine that Google, Microsoft, Apple would just sit around and say, "Well, I guess we'll let Nvidia do all this." Like, it doesn't seem like that sustainable for that long. Like, they they're going to compete on some level. Yeah. Well, that I mean that the argument though, the counterargument might be and this was the argument for Intel in >> 2000 to really until they started pooping the bed in whatever 2020ish. But it was always like, okay, Intel has can spend the most money on R&D and spread it over the biggest amount of fixed cost. How is anyone ever going to catch them? Like they can just >> How did anybody catch them? I know it's a great question. I mean that's that's a trillion dollar question there. >> Only the paranoids have off. >> I want semiconductor. They were able to do it. They were able to do it cheaper. >> Yeah. So what's the that's the lower labor, >> you know, and kind of feel like Nvidia's got similar. They got Intel 2000 vibes right now of like infinite demand, you know, 90 plus% market share. No one can catch them on R&D. They're so far ahead. Did anyone listen to his interview on Rogan? I haven't. Have either you? >> I just saw some clips. No, >> I don't think I've ever heard him talk before. >> Yeah, he apparently did a crazy interview. All I've seen is the tweets about it, but apparently he claims he reads 7,000 emails a day and he needs some [laughter] crazy claims. >> Is there room for that in the schedule while you're signing boobs? I feel like that's like [laughter] >> I didn't want to mention that. >> I should have I shouldn't have said that. >> That's that's a >> Toby take that out in post uh production. >> When people brag about that stuff, I just I find it so weird like I read this many books in a year. Like I would have no idea how many books I read in a year. Why Why would you go and count the number of books that you read in a year? That seems like bizarre behavior. Why would you have that much time in your day? >> And are you really aren't you reading books? >> Yeah. You don't have to take your shoes off to count that high, do you? [laughter] Like if you're reading if you're reading that many like when I hear people make that claim like how are you possibly retaining the information from the book and really internalizing it. >> He's getting like the Wall Street Journal headlines forwarded to him and he's getting all of his he counts every coupon 25% off. That's one more email read [laughter] the headline. >> All the junk mail that counts too. >> Yeah. So that's pretty bizarre. The other thing I've noticed about Nvidia, maybe Toby you would know more about it, but I've just seen the Benanish MC score is trending towards the earnings manipulator >> territory. Did have you looked into the components of that for Nvidia or anything? There there are lots of there are lots of people out there who say that they're doing a little bit of channel stuffing and that they're financing some of their customers and there are lots of things there there's lots of rumors like that out there. I don't know how I don't know if you have so much demand why you would be doing that stuff. So that is a little bit of a funny story but the demand seems to be real. >> Mhm. But then there are those stories like that. I don't know. >> I think I would, if I was him, I'd be going I would hit the at the money equity issuance. I mean, they're doing that with SBC right now, but I'm I would be like, I'm going to bring in >> a trillion dollars onto the balance sheet right now. >> Doesn't matter. You could always use it. I mean, this you probably at some point will be glad you had it. >> And in this market, people wouldn't even care. No, >> they should be outraged, but [laughter] I think in this market it's it's makes a ton of sense and you probably get away with it. >> The the only point that I get to looking at those big stocks is that they do seem to me to be h at they're doing very very well. >> The business or the stock price? >> The business. Yeah. So the business is doing very very well but the stock price is already you know discounting 25 years of maybe not 25 years but the stock prices ahead of them a little bit they look expensive to me and I and I don't know that that doesn't necessarily mean they crash like it just means that they're a little bit more sideways for a period of time and then there's other stuff in there where there's really not much you know if you go down the even in under the in the S&P 500 there's a whole lot of stuff that hasn't worked and then as you go further down 400 and 600, it gets, you know, suicidal. They're they're they're kind of looking they're looking pretty sad down there and not much has to happen. I mean, oil and gas is just Are we still using oil and gas or is the economy just moved on completely? >> Pretty [laughter] sure we're still using it >> because it looks like it's being given away for free like you go and buy a bottle of water and you get you can fill your tank up at the same time at the >> That is quite impressive, isn't it? I've always been amazed by that. Like you can literally move what, you know, 1,500 lb of vehicle 40 miles for the same price as like a as as water. Like, wow, that's impressive. >> You got to pull it up out of the ground and process it and then >> boil it to oblivion and then >> Yeah, it's amazing. I always think about that when people complain about the price of gas because yeah, it sucks when gas prices go up, but at the same time, you got to zoom out and think like, look at how cheap this is compared to what it allows me to do to what most of my ancestors never even could imagine. It's it's pretty incredible when you think about it. >> Yeah. The the megajoule per kilogram, so the energy density of fossil fuels is just absolutely incredible. Can we if we don't have jet fuel, can we fly? Like our batteries aren't our batteries aren't energy dense enough, are they? >> No, they're like uh I did the math on this at one point and like they needed to be like 30 times denser than where they were at the at that time to get to the same uh energy density as jet fuel. >> We we are improve I we are improving them. Like I remember seeing a this is a few this is five or so years ago now where I saw an article that said that if you use the same battery out of an old school laptop and you put it into a modern laptop it lasts for like a second. >> It just torches the battery. >> Is that right? >> Instantly. So because there had been this this article was answering this narrative that had been out there that there had really been no improvement in battery technology for a very long period of time. And I don't know whether they meant like just literally we're still using the same style of battery, but clearly the the batteries are much more efficient and better than they used to be. >> Like to 30 times how far into the future is that in terms of improving battery technology? >> I don't know. I mean, um certainly >> is it impossible or is it just like wait 10 years and we'll we'll be doing it? >> That's above my pay grade. I don't know. It might even be like much closer now. Like I wouldn't be surprised if there's some Moor's law that's happening there that >> because there's some veto, right? There's those EV >> Yeah, >> Eevee >> is one of them. >> I could see how people could think that the batteries were better because like my old Nokia brick could last a week, you know, and I you wouldn't have to charge it. Remember that? But >> but yeah, but I mean at the same time, okay, my iPhone has to get charged every day, but I mean, okay, my iPhone is like hundred times better than the old Nokia. That might be underestimating it. >> Yeah. I mean, >> thousand 100 thousand >> and the battery has had to advance tremendously to just be able to keep up with all the new features and tech that's in it. >> BSG, do you have any uh you you had a weird portfolio for a while. Do you want to give us an update on the weird portfolio? What what is it? What goes into it? Oh yeah, it's it's it's basically like a risk parody style portfolio composed of treasuries, gold, and then um basically small cap value funds and a real estate fund. Um it's a vehicle for a lot of my savings. Um and it's actually having a banner year. It's up 20% this year. >> Yeah, gold's helped, huh? >> Yeah, the chunky gold allocation has helped. And then um international small cap value has actually >> quite well too. Yeah. >> Yeah. So like I own AVDV in there and that's up um over 30% this year and gold is at a banner year. So and I mean the philosophy behind that portfolio is to have something that can basically compound at a decent rate over the decades and then be able to handle multiple economic environments. So you want it to be able to deal with recessions. You want it to be able to deal with booms. You want it to deal with strong dollar, weak dollar. Um and there's components in the portfolio that should do well in all those environments. um it'll never be like the best performing portfolio ever, but if uh you know for me I was just looking for something that would be able to control the draw downs and it was almost like my spin on a passive portfolio. Like in a lot of the financial independent circles they all always talk about like just going pure market cap weighted and to me that didn't make a lot of sense. I wanted something more robust than that. I wanted something where it's not going to decline 50% in 2008. it's not going to have a lost decade. Um, yeah, and that's been the whole philosophy behind that. And it served me pretty well. I mean, really, it just helps me stay invested is the big thing. Um, when I'm freaking out about something macro happening, I can look at that portfolio and say, >> at least gold's going to work or at least this is going to work. >> It's so nice. It's nice to have had something that's rallied over the course of a year when you've got something like small cap value which has been crushed and you've had something that's rallied and you need to rebalance. taking away from >> Yeah. How much of the return is from that kind of uh farmer's fable rebalancing? >> Yeah, it's a lot of the return comes from that. So, you do generate excess returns by um rebalancing uncorrelated assets. Like that's the whole um Shannon demon Shannon's demon thing where like you could have two assets that don't return anything, but if they're unc over the long run, but if they're uncorrelated and you're regularly rebalancing, you can eek out a return from that. And you see that like if you were to just average the kagers of all the different assets in the portfolio, the actual kagger, the actual result that you get is always higher than just the sum of the parts because you do have those times like today when I'm when I rebalance that portfolio, I'm going to sell gold and I'm going to buy US small cap value when it's cheap and you know, potentially small cap value will have a good year or in a year like '08, you'll have long-term treasuries up 20%. %. Okay, well now you can buy some stocks um when they're cheap and that does produce a good effect over the long run. You limit volatility and you get a result that's more than the sum of the parts. >> Uh just on recessions, I I I know I've been talking about this for a long time, but the 103 inversion seems >> Well, where are we? >> Inverted again. >> Well, we we uninverted about a year ago, something like that. And then we've been like basically flat where the 10 and the three are trading at the same level for for even though they've both moved around quite a lot over the last 12 months they've been >> in >> um yeah in sync well they they've done it but just over the last few days they've diverged uh go going back to normalization pretty marketkedly. So, you're telling me there's a chance? >> I don't know what that means, but the the uninversion the uninversion seems to be here for the moment. >> I remember you had Cam Harvey on talking about this. That was an interesting podcast. >> I mean, we inverted in in 22. That's how long we've been inverted. It's the longest inversion by a mile on the record. I mean, on the data that we have, which is >> That's a lot of blood flowing to your head when you're upside down for that long. Huh. >> I don't know what that means. It's like it's a strange thing where the the front month yields less than the or the front month yields more than the 10 year. >> Mhm. >> Not good for banks. Maybe maybe banks look a little bit better when with the uh inversion gone, but it's also the harbinger of the recession. It shows up 3 to 6 months before the recession. >> The uninversion is what is the trigger is what >> the uninversion is the trigger. Yeah. And is the extent of the inversion is that an indicator of prob of the severity of the >> No, it doesn't seem to be there. There's not enough there's not enough data point. There's the n is like eight or something like that. So, it's not it's not big enough. >> But that won't stop us from extrapolating. >> Well, there [laughter] doesn't seem I just there doesn't seem to be although the length of the inversion does seem to have something to do with the length of the recession. >> Of the recession. Okay. >> And so, it's been a very long inversion. So, it's possible we're in for a very long recession. Here comes nuclear winner. >> Like I've been saying, there's been an earnings recession, and this is another weird thing from 22 where the 22 was just all of that profit being pulled forward, but there's been a little earnings recession since 22 to date. Certainly in small and mid >> and when you say that, that's EPS. >> Yes, EPS. Thank you. Yes, it is. Earnings per share. Yeah, >> earnings for share for uh for large topped out in uh like beginning of 22 as well. Haven't haven't quite recovered that ground yet. SBC does count for EPS. >> Wow. Really? It still hasn't recovered. >> No, it's below. >> Wow. >> Below where it was. >> Narative narrative violation. >> There's a little bit of [laughter] a narative vi violation, isn't it? Yeah. >> Oh, yeah. You wouldn't >> I was surprised by that, too. >> Interesting. Uh, dude, BSG, we're up on time. If folks want to follow along with what you're doing or get in touch, how do they do that? >> Um, yeah, you could go to um securityanalysis.org or valuestockgeeek.com directs to the same place. That's my Substack. That's where I look for wonderful companies at wonderful prices >> and I host a podcast on the same platform. >> Yeah, it's very good. >> Yeah, I'm a fan. you great ass >> JT uh promoting anything? >> No, just any wise words, >> human kindness, I guess, coming into the uh coming into the holiday season. And I think something that stuck with me from the last Bergkshire meeting was Buffett said like if if you've been lucky in life, then you you probably owe it to to be make sure someone else feels lucky. So, uh, do what you can to, uh, help your fellow human right now. >> Uh, we'll be back next week. Uh, we'll figure out whether we're here the week after that, but we'll certainly be here next week. So, we'll see everybody next week. Same time, same channel. Thanks for tuning in.