Dan Niles: The Bubble Is Going to Blow Up — It's Only a Question of When

Born in Sri Lanka. My parents came to the US when I was about five and a half years old. Dad was going to school at Boston University to get his PhD in theology. He was a minister, and so I lived on the East Coast for most of my life. And then I went to Boston University to get my systems engineering degree, and then went to Stanford to get my master's in electrical engineering. And that's kind of how, yeah, that was my start.

Well, there was a girl.

Yeah. The summer before college, I went to — they took the top two science students from every state and they put us all in a camp together in West Virginia. It was called National Youth Science Camp. So, I was representing Massachusetts, she was representing Minnesota. We met there. Absolutely hated each other to the point I threatened to drown her if she splashed me again when we were like in the swimming pool because I was cold. Just absolutely hated each other. But we ended up going to Boston University. We were both there for electrical engineering. And so when I was there, her mother said, 'Hey, go look her up. Go look them up.' You know, some of my best friends are people I didn't like in the beginning. Met there, became best friends, started going out. First girlfriend, got married. I've been with her since I've been 18.

And yeah, I'm 59. So, and we've been together ever since. And so she left electrical engineering, went into finance, and then started, you know, was in this finance class, investing class, introduced me to it. So we opened up an account together. I'll forever be grateful to Merrill Lynch because back then, if you didn't have enough money to buy a complete one share of a stock, they would pool your money with other people so you could buy a share. We like between the two of us, we scraped up like a hundred or 200 bucks and we started that. And I realized like I love doing that. And so then I ended up when I graduated, I'm like, I don't think I want to be an engineer for the rest of my life. So I went and got my master's at Stanford, but all my electives I took in the business school. And so then when I graduated with a master's in EE, you know, I was like, well, let me try this investing thing. And I got a shot to do it, spent four years in investment banking at a small West Coast firm called Roberts and Stevens. And then I was like, you know, I really hate this banking stuff where you have to suck up to all the companies to get deals. And so then I went into research, sell-side research, and I enjoyed that, but I was like, you know, the fun part is figuring out what companies you want to invest in. I hate having to call up a hundred top institutions and go visit them and tell them the same story of why they should be buying Dell Computer or Intel. So I'm like, so when I got the chance, I started a hedge fund, and that's what I've been running ever since. And so that was back in 2004. So I was on the sell side for about 10 years, and then I switched to doing that.

Yeah. No, I mean it's great because I learned a lot from it. So the first stock was probably the worst thing for me because I bought this company called Standard Oil, which you may remember, the John D. Rockefeller original company, because we were doing work in the — this was 1987. And so I was like, 'Oh, you know, I think like, you know, oil companies are really important as the economies are expanding, all that stuff.' It immediately gotten taken over by British Petroleum. And so I was like, 'Oh, this is easy.' The stock though that really helped me was my next investment where I bought this company called Worlds of Wonder. Now, back then, just so you remember, this is 1987. And so this company made a talking teddy bear. This is 1987. Like that's a huge deal. And then they made this game called Laser Tag where you're shooting each other with lasers, and if you're a kid, right? Like you're always running around doing that. And they were two of the top 10 most hot toys going into Christmas. So I'm like, 'Okay, this can't move.' So we put our money into that going into finals, and then I finished finals and I'm looking in the newspaper because you know there's no internet and none of that. So you're looking in the newspaper. I'm looking for the stock symbol. Where is it? I don't see it. They had filed for bankruptcy on December 22nd. So, we lost all of our money that we put into this thing. Now, you go, how on earth do you go bankrupt going into Christmas with two of the top 10 selling toys in the world? The management team hadn't focused on production, distribution, and so the companies couldn't get — Toys R Us couldn't get the toys to sell. So with two of the best products on the planet, they went to zero.

And that's a lesson that has been hugely helpful ever since then because it shows you the importance of not just the product but the management team and the whole business around that. And that's, you know, I even got that lesson this last year where, you know, I bought a company because it was deep value, was in the optical space, thought they would do well. And I remember thinking, this company's been around for a while, not done anything in 20-some-odd years, and you can look at it and go, every other company has done great. And you know, the management team — I should have, you know, I remember thinking that going, but this thing's like at a fourth of the valuation of everything else. But it was there for a reason.

Well, you're, you know, you're a kid, right? You're in college, you're starting off. And so it was more the Peter Lynch stuff because remember I'm growing up in Boston, right? Boston University, Fidelity's there. Peter Lynch is a legend. And so, you know, it's invest in what you know, what you feel like you have some understanding of. And so I'm just reading a lot. You go to the library, you know, you're pulling out books and that's how you're getting involved with that. And you know, I always loved, you know, technology and I was like, 'Oh, talking teddy bear, you know, laser tag, like this stuff should be hot.' And obviously now, like all those things are pretty big deals. And so it was then, it was invest in what you know.

I was — so 1990 to '94 I was in investment banking at Roberts and Stevens. Then from 1994 to 2000 I did sell-side research, so writing reports on Intel, you know, and the computer hardware names as well. From '94 to 2000 I went to Lehman Brothers, from 2000 to 2004, and then Lehman seeded a hedge fund in 2004, which did really well right up until it didn't.

Yeah. No, I mean I actually thankfully thought there was a problem at Lehman because if you looked at the charts back then in like '06, you could see that the housing prices were starting to slow down year-over-year and the inventory of homes out there was starting to skyrocket. And I remember going, 'This doesn't look good.' And so as soon as my stock would vest, I would sell it. And I remember it looked really stupid to be doing that, right? Because you get a bunch of stock for compensation, it would vest over time and then you'd sell it as you go, and it looked really dumb right up until it didn't. So thankfully that's what — you know, the hedge fund did fine, but Lehman itself obviously went to zero. And so that was just, you know, being there running it as this thing is imploding around me and everybody's withdrawing assets and, you know, like it was crazy.

No, I mean it's fine because, you know, you obviously have a track record. You know, people always wonder, well, how does a money manager perform in a crisis? Well, it doesn't get to be any bigger crisis than that. And so I think, you know, that's not — in some ways, a manager that hasn't been through a crisis — I always tell people like you need to judge a manager through a cycle, right? In an upcycle, everybody looks like a genius. It's, you know, when things get ugly, like can somebody protect you?

Yeah. I went through the dot-com as a sell-side analyst, right? And then I went through, you know, obviously was managing money starting in '04 and went through Lehman failing while I worked at Lehman and, you know, went through, you know, the multiple crises after that that we got. And so you've had COVID, you've had the failure of, you know, Credit Suisse back, you know, a few years ago. And so there have been a few, you know, obviously major speed bumps along the way. And, you know, there's going to be more. So...

Yeah, I still do because, you know, you go back to that internet infrastructure buildout and from end of '94, which is when Netscape Navigator came out, and that's kind of the ChatGPT moment for the internet bubble, right? The first internet web browser. What the heck's the internet? Like it was a web browser. And so from the end of '94 to the peak in March of 2000, the semiconductor index — and I'll just pick that one because it's kind of very analogous to what's going on today — went up 850%. But in the middle of '95, right around September-October, the semiconductor index had a drawdown of over 50% from about September-October of '95 till it hit bottom. Then it ran again, and then starting in kind of late '97 it had another over 50% drawdown that lasted also close to a year. But over that whole period with two 50% drawdowns punctuating that, it went up 850% from the end of '94 to March of 2000. And so I think what we're going through right now is another one of those periods caused by the fact that a lot of companies in March were saying, 'Hey, how do I get my engineers to use as many tokens as possible?' to now they're going, 'Oh my god, I blew through my entire budget in four months like Uber said, and now I got to get those costs under control.' And so I think you're going through another speed bump and it could be ugly. You know, it wouldn't surprise me if you have a 50% move lower in the semiconductor index before it picks up. And I think, you know, you saw results last week and how the stocks reacted to Google and Tesla and Intel and, you know, I mean, Intel really was surprising. And I think you're going to have to work your way through this. So it's not, you know, it's not going to be smooth sailing. I think it's a speed bump because agentic rolled around on January 30th, requires 10 to 100 times more tokens. I think it's real, it's going to change everything, and I think it's just too early for this to be the end of this buildout. But you never know, right? You only know in hindsight. But thankfully, I thought there would be a speed bump and that was definitely helpful predicting that back on June 20th.

I don't know that there's any of it that's overlooked. I think, you know, I think if you're asking kind of what the next things are, agentic — so when you had training first, and inference, that's very GPU intensive because you're doing the same thing over and over again, right? What's the relationship of this word to that word? That's training, and you're doing it for every word. That's GPU intensive. Give me the next word, predict the next word, predict the next word — it's the same thing. That's the inference piece. That's also GPU intensive. You get to agentic and you go ahead and say, 'Hey, I'm having this interview with Dan Niles. Go pull me all the videos he's done. Then go ahead and summarize it. Then go pull the stock charts. Then make me a nice PowerPoint deck that I can read when I'm getting on my airplane to go interview Dan. And oh, by the way, book me a plane flight, get me a hotel, book me the studio.' That requires a bunch of distinct steps that have to be organized. There's a path to that, and that requires a lot of CPUs. And so that just started on January 30th. So I think the CPU companies still have a lot of room to grow because of that. And then I think next year you're going to have a lot of optical because now, you know, nobody wants a big data center in their backyard. You have all these data centers geographically dispersed. You turn them into one gigantic data center by connecting them with high-speed fiber. And so I think next year optical is going to become, you know, a lot more interesting where right now it's all copper, right? And so those are kind of a few things that are, you know, percolating in the back of my brain, but it doesn't mean the near term they don't all get killed.

Yeah. Well, here's where it gets tricky, right? You have three vectors at the same time. You have the top 1% like Coinbase, which you just brought up. Their CEO comes out on June 26th and says, 'Hey, I went ahead and I'm now routing a lot of queries to these cheaper models. I cut my AI spend by 50% — almost 50 — but I'm still needing more tokens.' So you go, well, how's that going to play out? Because if you take the top 1% of companies, they're all trying to do that, right? Uber was the first shot across the bow in April when they came out and said, 'We blew through our entire budget for the year in four months.' Uber doesn't want to miss their quarter when they report, right? So they got to get their SG&A and R&D under control. So they're going to be cutting back. So that top 1% cutting back, but you have the other 99% — you, me, all the firms across America — we're using more and more as we're figuring out how useful it is. How does that play with the top 1% cutting back, but 99% growing? But it's really that 1% right now that's driving everything, right? How many hyperscalers do you talk about? There's like five, maybe six now. Like that's it, right? You got Amazon, Google, Microsoft, and you kind of threw in Oracle over the last year or two, and now maybe you throw in SpaceX, and then Meta looks like they'll be trying to launch their own cloud. That's what, six companies. How many firms do you have in the world, right? There's only six you talk about. So that's the balancing act that's going on right now, which is why I'm very curious to see what Amazon and Microsoft sound like this next week because Google I thought would be the strongest of the bunch and they were so far, and that their growth in cloud accelerated to I think it was from 65 to 82% or so this quarter. I don't think the other two are going to be able to grow as fast. And you saw how much Google got hammered because they brought up their capex, but their EPS — so they're increasing their spend, but the amount of money that people expected them to earn isn't going up that much. Stock got pummeled. Tesla was obviously worse. Intel surprised me because their capex went up, but by like 18%. But the EPS went up by 40. The stock still got hit, right? And so you go, okay, investors, as they rightly should be, are being very focused on: you're spending all this money, but are you actually earning more on that? So what's my return on invested capital?

It's a great question. I'm increasingly coming to this view, and it's been sort of percolating in my head because think about early on, right? Who was the leader? Well, it was undisputedly OpenAI, and Google was going to get killed, right? Search was going to get replaced, etc. Then all of a sudden, Google — you know, the founders come back, they start getting the company back on the right track, and then it's like Google's going to win. And then there was a period of time when Grok was doing very, very well, Elon Musk's company. And then it swapped. Then you had this DeepSeek moment last year and you go, the Chinese can actually do this, and they're doing it on inferior hardware because they're restricted from getting the best chips. And then who had ever heard of Kimi K3? Right. Like, and then what about Qwen? Like that's Alibaba's model. And so you have this constant change of who's in the lead that's going on right now. And I think it's going to continue. But where I think this is ultimately going is the model layer gets commoditized. And so these guys, especially because in the example we talked about before, right, Coinbase — how do they cut their AI spend by 50%? Well, I don't need to ask the PhD-level model what 3 plus 3 is, right? They can ask us, right? Or the PhD-level questions — well, let's ask the PhD-level model how to design the nuclear reactor, and for 3 plus 3 it'll give it to us, right, for a tenth of the cost. And so that's kind of what's going on. And I think ultimately the model layer becomes like electricity. It gets commoditized and the value accrues to the hardware infrastructure layer, and that's where more of the value is because you still need a chip. You can't like open-source that away per se, right? And so I think more of the value is going to accrue to that and then ultimately to the application layer, the companies that sit on top of it that we use all the time. And so to some degree that's where my head is right now and it'll evolve over time. But I think when you've got open-source at 90% less cost than the leading-edge models and companies having to control their AI bills, I think that's where it's going because you got this constant leapfrogging going on at the model layer.

Yeah. For me, I like Gemini, and I'm also fortunate that one of my kids, Warren, who's working on his own startup — he was, you know, hacking into things when he was in grade school, and he's always been interested in, you know, artificial intelligence programming before it became a thing. And so he's constantly testing out all the new stuff that has come out. He's, you know, he's in the Seattle area, but he'll call me and say, 'Hey, I read this paper.' And like I remember when Kimi K3 came out and he was testing it out and he asked it like, 'Who are you?' And it said, 'I'm Anthropic.' He goes, 'Well, this is Kimi K3.' So, you know, they distilled the Anthropic model. So, you know, he was telling me that and I was like, 'Oh, well, that's interesting.' And so it really helps because he's always reading the latest stuff, using the latest stuff, and so I get a very firsthand view from an actual user of this thing in the weeds on what's the best. And but as he said, like it keeps moving, right? And he moves from what he's been using for his own startup, like the models, and right now Gemini is the one he likes the best. Interesting because he says for somebody who really knows what they're doing — at least this is his view — like Gemini, like you can get into the weeds. And he's like, for one layer up from that, Anthropic is probably the best. But you know that that has changed over the last several years too, and it may change.

Oh, that's easy. He focused on the consumer. Will we pay for anything? We get social media for free. We do Google searches for free. Amazon — you know, you're paying for the product, but you get to use the app for free. Why are we going to pay for artificial intelligence? And so only 5% of his customers pay. Anthropic looked at it and said, consumers won't pay for anything. Let's focus on enterprises. Enterprises, they'll pay for stuff.

Business to business. Which is why I think that's where the screw-up was. Which is why now you've seen OpenAI completely pivot and go — and by the way, the new coding stuff they have out is fantastic, but they started late.

So if you get that person locked in, it's going to be hard to get those engineers to switch to something else because each model, the way you interact with it, is a smidge different. And so I think that makes it hard to dislodge Anthropic. Not to say that they can't, because I also look at it and there's some regulations — like in our business is highly regulated in what we can do and what we can't do. And if you think about most corporations, what do they have? They have Microsoft Office. So, Microsoft's trying to force-feed Copilot to them and then what that sits on top of. And you go, well, if it's offered to your company in a suite of businesses, you don't have to pay separately for something else. So, you know, I kind of do wonder how Microsoft evolves through this because the bad news is if you're an investor — so if you look at it from an investor standpoint, you go, OpenAI has a problem. Microsoft owns 27% of that. They've got an issue. I ultimately think, and we'll see what happens, is that I think Microsoft at some point buys out the rest of OpenAI because I think OpenAI, if you're going through a model commoditization layer, they end up in trouble because they're squeezed between Anthropic and corporate Google and consumer, which we all trust and have abused Google for decades. And I think OpenAI gets squeezed between those two. Not to mention now these cheap Chinese models. And I think that's what happens. But like who knows, right? Like you never know who's going to survive these things or not. But I've said for a while I think OpenAI is going to be the single point of failure when this thing ultimately blows up. But we'll see.

Well, Anthropic is this year. But you're right, OpenAI. It looks like they postponed it till next year. And I look at that and I go, you know, they need a lot of money. Why would they have to postpone it? Well, probably because they don't feel like they can get public right now, which tells you how like the world has changed, right? But that happened during the internet too, right? You had Netscape. Where's that? You had Lycos, Excite, you know, AOL. Let's not forget that, right? You've Got Mail. I love that movie, by the way. It's one of my favorites. But like, where is it? Right. So you look at all this. So that's one of the big fallacies is that I hate it when people say, 'Oh, you know, just buy it and own it forever.' Because like, how do you know? You can't do it. Like the leaders change. Nokia had 40% market share in cell phones. 40%. They don't make cell phones anymore.

And that was going to be the king, right? The CrackBerry, right? That was a nickname because it was so addictive. I loved that. But you know, where's that, right? So I think you have to be adaptable. It's one of the things that I try to tell people is like, yeah, you hear the stories of say, 'Oh, I always knew that, you know, Apple was going to be great.' Well, it had a near-death experience. By the way, Amazon survived, but the stock went down 95% when the bubble broke, right? It survived. And so you got to be flexible and, you know, protect on the downside because like when I downgraded all the stocks I covered in 2000 because I thought we were going to have an issue — by the way, part of it was oil prices had surged. People forget that in 2000. I would have never guessed that NASDAQ would go down 78% from peak to trough. It happened over two and a half punishing years, you know, and so you have to be very adaptable and flexible as you go through things.

Here's the thing. I'm a major Elon Musk fanboy because I look at that guy and I go, we haven't seen somebody like him since, you could argue, Leonardo da Vinci — not even Edison — because, you know, he's making literally blind people see with Neuralink, right? Like the last guy that did that had a religion pop up around him, right? So he's doing that. He's got the Boring Company. He's putting up reusable rockets. Like we thought that was possible — certainly not NASA, you know? He's got electric vehicles. He's got all these different things. And he's just like literally willing companies into existence. So normally I would say if it was anybody else, and this goes back to that leadership thing with Worlds of Wonder, I would think that they might have a problem. But it's Elon. And I never want to bet against that guy being able to change something.

No, I think artificial intelligence is super important and it's central to a lot of the things like robotics, electric vehicles, all these other things he's trying to do. And so, I mean, he's building the whole infrastructure layer from putting the satellites into space, having the artificial intelligence, all of this stuff. And so yeah, normally I would think he would have a problem, but he's also got a lot of training data, which is important, right, on Twitter. So he's got all this training data for free.

And that's the other area that OpenAI and Anthropic had problems initially is like Google's got the most training data on the planet because we get to use their stuff for free, but they get to train on us using it for free, right? And that's a huge source of data for them, which is part of the reason why, you know, you can't count Grok out for the Elon factor plus all the data they have to train on.

Well, there's two different questions. There's the stock and then there's the company. So again, I'm an Elon Musk fanboy and so I go, I'm sure it'll be successful, but I doubt it'll be successful in the time frame he laid out. Elon's made every forecast he's ever given — it's just usually he's late. But that's fine because I go, when you're trying to change the world, do you really care if it's late? Now, that's the company. As for the stock, I'm a huge believer in margin of safety. And I look at the valuation on an EV-to-sales basis and I go, this is so overvalued that any hiccups, you've got a problem. That's why the stock has broken issue. I like Tesla much better. Obviously I didn't like it last week, but I like Tesla much better to some degree because I feel like you can almost hedge the two because Tesla is trading at, you know, whatever it is today, you know, 14, 12 times sales. You got SpaceX, which if you want to give it credit for all the deals they've signed with, you know, Google and all the other stuff, Cursor, etc., like you can maybe get it down to like 30-something times. And so to some degree, you can almost say, hey, you know, I can be bullish on Tesla, I can be bearish on the stocks — again, I'm talking stocks — because of the relative valuation differences. And I also 99% believe that SpaceX will at some point buy out Tesla. And so I feel like you can hopefully put those two against each other and do okay over time because they both have the Elon Musk valuation premium. And so you can kind of arb those two against each other and go, you know what, do I think robotics is closer, full self-driving is closer today than colonies on Mars? Yeah, I kind of do. So like I'm more willing to underwrite that versus, you know, space data centers and Starlink. I mean, Starlink's amazing, right? Like, so it's a great business. I think, you know, it's going to grow fantastically. I do believe in data centers in space. You know, they're going to cost a lot less because you don't have to deal with the not-in-my-backyard, you don't have to deal with the cooling issues, etc. And so I do think you're going to end up with that occurring. But again, for me it goes back to valuation. I go, like here's the best example, right? In 1999, Amazon had revenues of 1.6 billion. In 2001, Amazon had revenues of about, you know, 3.1 billion or so. So revenues doubled from '99 to 2001. The stock from its peak in 1999 to its bottom in 2001 went down 95% while the revenues doubled. Valuations matter. They may not matter in the short term, but they do matter when you end up in an eventual bust. And by the way, this AI thing is going to bust for one very simple reason. Every great transformative technology in mankind, whether you're talking canals, railroads, you know, electricity, TV, radio, has ended up — and obviously the internet — in a bust because everybody knows if you're the last guy standing, you're going to make a fortune. So you've got 20 guys that all think they're going to get 10% market share. And so ultimately it's going to end up in a shakeout. And we all know that AI is the most transformative technology most of our generations, arguably — with, you know, I put the internet in there, but I think so — you're going to end up with overinvestment. It's going to blow up. It's a question of when, right? Because either you think people are stupid and they don't recognize AI is great, or you think people are smart and there's an overinvestment because you want to be the last guy. You want to be Apple and not BlackBerry, right? Or Nokia. So you're going to go through this. And so that's why I think valuations matter ultimately. And so you have to be careful with that.

I think, but you know what? To me, I don't care because you don't know what is a speed bump versus the ultimate top. But, you know, even in a speed bump, right, semiconductors in 1995 started a drawdown of 50%. Semiconductors in '97 started a drawdown of 50%. The drawdown that started in 2000 was the real deal and they went down over 80%. But you don't know. So your thing should be to protect yourself if you think there's one coming for a very good reason. And which is why on June 20th I said token minimization I think is going to be an issue because you've now switched from token maximization in March to minimization in June. That's why I started to get cautious. Now, I don't think it is the top, but it doesn't matter. In either situation, I start selling stuff and start to get a lot more defensive. And I go, if I happen to be wrong, okay, well, then I'll get more data as I go through this. Clearly, last week with Google and the way Tesla and, you know, Intel reacted made me go, well, am I sure this is a speed bump? But, you know, like it doesn't really matter because you're getting data every day and you get to adjust to it. And so you work it out as you go.

Oh, that one's easy. I mean, what does everybody want to do besides being happy? You want to keep your job, right? You want to keep your job. I want to keep mine. You think the politicians want to lose their job? No. You know, oil prices are going to go down between now and the midterms in November. You know that for the simple reason of human beings want to keep their jobs. So oil prices will come down before then.

Yeah. Now the path is going to be like making a sausage. You don't want to see how it's made. Right. This has been ugly. And so, you know, I don't know what'll happen this weekend, right? My guess is there's a bunch of bombings and before 9:30 on Monday morning, we hear of another peace deal. That's my guess. But like I do know that they're not pushing back the midterms. The Republicans don't want to lose all those seats and they will, based on history, get absolutely destroyed if oil prices are up here and they know that. So that's how that has to work. There is no other solution. Now, it doesn't mean that oil doesn't go a lot higher before it comes down because, you know, President Trump's getting fed up with this. And so, you know, that happens. I also wonder if the Saudis enter the war because now they're allowed to enrich uranium, which hasn't happened before. So I do wonder if there's some backroom deals going on there. It's going to be messy as heck, but I, you know, that's something I feel, and I've been writing that forever, like the midterms are coming. You know, there's a deadline to this high oil prices.

That's a great question because people should really be concerned because rates are the highest they've been since 2007. Now, even when oil had dropped down, right, it had gotten back into the low 70s, interest rates were still near where they were in 2007. Now I believe that debt matters, right? Lehman Brothers, working there, you kind of, you know, you figure that out. The fact that debt-to-GDP across every major country on the planet has gone vertical and we're running deficits of 6 to 7% that you normally run during a recession, not when earnings growth for the S&P of 25% plus this year. So both sides of the political aisle everywhere have lost any kind of semblance of fiscal responsibility. And you know, the housing market was just fine right up until the point it wasn't. And so debt ultimately matters, but when that tipping point is, is hard to tell. But to your insightful question, I do sit there and I go, is the bond market — when oil was much lower, the fact that yields were so high — is the marginal buyer going, 'I'm not willing to lock up my money for 10 years at this rate?' So I kind of think that's going to be one of the triggers to the big problem that we're going to run into at some point. But you know, there's a reason they called it the widowmaker trade in Japan because people are like, well, they're going to blow up and, you know, people got carried out on stretchers betting on that too early.

Oh, yeah. There's always tons of them. And you know, the calls at the beginning of the year, like I have these top five picks, and those are ones that, you know, you hope by the end of the year you know that you're going to have two that are terrible, you hope you have three that are good. Because I always say that you have, you know, the best money managers maybe get 60% of their calls right. The best — the key is on the 40% that you screw up, you try not to lose a ton because math is really easy, right? You're down 50%. Being up 50 on that position doesn't get you back to even. You got to be up 100, right? So for me, I try to cut my losses quickly. And so the way, you know, I manage things is I try to be...

Very, you know, I have this phrase that I heard and, you know, I love it, which is like I try to have strong convictions but loosely held [laughter] like that. And so you want to be adaptable because you don't know if you own RIM or you own Apple, right? You don't know if you own AOL or you own Google. And so you want to be able to be flexible in your thinking as you go through. And if you lose all your money, you're down 100%. You can be up a bazillion percent and you still have no money.

There isn't one. Because, you know, again, I hate losing money. I hate losing money. And I also know I feel a lot dumber now at 59 than I did at 29. At 29, I thought I knew everything, right? I was one of the youngest managing directors on my path ever at my company where I worked. I had, you know, nailed a lot of my calls. And then I think as you get older, you figure out how many things you don't know. I didn't know what a credit default swap was, right? I figured it out, right? And these things out of left field come all the time and you go, I never heard of Archegos when that blew up. Like, what's an Archegos? Like, you know, so I think as you get older, you realize there's all of these things that you didn't even know mattered that all of a sudden matter. And so you feel a lot dumber because I think it's a Mark Twain quote or something. It's, you know, it's not what you think you know for sure, but it's what you think you know for sure but isn't. So it's something like that. I'm butchering this thing. But like, yeah, there's all of this other stuff. Like, you know a lot more things, but then I think you realize, wow, the amount of stuff I must not know that I don't even know I don't know must be really big.

Yeah. Well, it's a great example, right? Like this was a great trade of mine because I lost money on it. And it just reminds you of this. So GameStop had gone up to like 450. Then it came back down to like a couple of hundred bucks. And I was like, 'All right, I'm going to short some of this because I think it's going to drop below a hundred.' And I'm looking at the borrowing costs. Okay, they're okay. So I didn't think the squeeze risk was that high. Checking with my trader every day. Now, I only put a tiny bit. I think it was like maybe a half a percent of the portfolio, you know, into this. And I'm going, okay, I think I've, you know, now you go like, really? You put in a half a percent like the most you can make if it goes bankrupt is 50 basis points. So right there should tell you that wasn't very smart, right? But then it went back from like 200. It ripped again, or it might have been 100, and it ripped again. And I'm going, how the heck is this ripping? Like the squeeze potential supposedly gone. The borrowing costs aren't that high. Like it shouldn't be doing this. And so, you know, that reminds you that the market can stay irrational longer than you can stay solvent. It's not even GameStop. Your viewers may not remember this, but during the great global financial crisis, Volkswagen stock quadrupled in two days because there was a squeeze with their relationship with Porsche. It became the most valuable company in the world. Not GameStop. GameStop never became the most valuable company in the world, but Volkswagen did. Like, so if a company can become the most valuable company on the planet in the middle of the global financial crisis, that tells you anything is possible. And so you need to be humble enough to accept that crazy things can happen that you never even imagined. Again, going back to the I feel a lot dumber now than I did back then because you asked me at 29, can that happen? Can a company become the most valuable company on the planet off of a short squeeze? Of course not.

Well, you definitely spend a lot more time looking at things like how many days to cover, what's the percentage of the float that's short. And so that goes into your thinking whenever you take on a short position because I mean shorting stocks is, you know, ridiculously dangerous, right? Because the math is the most you can ever make is 100%. And you can lose infinite amounts of money like with Volkswagen or GameStop. So to short a stock, that's why you don't hear me typically ever talk about shorting stocks when I'm doing interviews unless it's the mega cap names. And I don't even, you know, I won't even say I'm necessarily shorting them. I'll just say, you know, I'm negative on it or whatever because, you know, unless you're monitoring this stuff literally daily, you know, it's super dangerous to do it. And I'm an optimist by nature, right? So if you believe people are going to have babies and there's going to be some new productivity invention and there's going to be some inflation, each of those things will add 2 to 3% to growth every year, which means the stock markets by nature have to go up by high single digits every year over hundreds of years, right? And that by the way, that's what the stock market has done. So being short a stock, you're fighting that big push from those three factors. And so it's really dangerous. But when you get them right, if you do like in 2022 on the other side of that bubble, that meme stock craze, etc., when the Fed had to raise rates, then you know, you can protect yourself pretty well because even if your longs go down, your shorts go down more and you're also just kind of trying to balance things, you can do okay.

Best one was Facebook. Okay, back when it IPO'd, I literally shorted that thing the moment it came public. It was written up in, I don't know, USA Today or something because I was like, there's this new thing called smartphones. They don't make any money if somebody accesses Facebook through smartphones. They didn't have a mobile advertising thing. Now, of course, the common thing was, 'Oh, the social media thing, it's not real. It's a fad.' You might remember this. Like, it'll never last. Like, this is ridiculous. And I was like, 'That's not what's going on. Everybody's using this. It's just they're not making any money off of it because they have an app that advertises on if you access it through your desktop but they don't have anything for mobile.' Then when they actually introduced it, I went ahead and got long the stock and you know obviously we know what happened since then but like yeah that to me was...

Yeah, and how bad that was.

Yeah. Well, that happened early on, right?