Speaker unverified: belief

Transcript context

…and it matters how the government intervenes and due process matters. There's the famous case where many of the digital education apps were banned overnight. Did that really update your view of China? Yeah. Investors hate uncertainty. In the US, can something bad happen to social media companies? They get called in front of Congress, but there's due process. Donald Trump can't wave his hand and get rid of Meta because he doesn't like the fact that they're allowing people to say bad things about him. And so when you have a government that can act capriciously, that has a humongous gravitational pull on valuations. Some big picture questions related to China. So there's all these charts one could look at of the concentration of Mag 7 into the S&P 500 and you could look at the 1970s equivalent of the Nifty Fifty. There's the inflation, which seems to be tracking the “double hump” story. First of all, do you care at all about these analogies? Do these analogies mean anything to you? These charts with all the JPEG artifacts on them? Yeah, they're always very convincing. Look, history rhymes and it’s the same thing in the market. So it's interesting to look at what's happened historically to these companies. It's dangerous to look at it and just extrapolate to the point where you— So you can overcorrect on the chart. But then my question would be, what would be the sign for you actually that the bubble is close to peaking? Is there something, an anecdote or an actual piece of evidence you look at, you'd say like, “Okay, maybe we all believe in AI, but…” So you're saying AI is a bubble? No, I'm saying I believe in cyclicality. The internet worked, but it was also a bubble. Derek Thompson said this recently in a way that I liked, which is people associate using AI—that AI may currently be or in future be a bubble as some kind of negative statement. But because every major tech change—the canals, the railroads, the internet, whatever—has been accompanied by a huge speculative bubble, it has to, when you think about it. Americans are optimistic. But it's an AI-optimistic statement to make that there will be a bubble. We’re actually celebrating how important it is. So Daniel's saying he's an AGI believer and AI is clearly going to be a big thing. Therefore, is it this year? Is it next year? When's the bubble? bble. We’re actually celebrating how important it is. So Daniel's saying he's an AGI believer and AI is clearly going to be a big thing. Therefore, is it this year? Is it next year? When's the bubble? If you look at the checklist of things for what's going to be closer to the end of the bubble. If you went back a year ago, you'd be like, “Okay, yes, there's a lot of money being spent, but these companies have massive amounts of cash flow.” Microsoft kicks off a ton of cash flow. Once you start having debt-fueled investment, that's usually a bad leading indicator because obviously when you have a lot of debt there's not much room to make mistakes. And you are seeing some of these latest projects are debt finance. I think that nobody knows. I think even if you spoke to—I know when Dario was sitting here… It's like they invest, they keep investing in training but at some point the returns on that training are going to be disappointing and then people will be freaked out a little bit. But usually then the next time it's back to, “Okay, the next model's good.” At some point people are going to say, “Well, we're asymptoting out. The returns on this new investment are actually not working.” That's when I think you'll see a pretty big correction. Okay, so there's the theory that there's a correction because the returns to pre-training start asymptoting and then there's a theory of bubbles. If you look at the 1970s, and to some extent dot com, which I think was really hampered by the fact that the Fed raised rates into ‘99 and 2000. There's another view of bubbles which is like there's an awesome thing happening, everyone gets overexcited, but then some random thing hits the bubble and because things have run up so much, everyone starts panic selling all the way down. All the people that bought their cost base have NVIDIA—not to pick on anyone—it's like 4 trillion. And so my question is, do you think we're in that mode and is this a productive… Like at D1, do you guys ever think of, “Oh, what could be that thing and how do we prepare for that thing?” Or do you think— I think, to our detriment, we should have owned more AI stocks. There are definitely things you go down the list of what constitutes a bubble and you can check some things like massive debt, massive debt-fueled investment just like people were building— What else is on that checklist? things you go down the list of what constitutes a bubble and you can check some things like massive debt, massive debt-fueled investment just like people were building— What else is on that checklist? Valuations, and bad companies trading at crazy valuations. I would tell you, NVIDIA is… We can have a debate about what the earnings are going to be in a few years, but it's not expensive. NVIDIA trades at 20-something times multiple. That's I think within reason and I don't think there's anything I see in the public markets which is bubble-like from a valuation standpoint. I think if you look back in history, 70% of the time that you have this major breakthrough technology, there is a stock market bubble. And maybe what you're seeing right now in some of these retail stocks like where Opendoor goes from one to 10. Maybe that is things starting to bubble, but we haven't seen large cap… I don't think the large cap AI stocks are trading at crazy valuations at all, so we're not seeing that yet. So maybe it's 1997, 1996 and by the time that we're in the equivalent of 1999 and NVIDIA will be three times higher and it's possible. Someone said to me recently as well that for a proper crisis you also need things that people thought were safe to not turn out to be safe. In 2000 when the NASDAQ went down 85%, it's like, well that's a real bummer but we did know that we were buying these highly gassed tech stocks, whereas it's when the debt actually turns out to not be safe. The debt turns out to be equity. Exactly, yeah. That's when you get real issues. It's very different from 1999, though. There were horrible companies which had no real economic prospects trading crazy valuations. I'm just not seeing that in the public markets now. Well I am, just not in AI. You're a huge fan of SpaceX and a big holder. Why are you so excited about this? Everyone understands Elon is an amazing inventor and amazing entrepreneur. I think people underestimate how good of a business person he is. It's like okay, yes, does he invent great things? He does, but he is ruthless about bringing down costs to a point where his business becomes a natural monopoly because it is a low-cost provider. With SpaceX, the whole problem with space in general, doing anything in space, historically was that it was very expensive to launch anything into space. Because one, the rocket blew up and therefore it better be really high value if you're going to send it up[ there because it's like it if you're going to take a plane from New York to LA and every time you do it, the 747 blows up the plane ticket's going to be really expensive. You better really want to go to LA, right? So the first thing is the idea of bringing down the cost dramatically by making things reusable. He was five to 10 years ahead of everybody else.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.