Dwarkesh Patel: uncertainty

Transcript context

…l theory you say, “Okay, here's what's actually going on in the world. Here's what people don't understand.” but you also have to have this lower level theory of — “Here's what they think is going on. Here's why things keep ratifying the theory that they have.” and then the next step is “Okay, what actually breaks down that causes perception to collide with reality?” And then the other question is, can perception actually undershoot in the same direction? A lot of money was made by people who looked at the tech bubble in the late 90s said “This is gonna blow up and we're gonna figure out when to short it and then we'll short it. We know it'll just keep going down for a while.” It wasn't that the dot-coms were 20% too expensive, it's that most of them are worth zero. Those people made a lot of money by shorting after things started declining knowing they would keep on declining. But a lot more money was made by people in 2003-2004 saying, this didn't actually discredit the internet. It's still a good technology and it's not like we fundamentally can't make money online, it's that you can't make money online if you don't know what you're doing and you massively overspend and there just aren't that many people online and no one's spending money online yet. So a lot more money was made by people who were able to take advantage of the overshooting in the opposite direction rather than the people who figured out from first principles that the bubble had bubble characteristics and was eventually gonna pop. That just requires a lot more of this micro level analysis. A lot of macro people are looking at what individual companies are doing and what they're saying and how consumer sentiment is changing month to month and all these other very low level indicators where the indicators are not a thesis but the indicators tell you something about when your thesis will become true. That's really interesting. I'm just gonna do some rapid fire questions in the final few minutes. First, how can somebody be long AI but hedge for the possibility that Taiwan will be invaded? So, I don't know if I should put money into TSMC but I know that GPUs are going to be the next big thing or are going to be very important in the future. How do you make that position concrete? Oh man, that's really hard. I guess the next best thing would be looking at the Korean fabs because they're close but they don't produce the same chips. I guess the bet would be that Korea is the country most likely to catch up to Taiwan today if Taiwan is no longer an option. But if you're trying to make the AI bet conditional on Taiwan bet, I think a lot of what you want to do is actually think about how you underwrite the Taiwan invasion bet because that's probably the thing with the bigger long term impact. Maybe not. It gets tricky. Sometimes geopolitical changes can just lead to these permanent inflections. There's some kind of emissions that you can measure that are the result of copper mining and so we can see how much copper mining changes year to year throughout history and we do actually see it rising during the Roman Empire and then peaked and then went down and then didn't come back for a millennium. Sometimes there is a really unfortunate geopolitical inflection in underlying technology. But if you think about what that means in the real world, your big concern is not your portfolio in a world where we were an invasion is the reason AI does not happen. I would separate the invasion bet from the AI bet. And then I guess the sad answer is Intel and TSM are sort of America's last hopes on this. You can you can tell a story where invasion becomes more and more likely and the US does a sort of operation paperclip with no connotation about the the political views of the engineers of old but you know, operation paperclips all the best TSM engineers out to Arizona and has them all work on building those chips in the US in which case TSM is still you know, still a play although it's certainly lost some valuable assets. But that's a very tricky question. And it may be one of those things where it's kind of kind of hard to hit like. There are sufficiently bad things you know, there's not really a good meteor hedge in the minutes before the meteor hits us, maybe treasuries do outperform equities, but you don't really care. [Laughter]…

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