Invest Like the Best with Patrick O'Shaughnessy
· Ben Horowitz
[00:00:00] S1: On the positive, everything is up for grabs. I think people are kind of overreacting to that in the stock market and so forth, in that if you look at existing software companies, people think, "Oh, they're all dead." Some of these guys are extremely hard targets. It's not that easy to take out Salesforce or SAP. You would be surprised, even with AI, how much heavy lifting that is. Having said that, it is true that a lot of these things, you can just make your own, you can do it yourself. It's going to be a lot easier. The number of possible interesting companies, I think, went up a lot. The other thing we're seeing is the products work so much better than any technology products we've seen in the past. Revenue growth is so much faster for these AI companies. And there's many such cases of companies coming out. Cursor, which is ostensibly an IDE, like what's the biggest IDE before Cursor? I don't know, but it wasn't big. And it took probably 12 or 15 years to get to that revenue level. And yeah, they went over a billion dollars in revenue in no time. So that's super interesting. I would say though, from an investing standpoint, the laws of physics of company building changed, which is going to affect investing in what's currently, I would say, an unknown way. If you look at the one thing you knew, if you'd ever built a software company, is you cannot throw money at the problem. "What's a man-year? 700 IBMers before lunch." That phenomenon, everything was built on because you knew if somebody built a great product and it took them three years and they did it with a small team, Google's not going to hire 2,000 engineers and catch them. It's just not going to happen. That was a law of physics. Now, if you have the data and you have enough GPUs, you can solve damn near anything. And we've seen that with Elon catching the big models in no time. He just took a lot of money and a really good data center design and some smart engineers. And he got in the game very fast. That would have never happened in the past. The markets also seem to be much, much, much bigger than anything we've ever seen. It would cause you to think about valuations and long-term value and other sorts of things in a different way than we have in the past.
· Ben Horowitz, Patrick O'Shaughnessy
On the one hand, well, when you calculate the long-term value, what if this market wasn't $50 billion? What if it was $5 trillion? On the other hand, what if somebody could catch you? These are concepts we've not dealt with. [00:00:00] S0: So how would the conversations feel different to me if I came in – you've got all these great investors working at Andreessen Horowitz. The nature of the conversation amongst your teammates as they're debating this stuff versus four years ago or something – where does it feel most materially different internally?
· Ben Horowitz
[00:00:00] S1: I would say one of the most different things is when you look at AI researchers, it is really a different kind of thing. If you haven't been at Google or Facebook or OpenAI or Anthropic and somebody gave you hundreds of millions of dollars to try and build a giant model and you weren't one of the main people, then you probably don't know how to do it. Because you can't learn it in school. You can't learn it in school because it's a little bit alchemistic in nature. It's a little bit of an art. And so if you've never done it before, the chance of on your very first try of building some kind of large model, that it's going to work well, isn't that great. People are coming up to speed more. There's more companies. People are learning it. But that's kind of why you got to this, which from the outside world probably looked absolutely bananas, that why is somebody paying $100 million for an AI researcher or a billion dollars for an AI researcher? Like, that's the craziest thing I've ever heard. Well, what if there are only 40 of them in the world? Then it kind of changes the math on it a little bit. And I think that's where we were because it's kind of the first time we've had a need for a technologist that academia can't produce. That is probably one of the bigger things that changed in the conversation is, like, who are all these people? And like, we track all of them and know what they're doing. But it's very different.
· Patrick O'Shaughnessy
[00:00:00] S0: Everyone talks in venture about the power law. The thing underneath the power law is a sort of inequality. It seems like so many of the things that are happening are just massive multipliers on the trend of inequality, in every way. The billion-dollar researcher, the size of the biggest companies, the wealth of the people creating those companies. I would argue that inequality is a feature, not a bug of the American system. But I'm curious for you to riff on the nature of growing inequality and the good and the bad associated with that.
· Ben Horowitz
[00:00:00] S1: What's happening in AI is, I would just say, an extension of the Kobe Bryant effect, which is, you know, a basketball player. There was a limited amount of money you could make because you basically played the game in front of the people who could show up for the game. And that was it. That's the whole market. Whereas once you add television and a global audience and these kinds of things, you can be LeBron James. You can become a billionaire. And that was not at all possible before. And I think that we kind of first saw that with the internet where, "Okay, now I can build a product. I can get to global distribution very fast. Then all of a sudden I can become extremely rich." And then AI is another layer on top of that in that, okay, now take that same product and make it just a more valuable thing. Whoever invents that is whatever the internet company was, plus, plus, plus. And so that's going to make them even richer. That's the bad part of it. But I think the good part of it is it's starting out day one completely democratized. Anybody gets access to very powerful AI. Most people in the world at this point have smartphones. And now you've got superintelligence in your phone. So that's a big equalizer of the opportunity in a lot of ways. I don't think we've ever seen a bigger opportunity equalizer than AI. Every child can have a super advanced, amazing tutor or teacher. Great education is accessible to all. So I think it's an equalizing technology. And there's some drive in inequality there. This is another thing I learned from my father when I was very young was, he said "Look, son, life isn't fair." That's extremely good advice because it's just not going to be fair. No matter what government or anything tries to do, it's not going to be fair. And the problem is if you create a system that tries to correct that, it doesn't make things more fair. It just transfers all the power to the person running the system. And that's what happened with Stalin. That's what happened with Ceaușescu. That's what happened with Pol Pot. That's what happened with Mao. Not an accident that every single system like that went bad because it really ends up just being a power transfer. When you think about, well, what do you want? You'd like everybody to have a chance. Don't give me no chance. Give me some chance.
· Ben Horowitz, Patrick O'Shaughnessy
It may not be as big a chance as the other guy. It may not be a perfect chance. But if I have the desire, if I've got some capability, give me a chance to be something, to make my imprint on the world. A system like that is going to end up with a lot of inequality. By the way, all systems end up with a lot of inequality. But you can try systematically to give everybody an opportunity. I think AI does a really good job of that. [00:00:00] S0: One of the memes that's very popular today is that you have a couple years to get some capital or you're going to be a part of the permanent underclass, is the phrase that’s used on Twitter. And I certainly agree that now everyone has the best lawyer, accountant, advisor in their pocket, and that's amazing. But what do you think about this notion that because of that, we need less labor? It's going to be harder if you don't have some capital to begin with to accumulate capital and break in. I don't necessarily believe that. I'm just curious what you think about the challenges we'll face because of AI societally.