원문 보기
said Jacob is correct here. We really do honestly believe AI could kill all humans. I personally think it is a more than 10% chance within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence. And we are not clearly on track to do so.
Yes, that's the threat. Yes, of course. Do you agree? Yes, I I think it's very hard to do the the probability calculation. Um but the pace of change here is faster than any of us have ever seen. You have a global race for superintelligence, right? The problem is we can't stop because then the Chinese or an enemy of the country will beat us.
We have to win. And you know, you can be very well-meaning, but we don't know what it means to create something that superintelligence, what the implications are. And these these models operate in such a manner that, you know, they need more resources. They go find a way to solve solve that problem.
So, once you have agents that can access bank accounts, and they can get funding, and you know, they they become independent [snorts] entities, and we can lose control. I mean, it's I think it's a a very reasonable and frightening scenario. And then combine AI with robotics, the Terminator prediction is becoming more and more credible.
So, that's the frightening part. Can you give me a top-line view of your credentials as it relates to your investing over the last couple of decades? We've generated about twice the compound return of the S&P over the last 22 years. We recently launched an entity called Pershing Square USA. It's a publicly traded company that's listed on the New York Stock Exchange.
We raised $5 billion of capital. It's a bit of the democratization of hedge funds cuz it's the first chance a US investor retail investor can invest in a in a hedge fund. What is the the most pressing issue of our our as it relates to both investing business, but also just being a human here on the planet in your view.
The opportunity and the threat from AI. Let's start with the threat. Um I I start there because we were just chatting before we started recording about a tweet that has gained a ton of steam since yesterday. It is from I'll throw it up on the screen for anyone to see, but it's from this guy called Jacob, who said, "I resigned from Anthropic today.
I spent the last 3 years doing pre-training research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts to follow." The other side, of course, is that quality of intelligence is going up massively and the cost is coming down massively, which means that every human's going to have access to something approaching superintelligence over time.
So, we have a a bunch of superintelligent people operating in the world. I think the vast majority of the world are pro-humanity. You would think that's very bullish for humanity. When you think about this first principles, is there any logical outcome in your view where it doesn't end badly? Like this I can't cuz I That's kind of something I've tussled with.
I sat with Geoffrey Hinton, who's considered one of the godfathers of AI, and he says like could you imagine a more intelligent being than us being like subservient to us? And I just I've not been able to find a way that it doesn't end not so well. I mean, hopefully the the the agents are our partners as opposed to who who we report to.
Yes. I mean, look, there are plenty of great examples in history of of benign leadership. That's why we still exist. I Of course, there's evil uh leadership that the People seek power. That's a kind of a natural human uh uh target or destiny. The AI seems to do the same. But lots of people pursue power for good.
And it seems I mean, we operate in a world in which more people seek power for good than for evil. So, yeah, I don't know exactly how the AI will operate. But it's that's that's the concern and it and it it is frightening and it's it's outside of our collective control. Well, it's outside of our control, but you're a very influential person and you know you know lots of these people that are building these companies.
I know some of them. Cuz you know, I've I've been privy to a few private conversations where people are hypothesizing why these AI leaders are marching forward when there's even a 1% chance that it could end very very badly. They have more confidence in their uh ability or their their moral character than than other people do.
Would you think that there is regulation required to mitigate the threat we're describing? I think regulation will inevitably slow things down. And does slowing things down create a greater risk than the potential protection afforded by regulation? And I I I don't have a huge amount of confidence in regulation, right?
You you've seen many many examples whether you can food industry, the drug industry, the securities industry where you know, regulation is certainly not efficient and flawed and leads to that outcome. So I don't know that that makes it better. I do think as a matter of it'd be great before, you know, Anthropic releases some super powerful model, it runs it by someone.
You know, that would Right. And And but you saw a little bit of this where Anthropic did an early release to a few people and then I think Andy Jassy kind of raised alarms with the administration and they kind of slowed down the release of a or of model. That's the first domino, cascading effect downwards to society more generally.
And as an investor, you must be thinking about how society looks on the other side of robotics and super abundance super intelligence and that must be feeding into your investment thesis. So I guess the question is, how do you think broadly this super intelligence and robotics, you know, if Elon I was listening to me the other day say that he thinks there's going to be a billion humanoid robots in fair relatively short order.
And then I think he said within the decade they'll be 10 billion humanoid robots. I sat with Dara from Uber here and Dara says to me that the 9 million sort of gig worker jobs that they have um will will go. My car drives itself in Los Angeles. I saw the cyber cabs roll out in Austin last week. When you're thinking about how society will look on the other side of robotics and superintelligence, do you have a thesis there?
What's interesting that we haven't seen from AI where I think there's opportunity for humanity or humans is creativity. What it What is creativity is coming up with something that hasn't appeared before. You know, all of the models are trained on the past. Right? You know, AI can know more about everything that's already happened.
But it's yet to prove creativity. So, I think that's where human beings are uh have have an edge. And if every you know, you combine the creativity of humanity with access to low-cost or you know, approaching free superintelligence, a lot of really beautiful things can come if if you do you sit in. I think AI is going to create economic abundance.
Uh how exactly that gets shared, I think is really important. Um you know, right now I think AI is creating more jobs that are that are being destroyed. I [snorts] would say AI today that it's most productive task is coding. People are as far as I'm understanding hiring more uh engineers as opposed to less.
And so, you would think that the area that AI is most disruptive becomes the area which is one of the large sources of growth. I The competition for engineers is is greater than ever before. So, now there will be massive displacement, right? The easy example is to think about the autonomous vehicle.
So, you know, a important thing is enabling people to transition from whatever they were doing before that gets disrupted and we've been through this kind of many times. By the way, AI, you know, the beauty of access to you know, low-cost high intelligence it it enables everyone to be some form of an entrepreneur.
Um so, maybe that's what we have to teach people. I've heard you say that a lot of the valuations surrounding these AI companies are inflated. Well, you're going to see a lot of companies, you know, with very big valuations go to zero. Very big, many multi-billion dollar valuations. Um but I don't know if Anthropic at 2 trillion is is cheap or expensive.
I don't know enough to know. With that in mind, I assume that there might be some kind of sort of catalyst moment that means investors in these technologies and these private companies get a little bit more bearish. Um all at once. I think that's kind of like from what I've I've observed in the last sort of 15 years is there's sometimes a bit of a catalyst moment where investors realize that their paper gains aren't actually there and then they kind of contract back and try and
Like if the Anthropic IPO went poorly, that would have a catalytic effect. For example. Do you think this is we are in a bit of an AI bubble per se? I don't think so. I I think, you know, a lot of people got excited about the internet as a technology. It took some time for the internet to turn into a technology that generated, you know, a lot of value, a lot of revenues.
It's taking much less time for that to happen with AI. You're seeing companies form 2 years ago, you know, with a billion of uh run rate revenue in 2 years. That's never happened before. And it's because these companies are adding enormous value and therefore you know, customers are signing up. I had a startup founder say to me a couple of months ago that he was concerned because of all the AI investment that there was going to be some kind of market pullback or some kind of collapse in the economic markets.
He said I'm going to go raise as much money as I can and he raised almost a quarter of a billion dollars for his AI company just valued at multiple billions of dollars and that's because he's scared. He's scared that at some point there's going to be a contraction. Do you think he's right to be? There's a lot of excitement around AI.
People are throwing it money at anything with an AI label. Some of these companies will be incredibly disruptive and become enormously valuable and some will go to zero. That's that's that's normally what happens. And do you think this is a good time to be starting a business? Yes. You can be 17 years old.
You have unlimited access to capital. You have unlimited access to, you know, realistic unlimited access to compute. It's still relatively low cost in a period of time in history. You have an an ability to recruit anyone in the world to a company. Again, you need certain things. You need a really good idea.
You need a really good founder. But, I mean this is the enormous competitive advantage in the United States. A A literally a 19-year-old can recruit the world's expert in nuclear power or, you know, pick your field to come work for a company that didn't exist a month ago by giving him this little piece of paper that entitles them to an equity interest in this startup company, and you can aggregate resources, and you can go on a Zoom, you know, for an hour and raise, you know, infinite money.
So, it's the best time in history, for sure. And if you succeeded in your last startup, or you failed, you still have a you know, if you got a really good I mean Adam Neumann's a kind of great example of that, right? Yeah. Adam had one of the biggest blow-ups in in uh startup history. And uh he's on to his next one, raised a few hundred million of capital.
Does it not make it also the most competitive time in history? Sure, it's a very incredibly competitive world, but there's so many I mean AI opens up the aperture for business opportunity more for than ever before. What are the constraints on building a company? Used to be you have to hire massive teams, you have huge resources, everything else.
Here, you can have a one-person company. It They're going to There's There's already I think been a one-person company that's worth a billion or more. That's why I'm very bullish economically. There's going to be the the pace of business formation and value creation is the highest it's ever been. And that will enter to our collective benefit in a very meaningful way, unless the AI kills us.
If your children came to you, Bill, and they said that um what do you think are going to be the most important skills of the next decade? I want to make sure that I'm set up to be successful in whatever their subjective definition of success is. What would you say to them? How to use an AI, work with an AI, um public speaking, how to communication skills.
And if they said that I want to learn how to be um a great investor. I have, you know, I have $10,000. I'm just starting out. Would you just Would you suggest they invest in a particular thing or would No, I'd I'd I'd start reading if I were I were them. I'd watch podcasts and with with lots of different investors and get some perspective.
And then yes, I think 10,000, 1,000, doesn't matter. Uh I would start investing. I'd start reading annual reports. I'd start learning about businesses. And what's your investment approach in individual companies and how has that developed through time? I know you talked a little bit about diversification being important.
What are the other sort of principles of your approach your approach? Actually the opposite. I talk about concentration being important. Diversification is protection against ignorance. If you don't know what you're doing, you want you want to diversify your assets as much as possible cuz it minimizes the risk of a of a catastrophic loss.
I mean, think about uh in terms of wealth creation, right? If you read through the Forbes 400 wealthiest people in the world, most of them made their money in one business. The very I don't I don't think there probably no diversified investors in the in the um Forbes 400 magazine, right? They're people who built like one company.
I mean, Elon's got you know, five or six. So, that's a portfolio of one. Right? So, it's hard to get rich with a diversified portfolio. So, if someone does have a $1,000 now, you your advice to them would be to use the money to educate yourself Yes. through learning by doing. Yes. Absolutely. And And would that advice apply for people that don't want to become an investor?
I think everyone has to learn about investing. The problem in the world today in terms of why do we have wealth inequality is largely driven by the fact that you have a group of people who only make money while they're awake and are working. And they're paid as an Uber driver. And when they turn on the Uber and they get a ride, they get paid.
But when they're getting lunch or they're at home, they're not making any money. And so their their their opportunity for generating income is limited uh by you know how many hours they can work in the day. Whereas if you're an investor, you can make money while you're sleeping, you can make money while you're on vacation because other people are working on your behalf.
And money has compounded at a much faster rate than wages over time. And so you want to get yourself in a position where you can have assets invested in the stock market as early as possible. Then you get the power of compounding over time like you know Trump accounts which was an idea uh guy Alex von Furstenberg I don't know if you know Alex.
He said look we have biggest problem we have is this. We have half the country that when the stock market goes up they get resentful um and uh half the half the country obviously does well in that world. We need to democratize capitalism so everyone owns a piece of you know the success of the country, the success of AI, the success of Anthropic because have an index fund and Anthropic's going to be in it.
If you're not participating in compounding and you're only making money on an on an hourly basis or when you're working, you're going to fall behind. I guess my concern that's Yeah. is one of the number one of the nine. That when the stock market goes down and your portfolio's down 20% you can't panic.
To your question, what you'd be concerned about, one of the mistakes people make is they invest money they can't afford to lose. In order to be a successful investor in my view over the long term, you have to position yourself so that you can absorb the volatility without it affecting your you know daily uh okay.
That's the next Expect volatility, sure. I wanted to just tick off this um this nine-part framework for investing um because I wanted to make sure the audience heard you say it but one one of the ones in your nine-part framework for individual stock selection is to invest in public companies, choose established businesses with liquid stocks that are easy to buy and sell.
Understand how a company makes money, invest at a reasonable price, find a company with limited debt, look for high barriers to entry, Mhm. which is I think harder these days. They're going to force you into short-termism if they can. Yes. That's exactly right. There's something that's called time arbitrage.
Mhm. [clears throat] Right? If you're operating in a world where everyone else is forced to think short term, but you can be long term, you have a huge advantage. And that's true for almost everything. Many people say, "I would do that, but you know, my current economic circumstance, uh other constraints in my life." I mean, think think about it every aspect.
You think about nutrition or health or whatever. Right? There's so many examples of things that if you take the very long-term view, the outcome is vastly better. The short-term gain, whether it's a pleasure type gain, but the the long-term I mean, smoking, right? That's a simple example where obviously people took the short-term pleasure uh and suffered the long-term, you know, consequence.
There there's so many things like that, whether it's investing, health care, relationships, Public companies' quarterly reports I think of as creating a certain short-termism, which isn't conducive. You get the shareholders you deserve uh based on the way you present the company to the market, you know, to take risk.
I heard you talk about the the deal you did investing in Netflix in 2022. Yes. As being an interesting example of of potentially what we're describing here. Yeah, certainly a failure uh in that we bought a stock, we had a very clear thesis for what would transpire. Um and we were surprised very shortly after we bought our stake.
It was kind of an interesting case where we went out to see the management of the company a few weeks before they announced earnings. And we had a great meeting, very encouraged. With who? We met with Reed Hastings and the and the senior team. Who's the founder of Netflix? Yeah. And we met management.
We shared our thesis um and we left very bullish on the company. [snorts] The company announced earnings a few weeks later and management was I would say deer in the headlights. They had told us they would never consider advertising. And we asked them why not? They had a pretty clear answer and then a few weeks later you know, subscriber counts were very disappointing.
They had talked about launching advertising. You know, kind of surprised us. So our view is you invest in a company and then you learn new information that's inconsistent with the original thesis. You got to think very seriously about exiting. And we exited and we took a I don't know, 25% 30% loss on the position and invested in Google which turned out to be a good investment.
But we kept track of Netflix. What's interesting today is we think Netflix is a much better more certain business. And then the market gave us another opportunity. Upon meeting them you knew that they just had it. I've invested with a lot of great entrepreneurs. I was a day one investor with Bom Kim who built a company called Coupang.
My most successful venture investment ever, my first personal venture investment was a 1500X. The first 200,000 I put in the company became 330 million. The second 300,000 became three another 10 million shares. And then I put 2 million and another 385 million dollars. And it's the only time I invested with a founder.
I had a falling out with the founder. Crazy. You had a falling out with him? Yes. Yes, but um I'm pretty good at picking people. And what I saw with him was all the things we talked about and then I lost a little confidence in him at the very end. And crazy that I had an incredibly successful outcome and very disappointing relationship at the end of the day.
That's an opportunity and a responsibility. So these people that are tweeting these kinds of things about anthropic and open air and the dangers That's a very important tweet. Very important. You don't think that being doomers cuz these people Absolutely not. He's This is a totally authentic person who gave up his job because he's concerned and he's sending a very important message to the world.
Do you know it? I don't. Cuz these people are quickly called doomers. The I watched Silicon Valley taking the piss. Oh, another doomer, blah blah blah blah blah. No, no, no. This is a person who authentically is concerned for sure. And and it's we need these people. Because by the way, that the the virality of that tweet is going to cause Anthropic and OpenAI to be that much more careful. So, he's this person is helping humanity.
Comments 0