Howard Marks: how I make money while you worry about a market crash

**Howard Marks** (0:00)
If you wait until you have nothing to be afraid about, probably the opportunity has passed.

**Sam Parr** (0:15)
Howard, it's good to see you again. We had a lot of fun last time, and we were like, look, I don't know if other people would like that, but we loved that. And then over a million people listened to the last one. And so this morning, I was reading, you wrote this blog post about how you changed your mind about AI. You had written, I don't know, a couple months back about the possibility of an AI bubble, and then as a good thinker tends to do, you got new facts, you sort of reassessed the situation, you wrote a new post about AI. Do you want to summarize the story of how you changed your mind on AI?

**Howard Marks** (0:49)
Well, the story is very simple. I have this son named Andrew. He's a VC.
He's dealing with AI every day. His companies use AI, some of them create AI, et cetera.
I had written the first memo around December 9th, as I recall, and then in early February, he said, Dad, so much has happened. You have to update the memo. And so, I rewrote the memo entirely.

**Shaan Puri** (1:15)
I was rereading one of your old books, and you repeat this phrase a bunch, which is like, it's important to be rational, and you can't get seduced into thinking something is a good idea, because that's when smart people can make bad decisions, and you get emotional about something. But then when I was reading part two, I was reading it, and I was like, Howard, you sound a little seduced. You sound a little seduced. You sound like you're into this.
Are you at all approaching this in an emotional way, you think?

**Howard Marks** (1:41)
It depends on your definition of emotional. I upgraded my opinion of AI and its potential, because its ability to talk about its own strengths and weaknesses, to use humor, to put information in the context of me, to use what it knows about me. And you know, this is really exceptional stuff. There's a quality to AI, or more than one quality, which are unprecedented, in my opinion. The first, the obvious one, is autonomy.
All the other technological innovations, from the railroad, to computers, to the Internet, et cetera, were all tools, or things to speed up and increase productivity. There's never been anything with the quality of autonomy.
The idea that you can give it a job and not tell it how to do it and it will figure it out is really unique. And what comes with that, of course, is this nagging concern that it may take over. So that's really important. The other thing, and this is not kind of quantifiable, is there's never been anything in my opinion so unpredictable. I don't think anybody knows the shape of the future. So I have never had that sense before. I never said, I never thought that the Internet, for example, was beyond comprehension or beyond prediction.

**Sam Parr** (3:36)
Do you think that AI will be able to do what you do? And I know you talk about this in the memo.
And I got to be honest, when I read it, I almost felt like you read stories about Warren Buffett reading the Moody's Manual page by page, 800 companies and trying to digest that information. Well, AI can do that in a heartbeat, right? Like a lot of the things that go into making investment decisions, they can do very well, very fast. And then also, it's advancing so fast. So whatever we thought it could do three years ago is laughable compared to what it can do today, as you pointed out, even three months ago. So I guess in your heart of hearts, do you think in the future, the next Howard Marks is not a human, but maybe a human with AI or just AI?

**Howard Marks** (4:25)
Everything I say on the subject I preface with, I'm no expert, but I think I told the story in the memo about the fact that indexation put a lot of people out of the equity business because it disclosed that they couldn't do what they claimed to do. And most active equity investors underperformed the averages. And AI will unfrock or defrock another group of people whose talents are not as great as they purport, I used to say, about computers. When I went to school and learned about computers, all they could do was read, remember, add, subtract and compare. They could do it with a lot of data. They could do it really fast. They could do it without making arithmetic mistakes. They could do it without making emotional mistakes. So while the list was limited, it was still better than most people. Now, what's the list for AI? Is the list for AI unlimited or limited? That's a big part of the question right there.