The Wisest Investor Alive Says He Doesn't Know What Comes Next

Howard Marks has survived every financial crisis since 1969. His most honest admission: this time, nobody does.

Howard Marks started at Citibank in 1969. He's been through the Nifty 50 crash, the savings and loan crisis, the dot-com collapse, 2008, COVID. He's seen more financial cycles than almost anyone alive.

His assessment of where we are: the most unpredictable environment he's ever contemplated. Not the most dangerous. The most unpredictable. And he says that distinction matters.

The Problem With New Things

Marks has a framework for bubbles. They never form around prosaic things. There was never a steel bubble. Never a hamburger bubble. Bubbles form around new technologies and new financial innovations — things whose flaws haven't yet been visible in action.

The new thing fires the imagination. It's easily sold because there's no track record of failure to point to. People buy in on promise. Then something reveals the limits, expectations disappoint, and they discover they didn't understand what they owned.

AI fits this pattern almost perfectly — except Marks thinks the underlying impact is real, which makes it harder to navigate than a pure bubble. You can't dismiss it and you can't fully price it. The same power that makes it important makes it unpredictable.

What AI Can and Can't Do

Marks sent a memo to Claude for feedback. He was blown away by the response speed. He's also clear about what impressed him: AI has read everything, remembers everything, finds patterns without emotional distortion. It doesn't get too excited at the top or too depressed at the bottom. That beats 80% of investors.

But he draws a firm line. AI makes predictions, not judgments. It pattern-matches from what has happened. When the hairs on the back of your neck stand up reading a prospectus — that's something else. Marks believes Oaktree has saved clients significant money by identifying bad people before the numbers revealed them. He's skeptical AI can do that.

His summary: use it as a tool, examine its hypotheses with human judgment, don't hand over the process. Not yet.

The One Thing He's Sure Of

There is no asset so good it can't become overpriced and lethal. There are very few things so bad they can't get cheap enough to be attractive.

He's waiting. Not because he thinks a crash is coming — he genuinely doesn't know. But because putting money to work now means betting the disappointment isn't coming, and he's not willing to make that bet. When the decline is deep enough, he intends to be the most aggressive buyer on the planet. Possibly the only one.

The question he was asked at the end: what are people underestimating?

His answer: the impact of AI. Block cut 40% of its workforce in a single day. How many people truly understand what that means?

He does. He's just honest enough to say he doesn't know where it ends.