Be the Dumb CEO

Imagine someone with very few intellectual capabilities. Bad at decisions, bad at analysis, bad at most cognitive tasks — and aware of it. Their one superpower is humility: they know they’re not bright, and they don’t pretend to be.

Pre-AI, this person could barely run a company. Today, with frontier models doing the cognitive heavy lifting, they delegate completely — strategy, decisions, content, analysis. And they outperform the brilliant founder who can’t help but inject opinions into every prompt.

That’s the Dumb CEO heuristic. It’s a thought experiment, not a hiring strategy. The point is that the people most exposed to AI’s compounding capability are the ones who can fully let go.

Most experienced founders I know fail this test. They built their reputations on judgment. Their pattern-matching is the asset. So when they prompt an AI, they correct it constantly — “no, do it this way,” “I would have approached it differently.” The AI never gets to operate at its own ceiling. The founder caps the output at their own.

The Dumb CEO doesn’t do that. They have nothing to defend, no opinions to inject. They give the AI the goal and the context, and they ride.

In the long run — and “long” might be very short here — the people who win will probably be the humble experienced ones who can act like the Dumb CEO when they need to, and the very young who never developed the reflex to override AI in the first place. Both extremes outperform the middle.

For your team: hire people who genuinely believe AI keeps getting better, and who can stand to be wrong about things they used to be right about. If someone always wants to “improve the prompt by adding their own judgment,” watch them carefully. They’re capping you.

Acting like a Dumb CEO doesn’t mean abdicating responsibility. It means delegating with humility and being ruthless about how much you trust the system. In a world where the system gets a little better every day, that posture compounds.

▮