Everyone’s talking about what AI can do. I’m more interested in what it can’t.
After building AI-integrated operating systems for multiple businesses, I’ve developed a clear mental model: AI is extraordinary at the repeatable and terrible at the ambiguous.
The Repeatable
Status reports. Session logs. Calendar reviews. Goal tracking. Data aggregation. Pattern matching across documents. These are things AI does better than humans — not because it’s smarter, but because it doesn’t get tired, doesn’t forget, and doesn’t have a bad Monday.
I have 20+ AI agents running across my businesses. They handle session logging, status reporting, calendar audits, goal reviews, and cross-project coordination. They’re tireless and consistent.
The Ambiguous
But here’s what none of them can do: decide whether to pivot the product. Read the room in a difficult conversation with a co-founder. Know that the reason Sarah’s been quiet in standups isn’t disengagement — it’s that she disagrees with the direction but doesn’t feel safe saying so.
Ambiguity requires judgment. Judgment requires context. Context requires presence. And presence requires a human in the room.
The future isn’t AI replacing operators. It’s operators with AI becoming unstoppable.
The Operator’s Edge
The operators who thrive in the next decade won’t be the ones who resist AI or the ones who delegate everything to it. They’ll be the ones who understand the boundary — who know exactly what to systematize and what to keep human.
That boundary is where I work.