The failure he's worried about isn't the agent that ignores you. It's the one that builds exactly what you asked for, and what you asked for was wrong. That risk doesn't sit with engineers. It sits with whoever decided what to ask for.
His other point lands closer to home for most organisations. The work that goes first is coordination: the status decks, the translating between teams, the meeting that exists so two groups stay in step. That work was never the value. It was the price of being big. We've written about that before.
So here's a test you can run today. Take the last thing you asked someone, or something, to build. Could a person who wasn't in the room tell from what you wrote what it's for, who it helps, and how you'd know it worked? Settling that before anything gets built is the Product Bench, and we publish the answer with the work.
He's clear that this is a skill, not a personality. It can be taught.
