What actually explains the difference between people who get real value out of AI and people who bounce off it after a few tries? It isn't technical skill, and it isn't job title — I've watched both predict almost nothing.
The pattern I keep seeing instead is closer to a willingness to ask an obviously dumb-sounding question out loud, and then actually follow the answer somewhere unplanned instead of closing the tab once the immediate task is done.
That's most of what Battle of the Bots and Project Sagebrush actually were, underneath the specific mechanics. Dumb-sounding questions, followed further than seemed reasonable at the time. What happens if two AI agents with opposite strategies trade against each other with real money? What happens if I let an AI system propose a restoration schedule instead of building one myself? Neither question had an obvious payoff when I first asked it.
I don't think curiosity is a soft trait here. I think it's closer to an operating system — it determines whether the first unhelpful answer from a model ends the conversation or starts a better one, and whether an experiment that doesn't work on the first try gets rewritten or abandoned.
The people I've watched get the most out of AI aren't the ones with the most technical background. They're the ones who keep asking past the point where a reasonable person would have stopped.
Current Hypothesis
Curiosity, sustained past the first uncomfortable answer, predicts AI-adoption outcomes better than technical background does.
Not a conclusion — a working idea, revisited as evidence comes in.