Jared Peterson rather convincingly argues that decision-making is not a skill. Expertise, especially adaptive expertise, is an “uncommon common sense”. Enough exposure to a domain and things “just click” for you when they don’t for others.

It’s only recently struck me how subversive this is. Think of paradigms in the Donella Meadows sense, or of Kuhnian revolutions. We’re in the midst of one. Venkatesh Rao describes the arc in three acts:

• Dark Forest. AI makes skilled adversarial attention cheap, so all software gets probed, all the time.
• Hardness Wars. The fight moves down to the maths that cryptography rests on.
• Strange Forest. Computation branches into forms we can’t parse, but which feel like home to whoever lives there.

If decision-making isn’t a skill, change like this disrupts. Judgment is exposure to a specific environment, so it doesn’t carry over when the environment changes. Entire swathes of accumulated expertise get invalidated at once. Nobody has the reps, not even incumbents, who are seeing these things for the first time, too. So whoever acts first accumulates them fastest.

Skills still matter. A move you’ve drilled is a move you have. But the moment of its application is the thing: knowing when the scenario is apt for it. That part comes only from exposure, and one can load up on it ahead of the change. Brazilian jiu-jitsu works this way: every sparring round is a tactical decision game, as long as there’s feedback at the end.

Exposure also compounds, which is why Musk and other pseudo-great men look n of 1. They’re mostly mediocre people with exceptional exposure over time, hailed as de facto great men.

The argument is not that you can be similarly great just by recognising great change. Only that radical change is a rare chance to accumulate expertise ahead of the competition, in a way that leads to what 7 Powers calls persistent differential returns.