The Story Is Not the Wall. What high agency is, and why AI makes it matter more

George Mack’s essay at highagency.com calls it possibly the most important idea of the century. He describes it as three skills at once: clear thinking, a bias to action, and disagreeability. Remove any one and it stops working. He credits Eric Weinstein with showing him the idea.

The test he likes: if you were stuck in a prison abroad and could call one person to get you out, who would you call? That person is high agency.

Psychology has circled this for decades. Julian Rotter called it locus of control in 1966: do outcomes come from you or from outside? Albert Bandura called it self-efficacy in 1977: do you believe you can do the thing?

The Stoics got there first. Epictetus split the world into what is up to us and what is not, and spent his effort only on the first.

Dick Fosbury was a middling high jumper who went over the bar backwards. People laughed. He won gold in 1968, and now almost everyone jumps his way.

High agency is often this. Not trying harder inside the rules, but noticing the rules were only habit.

We now call software that takes actions an agent. Most of it has tools and no agency. It does what it is told, inside the frame it is given, and stops at the first wall. The hard part of building useful agents is the same as the hard part of being one: deciding the wall is negotiable, then checking whether it really is.

I suspect the people who get the most out of AI will be the high-agency ones. Not because the tools are hard. Because the tools remove excuses. When anyone can build anything, the scarce input is the decision to begin.