Unusual at the time. It surprised me, but I was reading LessWrong for 2 years before I read this, and have internalized the notion that probability theory is basically universally applicable.

But back then when Frequentism dominated, I believe we tended to limit probability theory to reproducible experiments.