My new paper taking a computational complexity lens to machine learning. arxiv.org/abs/2604.07233 We abstract ML models as non-uniformly efficiently computable distributions and show that learning computable distributions forces the use of probability to manage complexity.

→ View original post on X — @ceobillionaire, 2026-04-09 14:54 UTC