2 pages tagged

#reinforcement-learning

react concept Reasoning + Acting; the bridge that lets language priors generalize across RL environments. 2026-04-27-the-second-half-of-ai source Shunyu Yao's thesis: the recipe generalizes; evaluation is the bottleneck — the utility problem.
esc