Kay Kozaronek comments on Reinforcement Learning Study Group