Lucas Teixeira comments on Reinforcement Learning Study Group