Steven Byrnes comments on Reward is not the optimization target