Behavior: The Control of Perception — LessWrong