Gamification and rationality training — LessWrong