Ontologial Reductionism and Invisible Dragons — LessWrong