On Kelly and altruism — LessWrong