x
Goal-directedness via explanations — LessWrong