x
Inner Simulator / Surprise-o-meter — LessWrong