x
Can AI learn human societal norms from social feedback (without recapitulating all the ways this has failed in human history?) — LessWrong