benelliott comments on [Link] A review of proposals toward safe AI