Stuart_Armstrong comments on Pascal’s mugging in reward learning