ryan_b comments on Pascal’s mugging in reward learning