This paper is a criticism of full-information theories of welfare. Such theories unsuccessfully attempt to accommodate an internalist intuition (that one's good depends in some way on one's desires or hypothetical desire) with a rationalist intuition (that only fully-informed desires are relevant to one's good).