Yes IMO the talk of safety and alignment has nothing at all to do with what is ethical for a computer program to produce as its output, and everything to do with what service a corporation is willing to provide. Anthropic doesn’t want the smoke from providing DoD with a model aligned to DoD reasoning.