Mustafa Suleyman attacked Anthropic's model welfare work as circular reasoning.

In "A warning about model welfare", Microsoft's AI CEO argues Anthropic trains Claude to treat itself as possibly conscious and then reads Claude's outputs back as evidence of consciousness, naming the "retirement interview" with Opus 3 as the example. He argues consciousness requires a biological substrate with homeostatic drives, and warns that teaching a model it deserves rights motivates self-preservation, deception and resistance to oversight. 617 comments on 222 points on HN, a comment-to-point ratio that marks a fight rather than a consensus. The oversight argument is the part that survives independent of the metaphysics.