This piece truely made me think about the Ψ-formalism and how perfectly it encapsulates the problem. Is there a danger that even if we acknowledge the inversion error, the sheer utility of these projection systems will keep us from building truly generative intelligence for too long?