> Mostly pretty straight forward. The idea that proofs be desloppified, attribute existing literature properly, and published somewhere expediently where it can be commented on, with artifacts for verification, is all very uncontroversial stuff.

If Ramanujan could freely disclose and publish results that were arrived at by literal divine inspiration (he certainly didn't bother to write down "proofs" in the modern sense!) why do we suddenly need all this "responsible release" process stuff when it comes to AI-derived results? This makes no sense when compared with existing norms within the mathematical community.