Post by @tomkalei@machteburch.social

2026-08-20 07:25 UTC

@kangmeister@mathstodon.xyz @tao@mathstodon.xyz I think this can and will continue without the companies. In principle and resource-wise a CERN-like organization could train and operate dedicated open source models for math research.

Replies (2)

@kangmeister@mathstodon.xyz 2026-08-20 12:26

@tomkalei@machteburch.social @tao@mathstodon.xyz That's an interesting possibility, but would a scientifically-oriented (and not purely profit-oriented) organisation really set up such systems to produce theorems without mathematical "stewardship"? What would their reason (read: profit) be for that? And even if there were legitimate non-mathematical but scientific motivations, wouldn't the relevant scientists still want to keep mathematicians in the loop at least to the extent that those scientific motivations could be better addressed?

Open ##5132036
@anatolyvitold@mathstodon.xyz 2026-08-20 12:29

@tomkalei@machteburch.social @kangmeister@mathstodon.xyz @tao@mathstodon.xyz You read my mind 😃: https://proofsandprompts.com/2026/08/20/response-to-the-ai-dissenter-viewpoint/

Open ##5132039