Tao: "Math 1.0" vs "Math 2.0"; solutions are being "harvested at large scale in an unsustainable fashion"

Four posts on Mathstodon, 18:00 UTC Oct 6, before the 22:19 UTC release and about half an hour after WIRED (17:28 UTC) reported that OpenAI planned it (https://mathstodon.xyz/@tao/117395267721642920, …/117395268130901862, …/117395268889583356, …/117395269325940185). Tao contrasts "Math 1.0", where a breakthrough sets off talks, workshops and new collaborations, with what is happening now: problems "solved autonomously by AI prompters who have no interest in the broader field", fewer seminars, and open directions "withheld from the public in fear that this will cause their own research to be 'scooped'". Because a solved problem cannot be "reverted", solutions "are now being harvested at large scale in an unsustainable fashion, leaving entire fields of mathematics much less fertile". "Math 2.0" should "decenter the role of raw problem solving" and value exposition, community building and new directions. He believes AI "can contribute positively in all of these directions". Earlier that day (13:51 UTC) he shared a parody of the rumoured release (https://mathstodon.xyz/@tao/117394287977270632). As of 05:20 UTC Oct 7 he had posted nothing after the release.

(pending: to be filled by npm run fetch-posts; views quoted above are from fxtwitter on 2026-10-07 around 05:20 UTC)