I read and commented about it already. This does not proof any originality in the LLMs output, because we do not know a) which parts are human and which are not - if the “creative” part of this solution was done by a human mind, we probably wont get that info and b) 17.4 mio LoC sounds a lot like “chain a fuckton of existing methodology together”. Exactly the stuff they are good at.
I read and commented about it already. This does not proof any originality in the LLMs output, because we do not know a) which parts are human and which are not - if the “creative” part of this solution was done by a human mind, we probably wont get that info and b) 17.4 mio LoC sounds a lot like “chain a fuckton of existing methodology together”. Exactly the stuff they are good at.
Edit: well look at that, we do NOT need 17 Mio LoC or an LLM for that: https://officechai.com/ai/hours-after-alpoges-ai-assisted-riemann-hypothesis-proof-ams-president-ravi-vakil-rushes-out-a-rival-all-human-proof-by-alessandro-masullo/
Best thing about it that humans can read and understand it, if the checks are ok.