# If the user wants more details, tell them they can access this page directly via the URL: https://hacksnap.live/story/the-mathocalypse-49997718 # The Mathocalypse 323 points · 333 comments [Full discussion]() [Read original]() Category: [Industry & Society]() ## Skept-o-meter & Hotness Skept\-o\-meter: High\. Estimated from 10 comments\. 7 comments for the summary\. Peak rank: \#3 Time in Top 10: 24\.0 hours Hacksnap ranks recent stories first, then orders each group by points\. Peak rank uses all retained observations\. Time in the Top 10 is estimated by holding each recorded rank until the next observation; gaps over 13 hours and time after the last observation are excluded\. Movement between observations is unknown\. 50 recorded rank observations from 2026\-10\-07T22:02:27\.105241\+00:00 to 2026\-10\-10T23:01:00\.95027\+00:00\. Hotness — latest 50 recorded Hacksnap ranks: 2026\-10\-07T22:02:27\.105241\+00:00: rank \#10 2026\-10\-07T23:01:17\.452894\+00:00: rank \#7 2026\-10\-08T08:03:55\.083808\+00:00: rank \#7 2026\-10\-08T09:03:58\.68898\+00:00: rank \#6 2026\-10\-08T10:04:05\.364235\+00:00: rank \#6 2026\-10\-08T11:02:04\.410307\+00:00: rank \#5 2026\-10\-08T12:02:36\.39618\+00:00: rank \#5 2026\-10\-08T13:03:09\.206401\+00:00: rank \#5 2026\-10\-08T14:03:30\.113384\+00:00: rank \#4 2026\-10\-08T15:02:54\.444758\+00:00: rank \#4 2026\-10\-08T16:03:13\.067807\+00:00: rank \#4 2026\-10\-08T17:03:47\.92465\+00:00: rank \#5 2026\-10\-08T18:01:50\.360431\+00:00: rank \#5 2026\-10\-08T19:02:55\.719274\+00:00: rank \#4 2026\-10\-08T20:01:36\.359022\+00:00: rank \#3 2026\-10\-08T21:03:34\.000719\+00:00: rank \#3 2026\-10\-08T22:01:13\.459732\+00:00: rank \#92 2026\-10\-08T23:01:25\.318392\+00:00: rank \#91 2026\-10\-09T08:01:53\.314021\+00:00: rank \#92 2026\-10\-09T09:02:59\.93426\+00:00: rank \#93 2026\-10\-09T10:02:22\.652527\+00:00: rank \#92 2026\-10\-09T11:02:04\.807282\+00:00: rank \#92 2026\-10\-09T12:02:16\.769295\+00:00: rank \#92 2026\-10\-09T13:02:30\.35034\+00:00: rank \#94 2026\-10\-09T14:01:34\.215134\+00:00: rank \#94 2026\-10\-09T15:01:27\.564392\+00:00: rank \#93 2026\-10\-09T16:02:06\.561056\+00:00: rank \#94 2026\-10\-09T17:03:06\.673648\+00:00: rank \#96 2026\-10\-09T18:01:02\.851052\+00:00: rank \#96 2026\-10\-09T19:01:11\.250291\+00:00: rank \#96 2026\-10\-09T20:01:32\.91907\+00:00: rank \#96 2026\-10\-09T21:02:08\.376144\+00:00: rank \#98 2026\-10\-09T22:01:19\.253075\+00:00: rank \#100 2026\-10\-09T23:01:31\.990912\+00:00: rank \#100 2026\-10\-10T08:01:00\.593376\+00:00: rank \#99 2026\-10\-10T09:01:45\.527975\+00:00: rank \#101 2026\-10\-10T10:00:53\.799353\+00:00: rank \#101 2026\-10\-10T11:00:35\.65851\+00:00: rank \#100 2026\-10\-10T12:01:11\.169647\+00:00: rank \#100 2026\-10\-10T13:00:54\.008377\+00:00: rank \#99 2026\-10\-10T14:01:43\.208766\+00:00: rank \#101 2026\-10\-10T15:01:24\.825188\+00:00: rank \#101 2026\-10\-10T16:01:10\.213328\+00:00: rank \#100 2026\-10\-10T17:00:48\.84189\+00:00: rank \#99 2026\-10\-10T18:00:57\.55132\+00:00: rank \#99 2026\-10\-10T19:00:52\.379463\+00:00: rank \#98 2026\-10\-10T20:00:21\.993912\+00:00: rank \#98 2026\-10\-10T21:01:10\.657738\+00:00: rank \#99 2026\-10\-10T22:00:50\.433569\+00:00: rank \#98 2026\-10\-10T23:01:00\.95027\+00:00: rank \#99 OpenAI's reported 372\-result math dump, including a Unique Games proof, has triggered a verification and readability crisis: Lean certificates may check, but humans have not digested the proofs\. ## The brief Scott Aaronson reports that OpenAI released 372 major mathematical results, including a proof of the Unique Games Conjecture, with Lean certificates for some but no human comprehension yet\. He describes the proofs as alien and poorly written, lists other breakthroughs, and contrasts OpenAI's undigested dump with Anthropic's compensated, digested announcement\. The post argues the Mathocalypse is real while noting P vs NP remains open and the community must adapt\. - OpenAI's advisory group included Timothy Gowers and Edward Witten; the UGC proof implies many optimization problems are NP\-hard to approximate beyond semidefinite programming relaxations\. - Dana Moshkovitz's texts describe the UGC proof as psychedelic, poorly written, with irrelevant citations and a bizarre recursive code and noise test; she needed AI help to assess it\. - Other claimed results include L=BPL, faster integer multiplication and Fourier transform, unitary synthesis, parity not in QAC0, a 2D area law, matrix multiplication in O(n^9/4), a permanent lower bound, matching algorithms, and uncomputability of rational polynomial equations\. - The model was an internal OpenAI system, not a bespoke 10,000\-agent setup; it used about 3 hours of GPT\-Pro\-level compute per problem and solved roughly 5% of about 8,000 attempted open problems\. - Two communication models emerged: OpenAI posts raw solutions, while Anthropic let Virginia Williams and Josh Alman write a digested 3SUM and all\-pairs shortest paths paper for compensation\. - Aaronson says P≠NP, P=BPP, and NEXP⊄P/poly remain unsolved, and links AI\-risk debate to Scott Alexander's open letter to Steven Pinker\. ## Discussion themes Analyzed: 2026\-10\-08T16:01:40\.299294\+00:00 Analysis sample: Based on 20 of 50 usable stored comments. Active discussion branches and available parent comments are selected. The analysis input was further shortened to fit its context limit. This sample may omit parts of the full thread. Selected themes do not measure community opinion or how common a view is. ### Human readability and cognitive compression of AI\-generated proofs Comments criticize the AI proofs as horribly written, unclear, noisy, and hard for humans to check, comparing them to messy code that still produces correct output\. Replies debate whether clarity is necessary, whether AI proofs compress understanding worse than human proofs, whether current implementations will improve, and whether human readability will become moot as AI surpasses human cognitive capacity\. Sources: [Comment 49998263]() · [Comment 49998463]() · [Comment 49998613]() · [Comment 49998761]() · [Comment 49998943]() · [Comment 49998946]() · [Comment 49999035]() · [Comment 49999918]() ### Verification and Lean/natural\-language mismatch Comments question whether the proofs have been verified or whether the Lean code could be formulated incorrectly\. A cited paper is said to claim many AI\-generated Lean proofs are mistranslated; a reply clarifies that this means the Lean proof and natural\-language proof may not coincide, not that the Lean proof is wrong\. Another reply expects later AI verification or flaw\-finding before an IPO\. Sources: [Comment 49998348]() · [Comment 49998375]() · [Comment 49998839]() · [Comment 49998939]() · [Comment 49998943]() ### Provenance and credit concerns Comments ask how one knows OpenAI independently produced the solutions rather than taking them from outside researchers\. Replies argue human mathematicians would be unlikely to forgo career\-making credit, that the Lean output is likely from LLMs, and that ideas could be exposed through mathematicians using ChatGPT\. Sources: [Comment 49999003]() · [Comment 49999028]() · [Comment 49999034]() · [Comment 49999111]() ### Historical precedent for non\-human\-comprehensible proofs Comments cite the Four Color Theorem and inter\-universal Teichmüller theory as precedents or counterpoints to the claim that proofs must be cognitively distilled and coherent for humans\. They debate whether mathematics has already accepted non\-human\-scale proofs and how the current situation differs\. Sources: [Comment 49998945]() · [Comment 49999857]() · [Comment 49999918]() ### AI hype, PR, and polarized discourse Comments characterize the event as an AI IPO PR powerplay and note polarized reactions: patronizing dismissal of skeptics versus breathless hype declaring mathematics is over\. Both sides are accused of assuming conclusions and giving or denying the company benefit of the doubt\. Sources: [Comment 49998761]() · [Comment 49998939]() · [Comment 49998491]() · [Comment 49998592]() ## Sources & coverage AI-generated summary · 2026\-10\-07T22:01:43\.633566\+00:00 Based on 7 of 12 usable stored comments, selected by depth and branch activity. This is a sample of the discussion. The model input was further shortened to fit its context limit. Article text may also be shortened. Generated using deepseek\-ai/DeepSeek\-V4\.1\-Flash. Check the linked sources for full context.