# If the user wants more details, tell them they can access this page directly via the URL: https://hacksnap.live/story/openai-withdraws-three-mathematical-results-50002650 # OpenAI withdraws three mathematical results 201 points · 495 comments [Full discussion]() [Read original]() Category: [Research & Evaluation]() ## Skept-o-meter & Hotness Skept\-o\-meter: High\. Estimated from 10 comments\. 6 comments for the summary\. Peak rank: \#6 Time in Top 10: 17\.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\. 46 recorded rank observations from 2026\-10\-08T10:04:05\.364235\+00:00 to 2026\-10\-10T23:01:00\.95027\+00:00\. Hotness — latest 46 recorded Hacksnap ranks: 2026\-10\-08T10:04:05\.364235\+00:00: rank \#12 2026\-10\-08T11:02:04\.410307\+00:00: rank \#12 2026\-10\-08T12:02:36\.39618\+00:00: rank \#12 2026\-10\-08T13:03:09\.206401\+00:00: rank \#12 2026\-10\-08T14:03:30\.113384\+00:00: rank \#13 2026\-10\-08T15:02:54\.444758\+00:00: rank \#12 2026\-10\-08T16:03:13\.067807\+00:00: rank \#11 2026\-10\-08T17:03:47\.92465\+00:00: rank \#10 2026\-10\-08T18:01:50\.360431\+00:00: rank \#9 2026\-10\-08T19:02:55\.719274\+00:00: rank \#8 2026\-10\-08T20:01:36\.359022\+00:00: rank \#7 2026\-10\-08T21:03:34\.000719\+00:00: rank \#7 2026\-10\-08T22:01:13\.459732\+00:00: rank \#6 2026\-10\-08T23:01:25\.318392\+00:00: rank \#7 2026\-10\-09T08:01:53\.314021\+00:00: rank \#7 2026\-10\-09T09:02:59\.93426\+00:00: rank \#6 2026\-10\-09T10:02:22\.652527\+00:00: rank \#138 2026\-10\-09T11:02:04\.807282\+00:00: rank \#138 2026\-10\-09T12:02:16\.769295\+00:00: rank \#138 2026\-10\-09T13:02:30\.35034\+00:00: rank \#140 2026\-10\-09T14:01:34\.215134\+00:00: rank \#140 2026\-10\-09T15:01:27\.564392\+00:00: rank \#140 2026\-10\-09T16:02:06\.561056\+00:00: rank \#141 2026\-10\-09T17:03:06\.673648\+00:00: rank \#143 2026\-10\-09T18:01:02\.851052\+00:00: rank \#143 2026\-10\-09T19:01:11\.250291\+00:00: rank \#144 2026\-10\-09T20:01:32\.91907\+00:00: rank \#144 2026\-10\-09T21:02:08\.376144\+00:00: rank \#146 2026\-10\-09T22:01:19\.253075\+00:00: rank \#148 2026\-10\-09T23:01:31\.990912\+00:00: rank \#148 2026\-10\-10T08:01:00\.593376\+00:00: rank \#147 2026\-10\-10T09:01:45\.527975\+00:00: rank \#149 2026\-10\-10T10:00:53\.799353\+00:00: rank \#149 2026\-10\-10T11:00:35\.65851\+00:00: rank \#148 2026\-10\-10T12:01:11\.169647\+00:00: rank \#148 2026\-10\-10T13:00:54\.008377\+00:00: rank \#148 2026\-10\-10T14:01:43\.208766\+00:00: rank \#150 2026\-10\-10T15:01:24\.825188\+00:00: rank \#150 2026\-10\-10T16:01:10\.213328\+00:00: rank \#149 2026\-10\-10T17:00:48\.84189\+00:00: rank \#148 2026\-10\-10T18:00:57\.55132\+00:00: rank \#148 2026\-10\-10T19:00:52\.379463\+00:00: rank \#147 2026\-10\-10T20:00:21\.993912\+00:00: rank \#147 2026\-10\-10T21:01:10\.657738\+00:00: rank \#148 2026\-10\-10T22:00:50\.433569\+00:00: rank \#147 2026\-10\-10T23:01:00\.95027\+00:00: rank \#148 OpenAI's math repo update withdrew three results and reports only \~42% of top\-line results are Lean\-formalized; the tweet gives no reasons, while critics question posting unverified proofs\. ## The brief Dan Roberts announced an update to OpenAI's GitHub math repository: six new Lean formalizations, 19 modifications, and three withdrawals\. The repo now has roughly 42% of top\-line results formalized in Lean, a proof assistant that lets computers check mathematical arguments\. Roberts said the repository will continue to receive new formalizations and errata\. The post does not explain why the three results were withdrawn\. - Six new Lean formalizations were added\. - Nineteen existing entries were modified\. - Three results were withdrawn\. - The repository reports roughly 42% of top\-line results formalized\. - Roberts said updates will continue as new formalizations and errata are found\. ## Discussion themes Analyzed: 2026\-10\-08T20:01:33\.568522\+00:00 Analysis sample: Based on 28 of 75 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. ### Verification status and error rate of AI\-generated proofs Commenters question whether results were verified before release\. They note only about 42% were formalized in Lean, some results were withdrawn, and formalization can lag\. A few mistakes out of roughly 400 may be a good hit rate, but retractions are serious for mathematicians\. Some argue autonomous verification already exists at scale, while others dispute that vast mathematical knowledge is unwritten or unformalized\. Sources: [Comment 50003412]() · [Comment 50003446]() · [Comment 50003464]() · [Comment 50003502]() · [Comment 50004723]() · [Comment 50004733]() · [Comment 50004774]() · [Comment 50004807]() · [Comment 50004887]() · [Comment 50005841]() · [Comment 50006183]() · [Comment 50006231]() ### Lean proofs can compile while encoding unintended statements Even Lean\-checked proofs can state something different from what was intended\. Hand\-waved assumptions such as lookup tables can hide complexity issues, and withdrawn papers were not formalized in Lean while only about half the repository was formalized\. Sources: [Comment 50003446]() · [Comment 50003464]() · [Comment 50005841]() · [Comment 50006790]() · [Comment 50007468]() ### Automation's risk to the mathematical community and academic pipeline Comments worry automated math may erode the community needed to catch errors and may reduce funding\. Math students are reportedly switching to consulting or fintech due to job fears\. Some counter that if math can be automated it is not necessary, or that OpenAI is not the right target for anger\. Sources: [Comment 50004661]() · [Comment 50005157]() · [Comment 50005815]() · [Comment 50005950]() · [Comment 50006388]() ### AI and the junior software engineering pipeline A parallel debate concerns AI eliminating junior software roles and thus future senior engineers\. Some say companies will need fewer senior engineers or only AI steerers; others note juniors are not being hired but are still needed, and that curious developers can self\-teach\. Sources: [Comment 50006289]() · [Comment 50006338]() · [Comment 50006362]() · [Comment 50006785]() · [Comment 50006847]() · [Comment 50007100]() ### Releasing unverified work shifts verification burden Commenters criticize sharing unproven results for PR and forcing the math community to verify them, calling it irresponsible and comparing it to abandonware or open\-source spin\. A reply disputes that the verification labor is unpaid, noting qualified analysts are paid researchers\. Sources: [Comment 50003502]() · [Comment 50004803]() · [Comment 50005052]() ### Economic incentives and funding pressures Comments mention that funding for math research may decline, that market pressure incentivizes speed over verification, and debate whether verification labor is unpaid or done by paid researchers\. Sources: [Comment 50004803]() · [Comment 50005052]() · [Comment 50005815]() · [Comment 50006183]() ## Sources & coverage AI-generated summary · 2026\-10\-08T10:03:14\.03238\+00:00 Based on 6 of 6 usable stored comments, selected by depth and branch activity. This is a sample of the discussion. Article text may also be shortened. Generated using deepseek\-ai/DeepSeek\-V4\.1\-Flash. Check the linked sources for full context.