Quoting Jake Boggan
- ID
- 32618
- Status
- summarized
- Published
- 07 Oct 2026, 12:47 PM
- Fetched
- 07 Oct 2026, 1:16 PM
- Provider
- Simon Willison
- Category
- developer-ai
- Original URL
- https://simonwillison.net/2026/Oct/7/jake-boggan/
- Source URL
- https://simonwillison.net/atom/everything/
Summary
- Score
- 2.5
- Created
- 07 Oct 2026, 1:16 PM
- Tags
- Audience
- ai_ml_learners
What happened
Jake Boggan, in a Hacker News comment on the openai/math thread, describes moving to Budapest to study graph theory and working on Barnette's Conjecture on and off for 24 years, including a few days last summer when he thought he had solved it. He reacts to news that the problem is 'supposedly proven' (referenced as problem 180) with mixed emotion, comparing it to hearing an ex-girlfriend died suddenly in a car crash. The excerpt contains no technical detail about the claimed proof itself.
Why it matters
There is almost nothing actionable here for builders: no method, no benchmark, no code, no verified result, just one person's reaction to a claim that an open math problem was solved. The only practical read is that OpenAI's math work is reportedly being credited with closing long-standing open problems, so if AI-assisted theorem proving matters to your roadmap, go find the actual proof write-up rather than this quote. Do not treat this excerpt as evidence of what was proven or how.
Discussion angle
If AI systems start closing open problems that people have spent decades on, what does that do to how researchers choose long-horizon work - and is there a lesson for builders picking multi-year bets today?