STORY RECORD
Former OpenAI insider jokes on Jacobian claim
Will Depue, ex-OpenAI researcher who worked on Sora 1/2 and o3/4o post-training, posted a deadpan joke about being unable to get 'this thing' to solve the Jacobian conjecture.
The gag lands because five days earlier, Anthropic researcher Levent Alpoge claimed his 'close friend fable' (widely understood as Anthropic's Claude Fable 5) helped produce a counterexample to the 87-year-old open problem—a claim that hit roughly 39 million views on X, drew a Lean formalization within hours, independent reproduction by OpenAI Codex, and a refined conjecture from GPT-5.6 with no preprint or peer review yet.
Depue's tone suggests playful frustration: as someone who helped build frontier models, he is publicly acting like he cannot replicate the feat himself, undercutting the hype. The photo attached to the post could not be retrieved, so its content is unknown.
Why It Matters
The post captures a mood shift inside the AI building community. Depue has credible insider status—he worked on post-training for o3/4o and pretraining moonshots at OpenAI.
That he is joking about not getting 'this thing' to solve the Jacobian conjecture, rather than celebrating an AI triumph, signals that even for the people who built these models, the Alpoge/Fable 5 claim feels like a singular, non-reproducible moment rather than a capability anyone can just dial up.
It is social texture from inside the tent, not outside commentary.
The Facts
11Will Depue posted on X at 06:17:34 UTC on July 25, 2026: 'guys i don't get it how do you get this thing to solve the jacobian conjecture' with an attached photo.
direct evidence · confidence 1
Depue's account bio describes him as 'dei ex machina, prev @ openai (sora 1 & 2, posttraining o3/4o, pretraining moonshots)' with 70,974 followers, Blue verified.
direct evidence · confidence 1
Five days earlier on July 20, 2026, Anthropic researcher Levent Alpoge posted that 'the Jacobian conjecture is false,' crediting 'close friend akhil' (mathematician Akhil Mathew) and 'close friend fable' (widely understood as Anthropic's Claude Fable 5 AI model) for the work.
direct evidence · confidence 1
Alpoge provided an explicit polynomial map F(x,y,z) = ((1+xy)^3 z + y^2(1+xy)(4+3xy), y + 3x(1+xy)^2 z + 3xy^2(4+3xy), 2x - 3x^2 y - x^3 z) from C^3 to C^3 with claimed Jacobian determinant -2, asserting it sends three distinct points to the same image, demonstrating non-injectivity.
direct evidence · confidence 1
Alpoge's original post reached approximately 39 million views, 42,000+ likes, 5,300+ retweets, and 3,000+ quotes as of later observation.
direct evidence · confidence 1
Within roughly two hours of Alpoge's post, Stanford math student Dean Cureton posted a Lean formalization of the counterexample.
direct evidence · confidence 1
OpenAI researcher Aaron Lou reported that an internal version of OpenAI Codex independently discovered essentially the same counterexample without web search.
direct evidence · confidence 1
Fields Medalist Timothy Gowers called the claimed counterexample 'pretty amazing' but said it is 'not end-of-mathematics territory.'
direct evidence · confidence 1
No preprint, code repository, or independent peer review has been published for the Jacobian conjecture disproof claim as of July 25, 2026.
direct evidence · confidence 1
The Jacobian conjecture, posed in 1939 by Ott-Heinrich Keller, is a central open problem in algebraic geometry roughly stating that a polynomial map from C^n to C^n with a non-singular Jacobian determinant has a polynomial inverse.
direct evidence · confidence 1
Depue's post had 106 likes, 5 replies, 2 retweets, 7 bookmarks, and 2,678 views as of initial observation approximately 30 minutes after posting, with a 'rising' attention tier.
direct evidence · confidence 1
Still Open
4- OpenThe content of the photo attached to Depue's post could not be retrieved and analyzed, so its role in the joke is unknown.
- OpenWhether Depue is genuinely expressing frustration about AI math capabilities, performing satire for an audience, or specifically referring to a different AI model (not Fable 5) is inferred from context, not explicitly stated.
- OpenThe Jacobian conjecture disproof claim by Alpoge has not undergone peer review and the mathematical community has not formally verified it; the Jacobian conjecture is noted as 'infamous for attracting wrong proofs' by mathematician Qiaochu Yuan.
- OpenThe full chat history with Fable/Claude 5 that produced the counterexample has not been released, and it is unclear how autonomous the AI's contribution was versus directed by Alpoge.