OpenAI's Navier-Stokes claim, and the fight over credit
OpenAI says its agents proved Navier-Stokes blowup. The mathematicians on the neighbouring result say the credit talks turned ugly.
- Observed threads
- 9
- Source clusters
- 3
- Observed
- Updated
The proofs
Alpöge and Buckmaster published Lean-checked blowup for forced Euler on Sept 7; Terence Tao called it a likely route to Navier-Stokes. OpenAI announced its own proof at 17:20 UTC, by agents on an unreleased model, not yet checked outside.
The dispute
Buckmaster says OpenAI began days after learning of their route, would not say whether their private Codex sessions were used, proposed an authorship without Alpöge, and asked why he would ruin his career.
The reply
Bubeck called the allegations false, then at 17:39 UTC said OpenAI started on rumours about Anthropic, saw none of their work, never asked for Alpöge's removal, and apologised for the career remark. The training question went unaddressed.
Sources
- OpenAI announcement on X ↗
- Bubeck first response ↗
- Bubeck fuller response ↗
- Terence Tao on the Euler result ↗
- Scientific American ↗
- OfficeChai on Bubeck's denial ↗
- Quomodocumque on the blowup ↗
- X: konstiwohlwend timeline ↗
- X: rynorhn on the statement ↗
- X: Thom_Wolf ↗
- X: kimmonismus explainer ↗
- X: ns123abc ↗
- X: bubbleboi ↗