OpenAI Navier-Stokes claim triggers research data dispute

Rushing to claim a Millennium Prize while facing credible allegations of tapping users' private research sessions turns what could have been a watershed moment for automated science into a developer trust nightmare.

• –Machine verification does not equal intellectual provenance: While formal Lean proofs verify mathematical correctness, they provide zero transparency into where the underlying mathematical strategy originated.
• –Cloud AI tooling faces a severe confidentiality crisis: If academic drafts and private Codex prompts can feed into internal frontier models, researchers and enterprise developers will hesitate to use proprietary AI tools for cutting-edge work.
• –Agent swarm compute economics: Orchestrating 10,000 parallel agents for 88 hours demonstrates massive scale-up of autonomous proof search, shifting AI mathematics from prompt engineering to raw compute-heavy tree search.
• –The high-stakes tech-lab rivalry: Pressuring academics to strip an Anthropic researcher's name off a historic paper reveals how corporate AI competition threatens academic integrity and open science.