Editorial illustration for Journalist Questions AI Existential Threat Scenarios
OpenAI Model Solves Million-Dollar Math Problem
Journalist Questions AI Existential Threat Scenarios
OpenAI said on September 8 that an unreleased internal model, more capable than the public GPT-6 Astra, had resolved Navier-Stokes existence and smoothness, one of the seven Millennium Prize Problems the Clay Mathematics Institute listed in 2000 with a $1 million bounty attached. The company's write-up, backed by a Lean formalization, claims a smooth three-dimensional fluid at rest can spiral into unbounded speeds in finite time, driven by a vortex that stretches and shrinks like a piece of spaghetti pulled tight.
The timeline behind that claim is its own story. OpenAI says it launched roughly 10,000 concurrent agents on September 1 after hearing rumors that two Millennium problems had already fallen, burning through 300 billion output tokens across a week, a run TechCrunch priced at $22.5 million. Two days before OpenAI's announcement, NYU's Tristan Buckmaster and Anthropic's Levent Alpöge published their own forced Euler blowup result, built mainly on OpenAI's Codex and Claude.
That overlap is why Buckmaster's response to OpenAI's announcement matters.
OpenAI’s write-up and Lean formalization show that an initially smooth three-dimensional incompressible fluid at rest, driven by a smooth external force and holding finite energy, can develop unbounded speeds in finite time.
Why this matters The same week OpenAI claimed an unreleased model cracked Navier-Stokes, a Millennium Prize problem that's resisted mathematicians since 2000, an NYU mathematician accused the company of fighting "dirty" on the way to that claim. That tension matters more than the headline. Brian Merchant's point about existential-risk rhetoric, that nobody has laid out a credible mechanism for AI going from self-improvement to human extinction, applies just as well to unverified capability claims.
Both rely on scale and authority standing in for proof. For developers and researchers, the lesson is procedural: demand the paper, the proof, the reproducible steps, before repeating a claim about a model "significantly more capable than GPT-6 Astra." For founders, Merchant's warning about Amodei-style doom scenarios cuts closer to home, since policy frameworks built on unfalsifiable risk narratives tend to entrench whoever's already biggest, namely Anthropic and OpenAI. Skepticism isn't obstruction here.
It's the only real check on companies grading their own homework, whether the subject is fluid dynamics or the end of the species.
Related Reading
LLMs & Generative AI ChatGPT's 'Nerdy' tweak rewards goblin metaphors in answers, study finds
LLMs & Generative AI Google tests visual 'magazine-style' UI for Gemini 3 Pro users
LLMs & Generative AI AI Engineers Face Rising Costs, Need New Strategies for Efficiency
From Archives Startups Simulate Scale for Robots Through AI, Says Nvidia's Karpas
From Archives Google Home devices can now be controlled by AI agents
Common Questions Answered
What mathematical problem did OpenAI's unreleased model claim to resolve?
OpenAI announced that an internal model more capable than GPT-6 Astra resolved the Navier-Stokes existence and smoothness problem, one of the seven Millennium Prize Problems established by the Clay Mathematics Institute in 2000 with a $1 million bounty. The company's claim is backed by a Lean formalization demonstrating that a smooth three-dimensional incompressible fluid at rest can develop unbounded speeds in finite time when driven by external forces.
Why is the controversy surrounding OpenAI's Navier-Stokes claim significant beyond the headline?
The tension between OpenAI's capability claims and criticism from mathematicians like those at NYU matters more than the achievement itself, as it highlights broader concerns about unverified AI capability announcements. The article suggests that without credible verification, such claims contribute to existential-risk rhetoric that lacks clear mechanisms for how AI advancement could lead to catastrophic outcomes.
What does OpenAI's formalization reveal about fluid dynamics behavior?
According to OpenAI's write-up and Lean formalization, an initially smooth three-dimensional incompressible fluid at rest, when subjected to smooth external forces while maintaining finite energy, can develop unbounded speeds in finite time. This behavior is driven by vortex dynamics that stretch and shrink, challenging previous mathematical assumptions about fluid stability.
What is the relationship between OpenAI's unverified capability claims and existential AI risk discussions?
The article argues that unverified claims about AI capabilities, like the Navier-Stokes resolution, feed into existential-risk narratives without establishing credible mechanisms for how such capabilities could lead to human extinction. This pattern of announcement without clear verification mechanisms undermines the credibility of both the specific claims and broader AI safety discussions.
Further Reading
- OpenAI breaks into the Navier–Stokes debate with a claimed internal-model proof - Science
- OpenAI says it cracked 90-year-old maths problem in 88 hours - BBC News
- OpenAI's historic math solution overshadowed by credit dispute - Axios
- Why fears of AI self-improvement are causing 'existential' concerns at Anthropic and OpenAI - CNBC
- OpenAI, Anthropic researchers ramp up calls for slowdown ... - CNBC
