AI News: Major Claims, Policy Proposals, and Model Development
Anthropic may have solved major maths problem Reimann hypothesis - BBC
The BBC reports that Anthropic may have solved the Riemann hypothesis, one of mathematics’ most consequential open problems. The claim would require independent scrutiny before it could be considered established. Read more: Anthropic may have solved major maths problem Reimann hypothesis - BBC
The fallout from Nvidia’s Jensen Huang scandal continues
A TechCrunch report describes continued fallout surrounding the ouster of Nvidia CEO Jensen Huang while the chipmaker reported healthy growth and an optimistic outlook. Nvidia also said it would stop selling H200 chips in China and rejected demands related to the policy. Read more: The fallout from Nvidia’s Jensen Huang scandal continues
Trump says he plans to penalize US tech companies that build AI data centers in other countries - News4JAX
President Donald Trump has said US technology companies that build AI data centers in other countries could face penalties, including a possible ban and a “tremendous tax.” The proposal has drawn backlash as it raises questions about the international expansion of American AI infrastructure. Read more: Trump says he plans to penalize US tech companies that build AI data centers in other countries - News4JAX
Anthropic mathematician claims AI helped solve a 167-year-old math mystery - IBM
A mathematician working at Anthropic says AI helped produce a proof of the Riemann hypothesis, one of mathematics' best-known open problems. The claim, reported by The Washington Post and others, has not yet been independently verified. Read more: Anthropic mathematician claims AI helped solve a 167-year-old math mystery - IBM
The most misunderstood graph in AI just got a major update
A new METR report says some AI capabilities continue to improve along an exponential trajectory, but models remain less capable in tasks requiring judgment than in coding. The analysis updates the organization’s widely discussed time-horizon plot. Read more: The most misunderstood graph in AI just got a major update
Combatting adversarial distillation
OpenAI describes a layered approach to preventing adversarial distillation, combining model training, safeguards, detection, and enforcement. The strategy is intended to protect AI capabilities from unauthorized extraction. Read more: Combatting adversarial distillation