3+ day, 6+ hour ago   (391+ words) Charles Hoskinson, the founder of the Cardano blockchain, said artificial intelligence has advanced in mathematical reasoning far faster than he anticipated, with large language models now capable of producing and formally verifying proofs that were once considered beyond the reach…...

3+ day, 3+ hour ago   (1154+ words) In one week, OpenAI’s internal model solved the Navier-Stokes Millennium Prize problem — a 200-year-old fluid dynamics puzzle — using 10,000 AI agents, 88 hours of compute, and roughly $6.5 million in inference costs. The achievement, confirmed on the Moonshots podcast by physicist Alexander Wissner-Gross and…...

3+ day, 20+ hour ago   (647+ words) The sequence of events, according to Nathan Labenz and Prakash Sundaresan speaking on The Cognitive Revolution, exposes what they see as the defining problem now facing the frontier AI ecosystem: not technical capability — which is advancing at a pace that…...

3+ day, 23+ hour ago   (863+ words) OpenAI announced that an unreleased AI model solved the Navier-Stokes existence and smoothness problem, one of seven Millennium Prize Problems, in 88 hours using up to 10,000 AI agents. The company said the effort cost millions of dollars in computing power and…...

4+ day, 11+ hour ago   (263+ words) OpenAI mathematicians discuss how AI models like GPT-5 and Astra assist in mathematical research, demonstrating that AI can excel in detailed reasoning and problem persistence, and Astra's proofs in sphere packing and non-sofic groups highlight AI's potential to make significant…...

6+ day, 3+ hour ago   (857+ words) A viral social media post claiming Anthropic's Claude AI has solved the Navier-Stokes Millennium Prize problem has drawn over 2.5 million views, though no proof or review materials have appeared publicly. The Clay Mathematics Institute still lists the problem as unsolved....

2+ week, 4+ day ago   (820+ words) Zou, who works with Together AI and holds a Stanford appointment, is making a deliberate philosophical pivot. The idea is that you shouldn't tell a powerful agent how to solve a problem. You should give it a place to work…...