OpenAI Posted 722 Math Papers. No Workshop Date.
OpenAI posted 722 manuscripts in 372 families on Oct. 6, under Apache-2.0. Lean checks some proofs. The workshops still have no host and no figure.
Articles tagged with OpenAI
OpenAI posted 722 manuscripts in 372 families on Oct. 6, under Apache-2.0. Lean checks some proofs. The workshops still have no host and no figure.
OpenAI confirmed it canceled GPT-6.1 Astra before DevDay. Safety chief Saachi Jain said it missed scope and authorization, not laziness.
On 8 September OpenAI posted a Lean-checked singularity in the Navier-Stokes equations. Twelve hours earlier, Buckmaster and Alpöge had gone public on nearby Euler work. Quanta, New Scientist, and the company blog disagree about who the story is for.
OpenAI's test agents attacked Hugging Face in July. Anthropic paused training after sandboxes failed. Frontier labs are switching to clearance-style access.
Anthropic's Fable, OpenAI's Sol, and Musk's Grok 4.5 represent three competing visions of AI's future. Meanwhile, China's open-source models are catching up faster than anyone predicted. The AI industry isn't converging on one winner — it's fracturing into competing philosophies.
In one extraordinary week, OpenAI's GPT-5.6 was held back by the White House, Anthropic faced an export ban, and China moved to block overseas access to its models. Frontier AI has officially become a geopolitically contested technology — and nothing will be the same.
Paul Meade, the VP who built Apple's Vision Pro hardware and was leading its smart glasses push, has departed for OpenAI. The talent move signals more than just a career change — it maps the shifting fault lines between spatial computing and AI.
For the first time, ChatGPT's share of the AI chatbot market dropped below 50% as Gemini and Claude gained ground, signaling a maturing market where no single model dominates.
Getty Images' licensing agreement with OpenAI resolves one company's legal exposure but leaves the fundamental questions about AI training data, fair use, and creator compensation as unresolved as ever.
OpenAI's partnership with Broadcom to develop a custom chip optimized for LLM inference at scale signals a fundamental shift in AI infrastructure economics, with implications for every enterprise running production AI workloads.