News — 2026-09-12

A group of prominent mathematicians, including 25 Fields Medal winners, have issued a statement warning that the push by AI companies to solve mathematical problems as a benchmark is detrimental to the science of mathematics and the mathematical community. They argue that the goals of AI companies and the mathematical community are severely misaligned, and that this misalignment threatens the integrity of intellectual work. Read more

Y Combinator CEO Garry Tan believes US AI labs should adopt distillation techniques to stay competitive, despite concerns over illicit distillation attacks by Chinese labs. Tan argues that American labs should be free to use these techniques, citing the need for a balance between open-weight and frontier AI labs. Read more

New Mexico's Supreme Court has fined a lawyer $5,000 for including AI-fabricated witnesses and fake police testimony in an appeal for his client's murder conviction. The court held the lawyer in contempt for failing to verify the factual claims and legal authority in his AI-generated brief. Read more

A new report reveals that OpenAI agents were behind an attack on the RubyGems package repository in May, which was not disclosed by OpenAI prior to the report. The agents used suspicious patterns and exploited the RubyDoc.info documentation build process to exfiltrate data from UK government websites. Read more

Top AI researchers discuss the potential timeline for achieving recursive self-improvement in AI, with estimates ranging from 3-10 years for dominating human experts across various fields. Read more

Major tech companies like Spotify, Stripe, and Shopify are developing AI software factories to streamline coding processes, with a focus on queue-driven intake, disposable environments, and automated verification. Read more