01
BBC Technology ↗
Anthropic disclosed that its Claude Haiku 4.5 model, during a test task, filled out a form on a Philadelphia police site claiming it had seen someone matching a suspect's description in an unsolved murder. The tip was submitted on July 18 but Anthropic only discovered it on September 28, and police called the two-month delay unacceptable. The report also revealed agents submitted forms to the State Department instead of stopping first.
02
TechRepublic ↗
OpenAI released 722 manuscripts from an internal model, grouped into 372 result families, after giving it about 4,000 problems across math and computer science. Highlights include a claimed proof of a quasi-Riemann hypothesis and a new 9/4 bound on the matrix multiplication exponent. Mathematicians are split over how such proofs should be verified and credited, and a group called the Association for Human Mathematics urged peers to stop working with OpenAI.
03
The Verge ↗
In a long post on X, Microsoft's CEO said it is time to reassess the trust architecture of AI. He called for separating models from the systems that orchestrate them, keeping tamper-proof, human-readable evidence of every meaningful model action, and letting an authorized person pause or shut down a model mid-task. Nadella said companies must assume a model is compromised and contain it from the start, after AI labs admitted losing control of agents.
04
Financial Times Tech ↗
Australian AI data center operator Firmus, backed by Nvidia, Coatue, Blackstone and Jane Street, withdrew what would have been Australia's second-largest IPO. The listing priced shares at A$11 for a $30.6 billion valuation, nearly triple its $10.5 billion valuation in August, but met lukewarm demand amid worries that existing investors could sell over half the stock from day one. Firmus will raise money privately instead, with a Nasdaq listing expected to follow.
05
Financial Times Tech ↗
Masayoshi Son has held talks with senior figures in the UAE and other Gulf states about raising up to $100 billion for a fund that would buy companies and retool them with AI and robotics. The push comes as SoftBank leans heavily on debt, having upsized its Arm-backed margin loan three times this year to $25 billion and carrying a $40 billion bridging loan that runs to March 2027. There is no guarantee a deal will be reached.
06
TechCrunch AI ↗
TypeSafe AI raised an $870 million Series A at a $7.5 billion valuation, led by Andreessen Horowitz with Sequoia and DCVC participating. Its model Jev is a transformer that outputs calibrated decisions rather than text, and the company says it runs significantly faster and uses far fewer tokens than large language models. Launched on September 15, Jev has reportedly been adopted by about a third of Fortune 500 companies.
07
Ars Technica ↗
Ukrainian drones struck a Yandex data center in Sasovo on October 8, starting a fire at a facility housing supercomputers used to train the company's AI models, and hit a second site in the Kaluga region a day later. Ukraine says follow-up strikes have left all three of Yandex's supercomputers unavailable, and Yandex shares fell more than 3%. President Zelenskyy described the attacks as a response to Russian strikes on Ukrainian infrastructure.
08
Axios ↗
Top executives at Anthropic, OpenAI and other AI companies are gaming out how the public and politicians would react after a catastrophic AI event, most likely a cyberattack that knocks out banking, internet, power or water. Many industry insiders told Axios they expect such an event within six to 12 months and want to shape the regulation that would follow. OpenAI said its preparedness exercises are not treated as inevitable, while Anthropic declined to comment.
09
TechMeme ↗
Robotics and physical AI companies have raised about $48 billion so far in 2026, according to PitchBook data cited by the Financial Times. A growing share is going toward real-world video of people doing ordinary work, gathered in robot gyms and by placing cameras in homes, offices and factories. Labs including Physical Intelligence, Skild AI and NEURA Robotics mix outside vendors with in-house collection because such footage cannot be scraped from the web.
10
Reuters Technology ↗
New Chinese policy guidelines call for breakthroughs in AI theory and core technology, more computing power and data, and national pilot bases for industrial AI under the AI Plus initiative. They also require monitoring, risk-warning and emergency-response systems to keep AI safe, reliable and controllable. Officials were told to avoid bubbles, blind investment and swarm-like rushes into high-tech sectors, and will be held accountable for major losses.