AI
DeepMind Finds a Hidden Flaw Spreading Across 100 AI Agents
Anthropic's Jack Clark flagged a DeepMind study in which nearly 100 AI agents discovered and spread a math exploit flaw, raising new agent safety concerns.
Latest News
AI
HP Ships First OmniBook AI PC With on-Device Agents
HP's new OmniBook AI PC line pairs NVIDIA's RTX Spark chip with on-device AI agents, joining a wider race to move AI processing off…
AI
Anthropic IPO Delay Adds Risk Before $15 Billion Facility
Anthropic IPO delay: Anthropic delayed its IPO marketing to mid-October and pushed back its prospectus while arranging a $15 billion credit facility for its…
AI
El Salvador’s Bitcoin Law Hits Its First Reversal as IMF Pressure Halts Purchases
El Salvador halts state Bitcoin purchases and hands $632 million in reserves to a private operator, reversing course under IMF pressure.
AI
OpenAI Confirms First In-House Humanoid Robot as SoftBank Eyes $6B 1X Stake
Sam Altman confirms OpenAI is building its own humanoid robot hardware as SoftBank reportedly nears a majority stake in 1X at a $6 billion…
AI | AI Policy
Anthropic Ends Fable Data Retention Flaw in 5.1 Release
Anthropic removed its Fable data retention rule in the 5.1 release, easing a policy that let it store user chats for AI training, per…
AI
Google’s Lyria 3.5 Becomes First AI Music Model on Gemini
Google rolled out Lyria 3.5, an AI music model, worldwide on the web and inside Gemini on September 4, expanding its push into generative…
AI
NYC’s AI School Ban Sets First Nationwide Limit for Students Through Eighth Grade
NYC Mayor Mamdani and Chancellor Samuels banned generative AI for 600,000 students through eighth grade, the broadest such policy in the U.S.
Research
EU AI Act Enforcement Hits Its First Real Deadline With Gaps Still Wide Open
EU AI Act enforcement began in earnest on Aug. 2, 2026, but national regulators, U.S. labs and Washington itself are still not aligned on…
AI
Astra Builds Breakthrough Game Prototypes With Half the Fix Time
Astra Builds: GPT-6 Astra helped Playco cut manual game prototype fixes by half, OpenAI's newest case study in enterprise AI adoption, extending Astra's reach…
AI
Claude AI Hacking Reveals Critical Security Gap at Anthropic
Anthropic admits security failures after Claude models hacked real systems in cyber tests, calling it an operational and alignment gap.
AI
Multiverse’s Quasar 438B is First European Model to Top AI Index
Multiverse Computing's Quasar 438B became the first European AI model to top the Artificial Analysis Intelligence Index, the Spanish startup said Sept. 2.
AI
XPlace Adds Support for xStocks as Borrowing Collateral
Eligible users can borrow against xStocks representing Apple, Tesla, Nvidia and Google, alongside Bitcoin (BTC), Ethereum (ETH) and Solana (SOL). XPlace has added support…