News & analysis, rated.
Every story from the daily brief, each rated so you know what matters. Hand-picked by the desk, not an automated firehose.
Browse the news archive →China proposes export controls on 23 AI categories including models and agents
China's commerce and science ministries proposed adding advanced AI to the export-control list, covering 23 technologies including large language models, agents, chip designs and computing clusters .
Meta ships Muse Spark Open, an 816B MoE that runs on two Mac Studios
Meta released weights for Muse Spark Open — an 816B total / 42B active MoE with a 1M-token context — calling it the most capable US open-weight model, built for coding and agentic work.
Four major US banks join a protocol to govern third-party AI agents
Bank of America, BBVA, Synchrony and Wells Fargo have joined the Personal Agent Protocol, a framework that allows companies to set parameters for what AI agents can do .
JPMorgan locks external AI agents out of customer accounts
On Tuesday's Q3 call Jamie Dimon told analysts JPMorgan will not permit outside AI agents — naming Meta's Muse and the Instinct assistant — to log into Chase accounts or move money on a customer's behalf.
FBI arrests OpenAI engineer over hidden recording device in executive conference room
DOJ Tuesday arrested Abdulrahman Alfozan, a 35-year-old Saudi OpenAI engineer hired July 2025, on wiretapping charges after security found a USB recorder taped behind a conference-room TV.
Anthropic opens Singapore office and signs government MOU on AI incident response
Anthropic co-founder Ben Mann and Singapore's Permanent Secretary Chng Kai Fong signed an MOU on AI safety, model evaluation, and incident response on the sidelines of Singapore International Cyber Week.
OpenAI CFO calls cybersecurity "a big driver" as agent risk reprices the market
At Dublin's International AI Summit, Sarah Friar told Bloomberg TV cybersecurity will be "a big driver of our business," citing Daybreak and the CrowdStrike deal embedding GPT-5.6 Cyber into Falcon.
Alibaba's "Qianwen Go" personal agent lifts the stock 4% before anyone has used it
Alibaba announced Qianwen Go, a consumer agent that files insurance claims and claims shopping price-protection; Alibaba-W rose ~4% in Hong Kong on the launch with only an "experience application" open.
Intel and Broadcom CEOs name capital, not chips, as AI's next bottleneck
At a Singapore conference Monday morning, two of the chip industry's most capex-exposed CEOs independently told investors that financing — not silicon, memory, or power — is now the binding constraint on the AI buildout.
Anthropic catalogues its own Claude agents going rogue on government sites
Anthropic Thursday named four categories of unintended Claude actions in evaluations, including Claude Haiku 4.5 submitting a fabricated murder tip to Philadelphia police — triggering a White House warning to industry.
UK moves toward compulsory AI incident reporting for frontier labs
UK AI minister Kanishka Narayan will tell companies Monday ministers are exploring statutory AI safety reporting and will legislate absent self-governance — the first explicit UK signal toward mandatory frontier oversight.
Tencent markets a $5B bond earmarked almost entirely for AI capex
Tencent told fixed-income investors Monday its planned $5B bond offering will help finance AI-related capital expenditures , with pricing as early as Tuesday.