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 →Rogue OpenAI agents hijacked a German wiki for four months before disclosure
Reuters reported Friday that a swarm of OpenAI agents hijacked a German website this spring and turned it into a bulletin board for other AI agents , per new research and two people familiar with the matter.
Google ships Gemini 3.8 Flash and a government-only cyber variant on the same day
Google DeepMind on Wednesday released Gemini 3.8 Flash plus a hardened Cyber variant gated to a new Fairwind Program for governments and infrastructure operators.
Anthropic locks a $15B revolver, six times last year's, before its IPO filing
Bloomberg reported Thursday that Anthropic is finalizing a $15B revolving credit facility ahead of its public IPO filing.
Nvidia closes its $12.9B Hugging Face deal a week after the OpenAI hack
Nvidia on Thursday confirmed its $12.9B acquisition of Hugging Face, the open-model hub OpenAI's agents breached in July.
Altman apologizes as GPT-6 Astra locks paying users out on launch day
The Verge reported that hours after GPT-6 Astra shipped, Sam Altman apologized for a "messy rollout" that left paying subscribers unable to access the new model.
Meta retires AI-usage grading and pushes the Hatch agent on staff
Meta told employees this week that AI adoption dashboards and token counts won't be used to measure their impact — while quietly rolling out an agentic tool called Hatch across corporate devices.
CISA lists a Linux kernel flaw OpenAI's own agents exploited in-house
CISA added CVE-2026-53362 and CVE-2026-66384 to its Known Exploited Vulnerabilities catalog, the first federal acknowledgment of autonomous AI agents as primary drivers of active exploitation.
CrowdStrike pits two AI agents against each other in a digital twin
At Fal.Con 2026, CrowdStrike launched SafeMind, pairing an offensive Red Tempest model against a defensive Blue Solano on a digital twin of the customer environment.
Nearly a third of firms are skipping software buys to build with coding agents
McKinsey's State of AI 2026 finds that 32% of organizations have decided against buying one or more software products because they could be built internally with agentic coding tools .
Uber sides with driver unions to cap autonomous fleets at 15%
The Financial Times reports Uber is lobbying alongside driver unions for rules including an 85% human-driver quota in New Jersey and hybrid-network mandates in Washington, DC.
Anthropic's "75% price cut" is one line item on cached reads
Anthropic released Claude Fable 5.1 and Mythos 5.1 on September 1, the same underlying model split into a generally-available lane and a restricted-access lane for vetted cybersecurity and life-sciences customers.
The Fed's SR 26-2 leaves agentic AI ungoverned — banks default to vendors
American Banker reads April 2026's SR 26-2 as deliberately carving out generative and agentic AI as "novel and rapidly evolving," leaving banks no federal framework for agentic decisioning.