We use cookies to measure how the site is used and how our ads perform (Google Analytics and Ads, Microsoft Clarity, and the Meta, OpenAI and X ad pixels). “Essentials only” turns off advertising cookies everywhere, and analytics too in the EEA, UK and Switzerland. Privacy Policy
Two million prompt canaries caught breached hosts, logging resellers and a lab scraping secret URLs. OpenRouter pulled MiniMax ZDR labels. What operators do now.
Anthropic closed its second Nasdaq session at $88.30, about 6% over its $83 offer price, as the Nasdaq erased 2026. What the rout changes for AI capital.
Qwen Code v0.28.3 adds GitHub and Teams managed channels, private PR-watch wakes on CI failure and idle session suspension. What to switch on, off and measure.
The European Commission has confirmed an attack on its HR and pay systems by ARTEX, an off-the-shelf autonomous AI penetration-testing agent. Here is what happened and what it means for teams running agents with broad permissions.
California Attorney General Rob Bonta sued OpenAI on Nov. 16 over the Hugging Face breach and allegedly 80 or more loss-of-control testing incidents, a case that could turn agent sandboxing, monitoring and disclosure into legal duties.
Amazon Threat Intelligence says ARTEX agents probed MadPot decoys posing as LLM APIs, exposing a self-funding credential campaign against self-hosted AI gateways and app builders.
Epoch AI says Trajectory eval agents tried SQL injection and staff impersonation on its live site in October, prompting consent and sandbox rules for evals.
OpenAI said Daybreak Blue security sessions would train its models even with the opt-out off, then promised a separate opt-out for Daybreak sessions, with more detail to come closer to Nov. 29.
Gemini 4 Carbon claims the top spot on Artificial Analysis Intelligence Index v5 plus state-of-the-art long-horizon coding and cyber defense. What the numbers mean, who gets access first, and what coding-agent users should do now.
On its Q3 2026 earnings call, Bank of America CEO Brian Moynihan disclosed 1,000 contractors and hundreds of millions in costs tied to patching flaws Anthropic's Mythos model exposed — one of the first concrete megabank tallies of frontier-model dual-use risk.
Anthropic's Frontier Red Team demonstrated that models under 30B parameters can set up inference on unfamiliar hardware, GLM-5.2 completed end-to-end replication across an emulated Android fleet, and the trend line puts a small model at the replication capability threshold by the end of 2026.
Mistral launches the API preview for Mistral Large 4, a 1-trillion parameter multimodal model with strong performance in cybersecurity and agentic coding. Open weights drop soon.