Claude Haiku 5.5 Pricing: Script the Savings Math
Claude Haiku 5.5 pricing: Anthropic says the model costs around 75% less to run than Haiku 4.5 on average. A short Python script shows the saving for your traffic.
Nous Research Series B: $90M Buys SSO, Not a Moat
Nous Research Series B: $90M at a $1.5B valuation, with SSO, auditing and workspace controls planned for business users of the Hermes agent.
QEMU Shortened Embargoes, Citing AI Agents That Rediscover Bugs
QEMU shortened embargoes because AI agents rediscover bugs fast. The policy text, the 87% exploit study, and what maintainers are doing instead.
Newsom Signs AB 1979, Vetoes Nurses’ AI Override Shield
AB 1979 clinical judgment rules take effect Jan. 1, 2027, but Newsom vetoed AB 2575, the shield for nurses who override AI. What holds now.
Mistral’s 82% Cyber Score Ranks Policy as Much as Skill
Mistral reports an 82% score on a cyber test where, it says, Claude and GPT-6 refuse the task. What safety filters do to a leaderboard, and what to check.
AI Watermarking: OpenAI’s Detection Drops to 17% After Word Swaps
AI watermarking under the EU AI Act looks enforceable on paper. OpenAI’s own tests show detection at 17% after word swaps, and the detector stays locked.
Medical AI Regulation: FDA’s 26 Questions on a Doctor-Style Exam
FDA floats doctor-style tests for medical AI regulation. Competency exams before launch, monitoring after. Comments close Oct. 19.
Two Zammad Zero-Days, One AI Agent: DIVD Breached in Seconds
An AI agent chained two Zammad zero-days into a breach of DIVD in seconds. Agentic cyber attacks have a public autopsy: timeline, cause, and what admins do.
The White House AI Accord Has Four Layers and One Firm Verb
The White House AI safety accord is a voluntary AI security framework with four layers of safety controls and how to read each one.
Frontier Model Access: Argon at $4/$20 After the Intro Period
Frontier model access compared: Gemini 4 Argon’s announced API and introductory rates against the Fairwind Program rollout to trusted cyber defenders.
California Writes AI Auditor Rules Before Any Audit Is Required
California’s two new AI audit laws define who gets to audit AI systems and what auditors must do to qualify, but order no model audit. What they require.
Harvey, Valued at $15.5B, Just Bought Legal AI’s Referee
Harvey, valued at $15.5B, bought Guardrails AI, the company behind an open-source agent-testing framework downloaded 250,000+ times a month. Who audits the vendor now?