Search

1. US Reverses Course on Britain's Testing of AI Models: AISI Regains Pre-Release Access

   

   … matters: the UK institute regains pre-release access 17 days after the restriction was imposed — the first ti…

   Flash·Oct 12

2. Anthropic Launches Claude Haiku 5.5: From $0.10, About 75% Cheaper Than Its Predecessor

   

   … it matters: It is the first Haiku with effort controls and the first with built-in safeguards for high-risk c…

   Stories·Oct 8

3. White House Tells AI Companies Incident Reporting Is Mandatory, After Anthropic's Disclosure

   

   …ump administration has framed AI safety incident reporting as mandatory rather than voluntary — the voluntary …

   Flash·Oct 11

4. UK's ICO Names 10 AI Labs in Training-Data Privacy Crackdown: "No Justification" for Non-Compliance

   

   …, Anthropic, DeepSeek and OpenAI; the report states that broad interests such as "benefitting humanity" are no…

   Stories·Oct 9

5. One Man, One AI Stack, Nine Banks: CrowdStrike Details the AI-Driven Korean Bank Hacks

   

   …constructed the full attack chain from session logs the attacker left on a public server, and President Lee Ja…

   Stories·Oct 8

6. Google Tests Internal Gemini 4 "Carbon" for Coding; Staff Say It Rivals Claude Opus 5.5

   

   …us 5.5; staff had previously said early Argon versions lagged rivals on some coding tasks, and Google has publ…

   Flash·Oct 10

7. Anthropic Cuts Live Internet Access for All Internal Evals

   

   …ions and concedes alignment training is "not yet sufficient" for skills like search and computer use; it brief…

   Stories·Oct 10

8. Anthropic Expands Cyber Verification Program: After 129,000 Bugs Found, Its Strongest Models Open to Vetted Defenders — With Fewer Safeguards

   

   …e-tier system, letting vetted security teams use Claude Opus 5.5, Sonnet 5.5, Mythos 5.1 and future models wit…

   Stories·Oct 7