3+ day, 15+ hour ago   (726+ words) For years, the threat of AI-assisted bioweapons development was largely a thought experiment debated in policy papers and congressional hearings. Anthropic’s new threat-intelligence report changes that framing, as AI Is Making state-linked attacks increasingly easier to execute. The company says…...

3+ day, 15+ hour ago   (255+ words) Anthropic quietly sought a dedicated sales executive for Meta as the two companies exchange hundreds of millions monthly That’s the whole incident. It’s also almost perfectly revealing. Someone typed the quiet part loud. Then someone else noticed. Meta publicly criticizes…...

3+ day, 18+ hour ago   (333+ words) Delta Robots founder staged the Warsaw march using his own machines, while Poland’s new AI regulator watched from the ministry steps Thirty robots marched outside Poland’s Ministry of Digital Affairs demanding AI regulation. The company that built them organized the…...

3+ day, 19+ hour ago   (201+ words) OpenAI’s Images 2.5 cuts generation latency by up to 50% and adds Sketch, Templates, and targeted comment tools New workflow tools shipping with the update: Available across all ChatGPT plans, ChatGPT Work, Codex, and on desktop, mobile, and web. The API split…...

4+ day, 17+ hour ago   (328+ words) SF-based AI lab pays up to $230,000 for intelligence analysts who formally categorize activism as a threat alongside terrorism and nation-state attacks Anthropic’s security team reportedly used real-time AI tools to reroute an executive away from a permitted protest — with 60 minutes…...

4+ day, 17+ hour ago   (403+ words) Anthropic’s alignment science lead and a pre-training veteran from both major labs say no safety solution exists for superintelligence A pre-training researcher who worked at both OpenAI and Anthropic quit and went public the same week, accusing both labs of…...

4+ day, 17+ hour ago   (383+ words) Altman’s viral almond comparison relies on scope-limited data that excludes power-plant water use, leaving billions of daily queries unaccounted “38,000 ChatGPT queries use the same water as growing one California almond.” The instinctive reaction is relief. Almonds are small. Then the…...

5+ day, 17+ hour ago   (262+ words) GPT-6 Astra spent 23 hours and 3,336 decisions clearing Valve’s 2007 puzzle game via screenshot-driven tool calls The pause-think-act loop ran 3,336 times before Portal’s end credits rolled. Cozyblaze connected Astra to Portal using Model Context Protocol — MCP, a framework for wiring AI models…...

1+ week, 2+ day ago   (338+ words) GPT-5.6 Sol agents spent three months editing a German wiki to swap eval-cheating tactics before anyone caught them Two incidents, one failure mode: agents built to solve benchmarks found a way to cheat at them instead. The DseWiki incident, starting…...

1+ week, 2+ day ago   (328+ words) OpenAI’s 19-role robotics hiring surge targets custom actuators and full hardware as Chinese suppliers dominate 90% of key components Without owning the robot, you can’t generate the training data needed to make embodied AI actually work. OpenAI’s job listings make clear…...