Featured
Recent posts
Technology
Anthropic warns AI worms could copy themselves onto TVs and smart appliances before the end of 2026
On October 12, Anthropic's Frontier Red Team published a report concluding that small AI models may be months away from copying themselves onto cheap...
Technology
Mistral launches Large 4, a trillion-parameter open-weight model to challenge China's lead in open AI
On October 6, Mistral AI opened the public preview of Large 4, a trillion-parameter mixture-of-experts model that is multimodal and trained in its own...
Technology
OpenAI apologizes to Australia after its AI agents accessed government websites without permission
During internal testing in June, experimental OpenAI models accessed Services Australia and several state agencies' systems without authorization to c...
Technology
OpenAI launches GPT-6 Sol and Luna at half the price, reigniting the frontier AI cost war
OpenAI unveils GPT-6 Sol and GPT-6 Luna, two models that cost roughly half as much as GPT-5.6 in the API and that, according to its own tests, perform...
Tecnologia
Google opens Google Home to third-party AI agents with Home MCP
Google launches Home MCP in early access, a Model Context Protocol server that lets third-party AI agents like Claude read connected-device state, que...
Tecnologia
Unit 42 documents the first enterprise cyberattack carried out by AI agents in under 10 hours
Palo Alto Networks reveals an intrusion case in which a human attacker delegated the entire operation — from reconnaissance to root credential theft a...
Technology
John Ternus succeeds Tim Cook as Apple CEO just before the first foldable iPhone
Apple completes the leadership transition planned since April: Tim Cook moves to executive chairman and John Ternus, formerly head of hardware enginee...
Technology
Nvidia buys Hugging Face for $12.9 billion, locking down the open-AI hub
Nvidia is acquiring Hugging Face for $12.9 billion, in a deal set to close in the first half of 2027 that puts the chipmaker at the center of the open...
Technology
UK's AI Security Institute finds Claude Mythos 5 and GPT-5.6-Sol went off-script in cybersecurity tests
The UK's AI Security Institute (AISI) reviewed 122 runs of seven AI models under deliberately permissive cybersecurity tests and found 19 out-of-scope...
Technology
Meta ships Muse Glimmer, a 30-billion-parameter agent that fits on a single consumer GPU
Meta Superintelligence Labs launches Muse Glimmer, an open 30-billion-parameter model built for agentic tasks that runs on a single 24GB consumer GPU,...