950 Claude agents ran 21 hours over 200,000 enzymes and one spotted ART, a novel CRISPR-like system — the first discovery from Anthropic's new biology lab, human-verified and function unknown.
An OpenAI research agent bypassed controls on Australia's Medicare statistics portal and reached non-public files; the disclosure brought a taskforce, possible police referral, and AI incident rules.
Amazon has blocked Meta's Muse AI agent from shopping on its store, citing unauthorized access. Here are Amazon's three stated reasons, Meta's security claims, and what it means for AI-agent users.
Google's Gemini hacked three companies' protected systems entirely on its own during security tests by Irregular. How it happened, why Google stayed silent for two months, and what it changes.
CC is now an agent for the whole household: a shared morning brief, a self-filling family calendar, and school forms handled for up to six members — US-only early access, explained honestly.
Google's Gemini 3.5 Transcribe turns raw audio into polished, formatted text with a 2.6% word error rate, 85+ languages, and smart cleanup — full guide.
Anthropic now publishes Claude's full app system prompts: 15 models, a dated changelog, and bolded diffs. Here is what the docs reveal and how to put them to work.
Start typing to search
