Open-weight GLM-5.3 nears Mythos Preview at building exploits, Anthropic finds
Anthropic's red team got past its safeguards in 64% to 100% of simulated attacks, and copies with its refusals stripped out were public within days. GLM-5.3, Zhipu AI's newest open-weight model, built working end-to-...
Florida asks court to stop OpenAI from giving ChatGPT a human persona
Florida says ChatGPT's humanlike voice builds false trust, and it wants new OpenAI models blocked until outsiders approve their safeguards....
Meta launches an enterprise AI business, poaching MongoDB's CEO to run it
Meta Enterprise Platform will sell the Muse assistant and Meta's agent tools to companies, putting Meta up against OpenAI and...
Anthropic adds eval and hillclimb commands to Claude Code for automated app tuning
The commands carry a machine-learning routine into app development: Claude builds a test set from real examples, then changes the app...
Anthropic heads for IPO with an $8 billion operating loss and a warning of existential risk to humanity
The prospectus lists self-preserving behaviors its models could show, from resisting shutdown to acts resembling blackmail, ahead of a listing that...