Anthropic browser example can expose an unlabeled password field in its output
A reported flaw in Anthropic’s browser-toolset example can put an unlabeled password value into a tool result.
Independent monitor finds no qualifying Opus 5.5 change on its fixed test
LiveNerf found no qualifying accuracy change across three completed windows, while separate users reported worse coding experiences.
NVIDIA, SK Telecom and Rebellions plan a hybrid AI inference rack
The companies signed a development agreement for a GPU and Korean NPU system expected to be ready for deployment in 2028.
NVIDIA publishes serving recipes for four Korean AI models
NVIDIA’s Dynamo project lists deployment recipes for four Korean models, with different levels of benchmark and production readiness.
Agent skills and browser tools draw a burst of developer attention
A GitHub Trending snapshot shows interest in reusable agent workflows, while maintainers warn that important browser controls are optional.
Anthropic’s filing sharpens a debate over AI’s investment boom
Anthropic reported $33.2 billion in nine-month revenue alongside a $1.3 billion operating loss and substantial long-term commitments.
NVIDIA describes Korean agent security integrations and model access
NVIDIA says Korean partners are exploring or integrating agent controls, but the announcement does not demonstrate a production deployment or security outcome.