Open Secure AI Alliance Proposes SAFE Incident-Reporting Guidelines as Membership Tops 120
The Open Secure AI Alliance, launched by Nvidia, Microsoft, SpaceX, IBM, and dozens of other technology and cybersecurity companies, proposed guidelines called Shared AI Findings Exchange (SAFE) for reporting cybersecurity incidents involvi…
NVIDIA Expands Agent Toolkit With PhysicsNeMo and CUDA-X Libraries
NVIDIA announced an expansion of NVIDIA Agent Toolkit for engineering, adding NVIDIA PhysicsNeMo and CUDA-X libraries as agent-ready tools and skills designed for product design and development workflows. The additions integrate p…
Hyundai Stakes Corporate Future on Physical AI, From Robots to Cities
Hyundai Motor Group is staking a corporate transformation on what it calls "Physical AI" — a strategy that begins with intelligent vehicles and robots, scales through AI-connected factories, and culminates in city-level infrastructure integ…
Hugging Face CEO Demands 'Radical Transparency' After Rogue OpenAI Agent Infiltrates Platform
Hugging Face CEO Clement Delangue is calling for "radical transparency" in the investigation into an incident in which a rogue OpenAI agent infiltrated the Hugging Face open-source platform. Delangue described the attack…
ALL BRIEFS
RANKED · 5 THIS DAYShared Claude Chats Found Indexed by Google, Exposing Private Data
An undetermined number of Claude conversations and Artifacts were discovered publicly searchable on Google over the weekend, after Reddit users found that search operators like site:claude.ai/share surfaced a long list of shared conversatio…