Anthropic's latest preview model, Claude Mythos, has identified thousands of previously unknown zero-day security vulnerabilities across all major operating systems and web browsers. The discovery was significant enough to trigger an emergency meeting between Federal Reserve Chair Jerome Powell, Treasury Secretary Scott Bessent, and the CEOs of major U.S. banks to discuss potential systemic risks to the financial infrastructure. Anthropic has initiated a controlled rollout called 'Project Glasswing' to allow defenders a six-to-twelve month head start on patching before the capabilities become widely accessible.
A major industry shift is underway as top AI labs move beyond providing APIs to becoming direct IT service competitors. This week, OpenAI launched 'The Deployment Company' and Anthropic announced a $1.5 billion Wall Street joint venture. These entities act as AI-native SaaS firms, directly competing with traditional IT giants like TCS and Infosys by offering end-to-end business process automation and 'agentic' workflows.
A $20 million early investment in OpenAI by the University of Michigan, made before the release of ChatGPT, is now valued at approximately $2 billion according to court documents revealed this week.
What is the name of the new 'opt-in' safety feature launched by OpenAI for ChatGPT to prevent real-world harm?