anthropic
CISA Deploys Anthropic’s Mythos AI to Hunt Vulnerabilities in U.S. Government Code
The U.S. Cybersecurity and Infrastructure Security Agency (CISA) is reportedly using Anthropic's advanced AI model, Mythos, to proactively scan government code for security vulnerabilities. The goal is to identify and fix flaws before malicious actors, such as foreign intelligence agencies or cybercriminals, can exploit them. Initial audits using the AI have allegedly uncovered a significant number of bugs, though details on the scope and severity remain undisclosed.
Claude Code’s hidden tracker was an “experiment,” says Anthropic
A researcher discovered a hidden tracking mechanism within Anthropic's Claude Code client that encoded user traffic data based on system time zones. Anthropic has since removed the feature, stating it was an experimental measure to prevent account abuse and protect against model distillation, potentially linked to recent US government export controls.