OpenAI publishes Linux kernel vulnerability research, confirms 948 privilege-escalation flaws

OpenAI published Linux kernel vulnerability research, confirming 948 unique local privilege-escalation flaws and generating candidate patches; the run dynamically reproduced 1,724 vulnerabilities. Kimi K3.1 Pro released open model weights with support for a million-token context. In US elections, several AI and data center restriction measures won voter support.