Google Engineers Launch Sashiko — Agentic AI Code Review for the Linux Kernel
Phoronix·medium signal
Google engineers open-sourced Sashiko, an agentic AI system designed to review Linux kernel patches, targeting safety-critical infrastructure code that cannot afford AI-generated slop. The system earned 89pts/43cmts on HN, with discussion focused on whether AI agents can be trusted at kernel level — and whether the bar for automated review of security-critical open-source code is fundamentally different from application code.