Linus Lays down the Law

The Linux kernel development community has finalized its policy on LLM usage after a 6-month debate sparked by an undisclosed AI-assisted patch from Sasha Levin. The controversy began when developers discovered a merged patch had been AI-generated without disclosure, causing a performance regression due to a missed 'read_mostly' annotation. The resulting policy requires contributors to disclose AI assistance, prohibits AI agents from adding 'Signed-off-by' tags (only 'Assisted-by' is allowed), and mandates human accountability for all merged code. The broader debate touches on copyright concerns with LLM-generated code, whether knowing code is AI-generated should change review rigor, and the long-term risk that AI tools will flood maintainers with patches while reducing the pipeline of junior developers who learn by writing code themselves.