On Fri, 09 Oct 2026 09:20:18 -0600 Jonathan Corbet <[email protected]> wrote:
> Krzysztof Kozlowski <[email protected]> writes: > > > On 09/10/2026 10:31, Sasha Levin wrote: > >> The Maintainers Summit reached consensus to deprecate the Assisted-by > >> tag. Remove the documentation requiring it. > > > > Please provide here some sort of arguments used in that summit > > consensus/discussion. This basically reverts/changes previous consensus > > [1], which is fine itself, but it lacks reasoning. > > > > [1] > > https://lore.kernel.org/workflows/2026070224-unholy-commode-cf45@gregkh/ > > I'm working on a report on the discussion, don't worry...Monday or > Tuesday probably. But changes like this should not rely on LWN articles. They need to be presented in mailing list. Honestly, I would like to have a "Vibe-coded-with:" tag. Why?, when I do a git blame and want to ask the author why they did the change the way they did, I would likely not expect much if they vibe coded it. When we first started talking about what constitutes an Assisted-by tag (using AI to debug, or review, or simply add templates), there wasn't really any agreement to where it should stop. I have always been of the idea that I want to know if someone actually wrote the code. If all you did was use prompts, then you didn't write it. You simply instructed a computer what to do. Much like helping a junior engineer. But with a junior engineer, they would have the authorship and signed-off-by and the kernel developer would likely just get "suggested-by". But we don't do that for computers. But still it should be tagged that the author didn't write the code somehow. I wasn't at the maintainer summit to express my opinion so I'm doing it here. -- Steve
