Re: Linking Patchwork with Sashiko?

[Laurent Pinchart <laurent.pinchart@ideasonboard.com>]

On Thu, Jul 16, 2026 at 05:10:11PM +0200, August Zeller wrote:
> The frustrating part of this situation is how everyone arguing on both
> sides is right in some way.
> 
> 1. The tools work, yes. This used to be up for debate until recently,
> but now it's clear that AI can enhance contributor *and* maintainer
> productivity in the kernel if used correctly.
> 
> 2. On the other hand, it is also true that LLMs and derived tools are a
> comprehensive manifestation of the problems of the industry and there
> are valid ethical concerns regarding access discrimination, copyright,
> workload externalization, job loss, etc.
> 
> 3. However, a "no AI used" policy is unenforcible. Many people would
> use these tools regardless without announcing it, which is the worse
> outcome. If we're transparent about our AI usage, then at least
> everyone is aware and can treat the affected code with the right level
> of scrutiny.
> 
> 4. Moreover, a "no AI used" policy isn't the best thing to do
> technically, as AI has proven to be very useful to catch bugs, review
> code, and even write some new code. Some critical bugs that we could've
> caught with AI *will* slip under the radar if we refuse to use those
> tools, while bad actors will gladly use the newest models to find
> exploits.

I don't think there's much disagreement on all that. I'm not aware of
anyone trying to get the kernel to adopt a "no AI" policy.

> Overall, even though I've been known as a great AI skeptic in the past,
> I'm with Linus on this issue. *Accepting and embracing AI may not be a
> desirable thing to do, but the alternative is even worse.*

-- 
Regards,

Laurent Pinchart