Date: Sat, 21 Mar 2026 11:59:02 -0600
From: Theo de Raadt <[email protected]>
To: Renaud Allard <[email protected]>
cc: [email protected]
Subject: Re: [patch] ext4fs rw
In-reply-to: <[email protected]>

Renaud Allard <[email protected]> wrote:

> Maybe it should be made clear on the website that OpenBSD will only
> allow new code made by a human. Because I feel there might be more
> requests like this and there is no point in repeating the discussion.

Yes.


#openbsd
The context is NOT accepting code that has been output by an LLM.

So that is pretty close to a stance on OpenBSD NOT accepting LLM code.
(that "yes" is from the main guy)
Not that I had doubts :P
I just have seen people on here question it.
more:


Date: Sat, 21 Mar 2026 12:07:32 -0600
From: Theo de Raadt <[email protected]>
cc: Renaud Allard <[email protected]>, [email protected]
Subject: Re: [patch] ext4fs rw
In-reply-to: <[email protected]>

Theo de Raadt <[email protected]> wrote:

> Renaud Allard <[email protected]> wrote:
>
> > Maybe it should be made clear on the website that OpenBSD will only
> > allow new code made by a human. Because I feel there might be more
> > requests like this and there is no point in repeating the discussion.
>
> Yes.

If we point out that Copyright assumes human/artistic/creator output,
and that we depend upon various rights being specifically and legally
granted by the human/artistic/creator, then we barely need to say that
AI are not humans, and definately not able to legally grant the rights.

@pkw that's super succinct and clear... and hard to reasonably argue with. I'm going to try and copy that explanation when telling people I won't accept there LLM patches. I think this is the angle to lean into. Just tell them, it could endanger the copyright of the project, depending if the law keeps trending in the direction it has been with LLM copyright. So, it would be irresponsible for the future of the project to accept their patch, knowing it's LLM generated.