On Fri, 19 Jun 2026, Jim Hall via Freedos-devel wrote:
I think "take responsibility for the [AI-written] code" is a hard hill to climb. Even with the "No code gets accepted unless it is reviewed by a human" statement, you still don't know the origin of the AI-generated code. (i.e. what it was trained on, and if AI reproduced some code it was trained on.) As an example: Way back at the beginning of this thread, I mentioned that I've seen AI "vibe coding" regurgitate code I wrote elsewhere. (I recognized my own code from an article I'd written.)
This is the same thing I was referring to when I said "how do I know the LLM I used didn't plagiarize GCC to develop my C compiler?"
<snip>
I've always been very careful about contributions to FreeDOS that could be "tainted." (See "OpenDOS" and "Computer History Museum MS-DOS" from many years ago, for example.) But I only care about the core parts of FreeDOS. If a developer uses AI to work on a "third party" DOS tool that we happen to include, that's up to the developer -- as long as it's not included in a "plain DOS" install. Games, apps, editors, .. etc .. are outside the "no AI" policy.
I probably sound a bit weird here as someone who _is_ developing something outside of my specified "absolute zero" idea but I would argue even the third-party stuff should have a rule similar to...I think it's NetBSD: "if it's got AI, it's automatically considered copyright-tainted and refused."
I mentioned the C compiler I'm working on; it would be an absolute boon to the FreeDOS Project to have a lighter C99 compiler (and it might even end up being able to self-host) BUT:
1. Unless the compiler proper were rewritten from the ground up, there is _no way_ it can be included morally in FreeDOS, because there is no way to know the provenance of the code. This, irrespective of it being clanked.
2. The libc DOES have a known provenance, because I gave a directive to the LLM to specifically draw from a trusted body of code i.e. three versions of the BSD libc - so the copyright issue isn't there. But it's still clanked out (although not in a way that, apart from the startup code, I can't do anything about - and maybe MSC51 can be used for that). It could probably be argued that since any actual clanked-out code is either properly attributed BSD code or very obvious stuff that effectively translates RBIL into C. At least this is a more manageable problem.
Now, I'd be absolutely OK with someone taking and "de-clanking" the code once it's out there (i.e., completely rewriting the compiler; I'll probably rewrite significant chunks of the libc myself, which I at least can understand). Then maybe it would be actually useful.
-uso. _______________________________________________ Freedos-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/freedos-devel
