Sorry Eli, it seems you never got the point, and you keep asking what is
the purpose.
I really do not doubt in none of developers as in their characters.
You asked what is it about, so...
- A commit was made that appears to be LLM-generated (beyond just the
commit message)
- The author self-certified it was not LLM-generated (but proprietary
Claude submitted the commit)
- This self-certification was accepted without any way to verify it
This creates a situation where the policy (prohibiting LLM-generated
code) is unenforceable because there is no verification mechanism.
Then you Eli then responds by questioning mu character and purpose,
while ignoring the procedural problem you raised yourself and keep
managing it.
If you can't see there governance gap, I can't help.
I never said Yuan is not trustworthy or you. I said there is nothing you
can do to verify the claim, yet indication indicates it was proprietary
code that generated it.
And you tried hard-removing the commit from the git.
So you are trying to protect Emacs as software from the LLM generation
with the allegedly conflicting copyrights, and you miss to see that
there is vulnerable position of the project, to which I am pointing.
There is no boundary, or mean to control it. You got the honor system,
but in GNU project I simply can't believe in that honor system of
self-certification in that case where proprietary LLM Claude submitted
the commit message to Emacs.
Is it sufficient then to accept the honor word? The case itself shows it
is not.
Let us say that Yuan or whoever else, submits commit with the
proprietary Claude, but code itself is written by hand. That is not
logical. He could submit the code as well by hand, it is few keys to do.
It is obvious that it is not logical, and so people will not believe it.
Put few more cases like that, LLM submitted commit messages, and who
will believe it then?
So question is right, not attack on you or anybody, but question of
governance.
Obviously you can't see what I am telling you, yet you should be first
to understand it.
Address your procedural flow, not myself writing to you about it. I am
messenger. It is now for you to actually understand it, and not just be
defensive.
Jean Louis
On 2026-07-28 13:59, Eli Zaretskii wrote:
Date: Tue, 28 Jul 2026 00:38:08 +0300
From: Jean Louis <[email protected]>
Cc: [email protected], [email protected], [email protected],
[email protected]
My e-mail was not about the GNU policy. I also did not address it to
you
Eli. In fact I was very clear that I support whatever decision will
be
made by GNU.
I am sorry to hurt your feelings, yet it was you who accepted
obviously
something so suspicious that is contrary to whatever you are trying to
impose there. If you wish to handle things privately, then you could
do
so with the developer right?
My e-mail was not about GNU policy, but about the obvious lack of
integrity on the list: somebody submitted the LLM generated patch, and
then self-certified it, and you accepted it. And it is obviously not
possible to verify that self-certification.
No matter how much you may try to squeeze me to put on "moderate", it
doesn't change the fact that it was weird. So moderate as much as you
wish. God thanks everyone can have their own website and exercise
speech
on other places.
What was the purpose of that message (and this one)?
You point to various people by their names, you hint that they might
lie or should not be trusted, you say that there's something "weird"
going on in Emacs maintenance. What's the purpose of this? And why
do you think you have a right to post such ad-hominem attacks publicly
in the first place?
Yuan is a valuable member of the Emacs development team. He developed
one of the most important features Emacs gained during the last years,
and continues to develop and maintain it ever since. What reasons can
we possibly have not to trust him when he says that only the commit
log message was produced with LLM's help? Especially since Yuan is
not a native English speaker, so his desire to have help in expressing
himself in English can be easily understood. Why should we trust him
less than you or anyone else here?
If you contend that we should build this community on mistrust, then I
must reject that. IME, communities based on mistrust and suspicion
don't hold long, they collapse and disappear. So at least as long as
I'm one of the maintainers, we will trust our team members that they
answer our questions truthfully, as long as we don't have good reasons
to think otherwise. Not the other way around.
In any case, discussing the trustworthiness of people is not
appropriate here (or on any Emacs-related list), and will not be
tolerated.
---
via emacs-tangents mailing list
(https://lists.gnu.org/mailman/listinfo/emacs-tangents)
---
via emacs-tangents mailing list
(https://lists.gnu.org/mailman/listinfo/emacs-tangents)
