Re: LLM Commit in upstream `master`?!

Madhu via "Emacs development discussions." <[email protected]> writes:

> * Stefan Monnier <[email protected]> :
> Wrote on Sun, 26 Jul 2026 10:59:45 -0400:
>> I know Emacs maintainers are not crazy at all, so I'll file the whole
>> event under "blunders", but I have to speak up and say this was a very
>> bad decision.  I don't use LLMs and keep praying that all the AI data
>> centers burn up in flames, but can we stop about this LLM paranoia?
>> LLMs cause enough real immediate problems that we don't need any
>> extra paranoia.
>>
>> Yuan made a mistake because he misunderstood the policy, we learn, we
>> `git revert`, and we move on.  Force pushes are not for mere worries.
>
> UDP (useless data point)
> I happened to do a pull on master yesterday before i saw sean's post
> here. I'm glad I don't have the commit in the history (which I'm
> hosting on my local disk), and I'm glad I'm not forced to keep that
> commit on my disk because of some bureaucracy.

Don't you think that Stefan has a point?  As I see it, a good deal of
inconvenience was occasioned to many people in order to satisfy the
unreasonable antipathy of some users to have evidence of a mistake which
was corrected in their computers, i.e., some persons who desire to erase
history if it contains traces of something which displeases them.

How does it follow from the existence of a log message produced by an
LLM in someone's checkout of emacs.git that all of their alleged
mischief will thus be reinforced?  Isn't this senseless posturing?

My first response to the inability to update from git.sv.gnu.org was
that the repository had been compromised by malefactors, and it is easy
to see that it could have produced significant alarm in other quarters,
among users whose suspicions were less easily assuaged than mine (or who
had fewer means of assuaging them).

Please do _NOT_ force push in future, whether to remove LLM-generated
code or otherwise.