Just for clarity, there's a lot to critique about RMS and trust me I have critiqued RMS *plenty,* but I see some threads here seeing "emacs maintainer" which are coming from Eli and assuming they came from RMS. I am NOT defending RMS, just trying to make some emacs-devel drama clear
RMS hasn't seemed to have fully developed his stance on LLM things right now but is working it through. For example, I wrote this email:
https://lists.nongnu.org/archive/html/emacs-devel/2026-11/msg00192.html
RMS replied with:
https://lists.nongnu.org/archive/html/emacs-devel/2026-11/msg00271.html
And RMS at points called LLMs "bullshit generators"
https://lists.gnu.org/archive/html/emacs-devel/2025-03/msg01107.html
I know plenty of people on here have no inclination to take RMS seriously, but while Eli has been pretty frustratingly on the "if we're allowed to include LLM output in emacs let's do it", RMS seems to be trying to work through the issues, and hasn't come to conclusions yet, but there are emails like the above that mean that it might be worth distinguishing between the different players involved.
@cwebber@social.coop I feel like your reading of Armin's blog post is a lot more positive than mine; I read it as a cautionary tale of how this new technology is fully capable of bringing a codebase into the state of being commonly known as Legacy Code at a hitherto unseen pace, and how that state can arrive unnoticed because the costs are initially borne only by the hands of machines, not humans.