Conversation

Notices

1. Embed this notice
   Lykso (lykso@tiny.tilde.website)'s status on Sunday, 30-Aug-2026 14:10:45 JST Lykso

   @lianna Exactly. The BSDs will likely prove to be temporary refuges. More ideologically grounded projects will likely prove to be more stalwart. E.g., OpenBSD rejects LLM-based contributions due to copyright ambiguity, a temporary situation.

   • Embed this notice
     Graham Perrin (grahamperrin@mastodon.bsd.cafe)'s status on Sunday, 30-Aug-2026 14:10:40 JST Graham Perrin

     in reply to

     • Adrian Chadd <verified.png>
     • Raster

     @raster

     In my limited experience: use of AI in connection with commits to the FreeBSD trees – direct commits (not counting AI-assisted commits to OpenZFS and other upstream projects) – has been responsible, and I do not expect this to change.

     A thought from a FreeBSD committer, in response to a comment on a pull request where code is heavily AI-assisted:

     "I think its great, you figured out a bug, you reported it in what looks like a good spot and the PR submitter integrated it.

     The only suggestion I have is to be mindful that maintainers/PR submitters may not want to read AI output in PR comments and code. I'm not saying your PR comment sounds like Claude or not; it's just something to keep in mind. I suggest that when people submit stuff they actually take the time to go write their own short summary of it and understand it enough to answer questions that the maintainer/PR contributor may have."

     👍

     <https://www.reddit.com/r/freebsd/comments/1w1auhp/comment/p6odzdc/>

     The comment in GitHub:
     <https://github.com/freebsd/drm-kmod/pull/499#issuecomment-5459886264>.

     Please note that the drm-kmod repo is separate from the src tree for FreeBSD base, which includes the kernel.

     Cc @erikarn @lykso @lianna

   • Embed this notice
     Raster (raster@retro-gaiden.com)'s status on Sunday, 30-Aug-2026 14:10:42 JST Raster

     in reply to

     • Graham Perrin

     @grahamperrin @lykso @lianna In any case, I have the impression the FreeBSD and OpenBSD project are not really keen to accept LLMs neither in long future for what concern the code.
     OpenBSD in particular has like a paranoid obsession about code control, enough they also made their own version of the X11 server, that I doubt they would change idea about LLMs even if the copyright ambiguity would be solved somehow.

     feld repeated this.
   • Embed this notice
     Graham Perrin (grahamperrin@mastodon.bsd.cafe)'s status on Sunday, 30-Aug-2026 14:10:43 JST Graham Perrin

     in reply to

     The FreeBSD Project is surveying …

     https://www.reddit.com/r/freebsd/comments/1vthk9s/comment/p4uwki6/

     #FreeBSD #AI #LLM #LLMs

     @lykso @lianna

   • Embed this notice
     Adrian Chadd <verified.png> (erikarn@mstdn.social)'s status on Sunday, 30-Aug-2026 14:11:25 JST Adrian Chadd <verified.png>

     in reply to

     • Graham Perrin
     • Raster

     @raster @grahamperrin @lykso @lianna Our job as freebsd maintainers is to hold a way way higher bar regardless of source.

   • Embed this notice
     Raster (raster@retro-gaiden.com)'s status on Sunday, 30-Aug-2026 14:11:26 JST Raster

     in reply to

     • Graham Perrin
     • Adrian Chadd <verified.png>

     @grahamperrin @erikarn @lykso @lianna My concern is more when people turn off the brain and put the machine in the place, than have a machine that gives hints where poke your fingers in what does not work.
     And usually I do not put much trust in people.

     feld repeated this.