A community for everything relating to the GNU/Linux operating system (except the memes!)
Also, check out:
Original icon base courtesy of lewing@isc.tamu.edu and The GIMP
I think not allowing it at all would be worse, because then people start claiming not to use Ai while they secretly do. Allowing it with a disclosure at least makes this process a bit more transparent. You can think about ai what you want, at least handling it this way is better than not allowing it at all.
Only reasonable take in this thread
Awh yeah because the type of person to use a chatbot is definitely the type of person to be honest and not take credit for what the chatbot outputs. No fucker is gonna disclose this unless they're trying to sell a chatbot model to a greater fool.
fedora living up to its name i see
Don't know how bad this take is, but not using LLMs for coding assistance to some degree just for the sake of not using LLMs might not be the best option right now.
There has to be a middle ground between letting the thing spit out whole kernel modules and refusing to use it at all.
Also having it declared as AI assisted code might be better than the people doing it anyway undisclosed.
The middle ground, IMO, is not letting it spit out code.
Its almost certainly terrible, every time. Sometimes though... Its just mostly bad.
Ive found it useful for finding errors and potential optimizations though. Just not, you know, letting it actually write anything.
But letting it review and seeing:
This library is currently being considered for deprecation on this mailing list, where other library is being suggested instead.
Thats useful! Helpful, even.
Just not the nonsense it makes on its own.
The middle ground, IMO, is not letting it spit out code.
Are SPEC files for RPM creation code? How much actual code is even written under the Fedora umbrella, except maintenance scripts and such? Adjacent projects such as Anaconda are in the rhinstaller organization on Github: https://github.com/rhinstaller/anaconda
Either I overlooked the details or they aren't spelled out. From my experience of packaging software for myself as RPM (for openSUSE) the amount of actual code are a few lines of bash scripting to invoke sed and such.
@passepartout
Perhaps if AI were explicitly asked to write code in pidgin English (specifically something which is not meant to be compilable), and then that was used as the basis of actual code development?
@cm0002
bold move. Can't wait to see how this gets reversed within a year
!Remindme 6 months
FYI PieFed has a reminder function, in the 3 dot menu on a post / comment.
I literally can't keep up with how fast PieFed is improving You and the team are seriously impressive
Thanks!
Unfortunate. Announcing this when everyone is migrating from windows... a lot of which is because of the AI bullshit in it... seems like throwing the stick in the spokes move.
I guess it's time to go shopping for a new distro :(
This is about contributing code that was co-created with an llm like copilot. Not about adding "AI" features to fedora.
that's a fair point, though at least with projects like fedora, commits are public and can be broadly scrutinised, and since this stipulates that the use of LLMs must be disclosed, I'd hope that'd keep its use firmly under the spotlight.
Well sure... the point of a warning label on the side of a product is to allow you to make informed choices if you want to use that product. A warning label saying "we're condoning the use of unethical practices" allows me to decide I would rather seek a different product.
another fair point, by allowing it at all they're condining its use. i personally see it less as condoning and more that they acknowledge its wide use in the field, and that they probably cannot prevent its contributors from using it entirely.
I'd be interested in how many commits come in from now suggesting the use of llms.
Well that’s what “condone” means fwiw.
I see it the same as how steam does the same thing with requiring disclosure of ai use on store pages now. And I treat it the same way if I see it I make a consumer choice to not support the game.
What disturbed me was reading the minutes of the meeting that people seemed genuinely excited to include gen ai code. To me that speaks to an ethos that is highly divergent from what I would like to see and what should be happening. It doesn’t feel like it’s “welp I guess we gotta let people do it” and more “oh boy we can finally use it”. And with all the companies that make llm how long before some back door evil nonsense sneaks in. To say I’m dubious would be an understatement. 🤷
That's understandable, I'm inclined to react the same way to games on steam.
I hadn't checked in with how that discussion went down; that's disappointing. I just settled into fedora after fully moving on from windows earlier in this year (have been multi-booting for several years prior). Time will tell if this all goes to shit as well.
I guess it’s time to go shopping for a new distro :(
If you think that undisclosed AI contributions aren't happening everywhere, you're delusional.
Doesn't mean I have to support the ones that actively encourage it.
I guess it's time for some of the projects to start putting little stickers that say "hand crafted code" on them explicitly.
Properly attributed generated lines are easier to remove should courts declare them illegal.
What would projects with undeclared AI code do? Shut everything down? Revert everything until the commit before ChatGPT launched? Just say yolo and go on?
Counterpoint... there is no real enforcement beyond the honor system... so it changes very little other than expressly condoning the activity.
No thanks, good luck ,iam gone
But you were there right?
Much of distribution development is writing menial scripts and SPEC files. It's tedious work with little creativity. The last SPEC file for an RPM package I wrote from scratch was years ago but it was so tedious work. The Arch maintainers even argue that their PKGBUILD files are so simple, they don't pass the so-called threshold of originality and therefore are public domain anyway.
Much can be (and probably already is) automated. Compilation directives like CMake files already contain all the info needed to generate a workable if a bit bare bones SPEC file. I'd say an LLM might even be overkill for what a script could also achieve. The result is public domain anyway.
Whelp, it has been nice Fedora.
I honestly don't see the use case