M@ Bntn đ˝
厎ĺ塴ć Kritaâs Maintainer is awesome!
zivi@KitsuneofInari âas long as the contributor understands the codeâ understands is subjective, and theyâre guaranteed to understand it less, so really its just apologism for converting codebases into unmaintainable dumpster fires
@KitsuneofInari based maintainer we need more takes like that in the space ^^
đ¸ lumi-nhac 
đą
â
Tock@KitsuneofInari I love you Krita! And all the developers, too!
Been backing Affinity for the longest time, since nobody has to wonder on Adobe's position on AI ("we're already vibing our source code, use Firefly!") Affinity 3.0 was the biggest disappointment.
Thank you for this.
catraxx 
@KitsuneofInari Fuck yeah!
Anna Miriamsdochter@KitsuneofInari @halla â¤ď¸đ
RoseForge 
Li ~ Crystal System@KitsuneofInari im not sure id say that understanding of code is a boolean function of just either fully understanding everything including any and all bugs that could ever arise, and not understanding it at all, but rather a variable level of some understanding,
Halla Rempt@Li @KitsuneofInari So... Nobody can really understand code fully, not even the code they have written. Code has become too complex, and there is just too much of it!
But if you have thought about what your were writing, if you have recorded your thinking in commit messages and comments.
There is a chance you might have a recollection of having created that code when debugging it, ten years later.
If you asked Claude to regexp-slop it for you, not chance.
@Li @KitsuneofInari Also, the fun part of writing software is thinking, coding, testing, seeing people use it...
Doing code-review, not so much. That's what you do to help other people to level up.
But with LLM-generated code, where are the people you want to mentor?
But you still have to code-review the swill.
o lĐ°vrĐžvsky@halla @Li @KitsuneofInari This is tantamount to saying donât use tools, or libraries, or graphics cards, if you donât review their code. I canât agree with this sentiment. The real issue is the broken trust and overreach of agentic tools.
Should I ditch the Wacom and pick up a papyrus roll and feather quill instead because I donât review the Krita source code or my tabletâs circuits? No, I trust the authors and their community.
So the question in my mind is: in what skilled hands, with which LLM, and to what extent, does it fall within the boundaries of acceptable use?
@halla @Li @KitsuneofInari I completely respect the decision of the maintainers, and think itâs better to have a clear policy than let the debate stew and simmer. Given the user base and sentiment against the GenAI topic, itâs possibly even the right move.
I just want to see an even better justification, like: we donât add code, we tactically remove it. Our code base is a sculpture, and LLMs arenât much help here.
Or even: as you kind of wrote above, Krita is an expression of the joy of art through the joy of coding. We arenât feeling the vibes when you PR us your lobster soup, but feel free to make your own painting program from scratch. Call it Pincha, if you want!
M.
NullDott đŠ (Comms Open 3/4)@loleg I don't think is a tantamount at all.
LLMs aren't a vital part of the development process, and they have never been... You can pretty much get all the supposed "benefits" from "agentic tools" without actually using "agentic tools", and you avoid all the hassle that they can bring in the long run.
Also the examples you gave aren't really equivalent?
I can trust Krita maintainers and community, I cannot trust slop machines owned by private companies.
Moses Izumi
ck@halla @Li @KitsuneofInari Anyone who has ever tried knows that if you try to learn math by just reading a bunch of chapters in the text book, you might get a passable C on the test the next day, but you'll flunk any attempt as little as a week later.
As a human, you need to do the work.
As a human, you need to do the work.
Urwumpe@Li @KitsuneofInari I think the main issue is not the code you design and write, but that its interacting with code and users you don't fully know. Often requirements or the assumptions behind the code you change or the code you interact with are not documented. Often users don't tell you what they really want. Or don't know that a small action they don't tell you in the error report is the key to understand the bug.
Good software developers are not nerds or code monkeys, but good mentalists.
shared public horse .not bitey@KitsuneofInari I checked and here's the post âĽď¸ @halla
Policy on LLM code?
Speaking as the Krita maintainer, I donât want us to accept any LLM-generated code into Krita. Some may say, as long as the contributor understands the code itâs fine, but thatâs a fallacy. No programmer really fully understands the code they are working with (otherwise bugzilla would be empty); but with LLM-generated code they havenât even put any thought into it. That code will be unmaintainable. And thatâs even completely apart from the fact that using LLMâs to generate code lowers your co...
lps
Neville Park
ratfactor
Metin Seven đ¨@KitsuneofInari Nicely worded, @halla ! đ
mohs@KitsuneofInari thx Krita!
I think many fall for a misconception here.
The code good developers write is a result of their understanding of the code that already exists and the way they intend to improve it.
If that code is generated, it is no longer a testament of somebody understanding what is going on. It becomes a "this feels like it does what I want".
JesseBot@KitsuneofInari this is lovely to hear 
Xoa Gray@KitsuneofInari @gramarye I so wish that more people felt like this.
Hacker 
@KitsuneofInari Gedit developers too!
AzarilhâĽ@KitsuneofInari Thanks for making me love Krita more.
GezThePez@KitsuneofInari is it not just another layer of abstraction?
Noa@GezThePez It is not, no.
PÄteris KriĹĄjÄnis@KitsuneofInari preach it, well said.
David Fleetwood - RG Admin
Ulysses Almeida Neto@KitsuneofInari how do they detect LLM generated code?
kie/qui (planar) đŚ˝đš HLZ@ulyssesalmeida by its character
bluestarultor@ulyssesalmeida @KitsuneofInari Aside from them all slapping their branding on it? Poor code quality. LLMs don't know how to call existing functions; they just write new code with some stab at the same functionality. Heck, they're finding Claude Code is vibe-coded by Claude itself and it's so incapable of calling functions that it's literally re-prompting itself instead.
If that's any indication, detecting generated code is as simple as a gut check. If your guts are wringing your lunch back out, it's generated. XD
@bluestarultor @KitsuneofInari fair point. So, in the end of the day, they will accept good quality code and reject poor quality code.
@ulyssesalmeida @KitsuneofInari I mean one would also hope someone has morals to disclose it if it's for some reason not immediately caught, but then Claude Code also has a special stealth mode only accessible to Anthropic employees, so probably not.
Ultimately, "no" means "no."
And frankly anyone violating that consent can take a big leap on a balance beam and land on both sides.
Speed demon đŞđş đłđ´đşđŚđľđ¸@ulyssesalmeida You can't, not always. Dishonest dev's will exist. How many thieves donate to charity? Probably some. How will we know? We dont. @KitsuneofInari
@KitsuneofInari Thatâs a really eloquent way to describe my issue with âas long as you understand the codeâ argument.
That never mattered.
I wonât merge code written entirely by human just because they understand it. Because what matters is whether after an isekai accident someone new will be able to work with it.
Flux@KitsuneofInari KDE developers try not be based challenge, literally impossible.
happyborg@KitsuneofInari ...and removes copyright from the rest of the code unless it is explicitly marked as LLM generated.
snowyfox@KitsuneofInari I mean, those are pretty strong reasons
noodle@KitsuneofInari
"No programmer really fully understands the code they are working with".
put that next to
"You can and must understand computers NOW"
"No programmer really fully understands the code they are working with".
put that next to
"You can and must understand computers NOW"
wrath0110@KitsuneofInari I totally agree! Hold the line!
Cress@KitsuneofInari You are sooo right about cognitive decline. Its a worry that we may forget (qualiity) critical analysis, root cause and a million different ways to solve blockages and yet to make space for all wildlife!
Jigme Datse@KitsuneofInari Inkscape has said similar...
Xarizzar Phantasma (ç)@KitsuneofInari There is such a thing as making mistakes because of carelessness, and it is also one reason why bugzilla is not empty, aside from not understanding the code. That said, this is a nitpick and I pretty much agree with everything they said.
JauneBaguette@yunohost De fait, on doit s'attendre Ă ce que des logiciels comme Nextcloud soient retirĂŠs du catalogue ? Et d'autres comme LocalAI ?
mikeTesteLinuxQlub@JauneBaguette @yunohost Je crois que c'est plus une politique de dÊveloppement qu'autre chose. Chaque projet le gère comme il le veut et on le voit actuellement dans l'open source, certain acceptent sous certaines condition, d,autres le rejette totalement
Mais le dude de Krita a tout de mĂŞme un excellent point, ca ramollis le cerveau ces LLM.
@mikeTesteLinuxQlub @yunohost J'entends, mais si le compte Yunohost repouet ça est-ce qu'on doit le prendre comme une prise de position officielle ou bien ?
@JauneBaguette @yunohost Excellente question. SI c'est un des principaux maintainer du projet, je crois qu'on peu prendre cela comme une position officielle, mais ça n'engage que moi đ
tinsukE@KitsuneofInari @halla it is heartwarming to see such integrity from great FOSS projects.
Keep up the awesome work!
otto@openbsd@KitsuneofInari I can confirm. We were co-workers some 25 years back...
Dimitar Stoev@KitsuneofInari This is great. I fully respect it!
Ash@KitsuneofInari Not sure if thatâs on purpose so just in case: the last paragraph of the altâtext is not visible on the image?