On Get Involved page: https://krita.org/en/get-involved/ The link to Contact page is broken because it links to: https://krita.org/contact/ It should instead be https://krita.org/en/contact/
Sorry about focusing too much on my personal opinions on AIGC/LLM in general in my previous comments.
In this comment, I would like to comment on the CEHA itself. I hope this will make it clear why I don't like the current CEHA.
I understand you want to stand behind human artists by this CEHA, and I appreciate that. But it appears that the drive behind this Commitment is emotional rather than rational.
Like many others have already stated, I mainly took issues on the following two segments:
...rejecting any movement of the project towards technology that can be reasonably considered exploitative of unlicensed copyrighted works.
Who is considering? Why do these people have the right to judge, but not the others? By what standard? On what ethical/legal basis?
...to empower all people to learn, improve, express and work as artists.
You are not "empowering" anybody by putting obstacles on implementing new, useful tools. As I stated before -- do you consider helping less talented people to learn, improve, express themselves "empowering"?
Copyright (Legal)
You can't use Copyright for your legal argument. AI is not redistributing copies of the original artworks. You can't win a case with this argument, which makes it illegitimate.
If you use copyright as a reason to stop AIGC/LLM, you will be forced to also apply that standard to every human artist. You will end up with a world where human artists can't even reference each other's work. That benefits nobody, except the sorts of Disney.
Free Software (Philosophy)
You may say the CEHA is not "legally" in the conflict with FOSS or GPL. However:
Free Software is about the FREE distribution of knowledge and intellectual properties. GPL stands for General PUBLIC License. Creating obstacles to prevent usage/access/distribution of knowledge does not align with the spirit of our root philosophy.
Easily Challenged Ethical Basis
As how I demonstrated in my first comment, your ethical ground can be easily challenged from the angles of free culture, equal rights, public ownership of intellectual properties, anti-elitism sentiments and so on.
Luddites
The current Anti-AI discussion reminds me of the Luddites https://en.wikipedia.org/wiki/Luddite , a 19th-century movement of English textile workers which opposed the use of certain types of cost-saving machinery, and often destroyed the machines in clandestine raids.
Needless to say, they failed. Their actions were illegal, their cause did not align with the fundamental interests of the broad consumers who wanted cheaper goods -- despite well-meaning people might have rallied behind them.
The 19th century workers suffered not because of the machines. They suffered because of the unequal distribution of technology and wealth. They were barking on the wrong tree.
I can't help but seeing the similarities between the history and what is happening today.
A movement without legitimate means, without defendable self-justification, without tangible interests to connect with the broad population base, in my opinion, is destined to fail.
Public Perception
In my opinion, adopting this kind of Commitment is the equivalent of cutting your own arms off in exchange for some one-off applause. After the stunt, the crowd scatters and returns to their usual life -- while you, will have to live the rest of your life with no arms.
If we adopt this Commitment, do you expect people to change their art app to Krita immediately? I don't think so.
Artists are not going to support Krita in any meaningful way just because of some empty Anti-AI gesture. All we get at the end will be the same empty gesture -- a few hundred Likes and Reposts on social media, and that'd be about it.
The kind of Public Perception we really need, is "holy shit, Krita is such a handy tool to draw, look at this amazing demo!". Practicality over Politics.
Talking is cheap
I am an idealist in spirit, but I am also realist in action.
Can we really make an educated decision by listening to merely what people say? How often do we see people pretending to be concerned about ethics, but turned out to be the worst offenders themselves?
Do you remember the pure evil invention named "loot-boxes"? People used to be noisy about them, but they have become a norm these days. Look at how popular Genshin Impact is these days, a f-king "GachaGame" of all things. It is a walking proof how much people actually care about "being ethical" when their happiness and payment are on the line.
Sooner or later, people will forget about the AIGC/LLM discussion. Suddenly you'd find yourself living in a world where AIGC/LLM is a norm, or worst, as a requirement to stay relevant as an art tool. Artists are selling properly licensed LoRA models of their characters and art styles. The audience are going to town with these toys. Your users are nagging you about the lacking of AIGC/LLM infrastructure -- which BTW, every other competitor has fully established years before.
What are you going to do now? Are you going to walk back on your Commitment? How's that for your "Public Perception"?
AIGC/LLM as a threat to artists
As mentioned by Tymon Dąbrowski in the maillist, AIGC/LLM can custom-generate appealing art. Because of that, it is considered to be a threat to artists' livelihood. As an artist myself, I also felt deeply threatened by it. No deny about that.
However, creating obstacles to stop your competition is a foolish attempt that will never success. As artists, we can only improve ourselves to stay competitive.
For example:
Not only that, as @davidrevoy already pointed out, AIGC/LLM can't help but generates "same-looking things". The results have no real substance. They are either useful enough as art references, or good enough to serve as cheap eye-candies/pornography. Their novelty is quickly running out.
As such, I really don't see how human artists in general will be harmed by AIGC/LLM in the long run. If any, it urges us to up the game and become better artists.
As for the dwindling income -- maybe the deteriorating world economy is the real culprit. To fix that, we need to look elsewhere.
Can't agree more on all of your points here!
The repetitiveness of the AIGC/LLM results. They can be very appealing sometimes, but they do FEEL curiously THE SAME if you generate enough of them. I can kinda understand the cause of this -- if everyone wanted "the best masterpiece quality", it will certainly lead to the most popular aesthetics.
Text prompts are often useless to get what you want -- unless the model was prepared with all the tags you need.
Those clumsy controlnet tools...I sometimes think I might be better off just draw a sketch myself, lol. Or maybe one day AIGC Operator will become a new genre of tech artists?
I laughed so hard about the "Cloud of Rocks". That's the kind of creativity you don't often see in human artists XD! Jokes aside, I was immensely impressed by a Krita ComfyUI demo where they change the pose of a skeleton and gets a renewed reference layer immediately. The result looks good enough for a proof-of-concept. I can totally see why corporate artists are using this.
Judging by all the demo videos I watched, I consider AIGC/LLMs to be useful tools. They can really enhance the productivity of an artist. Just not very fun to use, and they are very limiting.
Although I personally would love AIGC/LLM technology to evolve further, I just can't shake off the fear that our AIGC/LLM technology is reaching to its limit.
Regardless of how people advertise their LLM models, under the hood, they are just different instances of the same Transformer technology, maybe with more sophisticated peripheral tools. They are not really "AI" after all. They do not "understand" what the input means, nor what the result means. They have no logic capabilities. They will never truly replace human artist without further breakthroughs.
I too, have concerns for new artists after AIGC/LLM. I fear the "shortcut/cheat" provided by AIGC/LLM will destroy many new comers' initiative to learn how to make art from scratch.
But on that note, isn't Digital art itself already a shortcut compared to traditional art? Will the new generation find their way, just like what we have found ours? I want to remain open-minded about that. After all, the true power of an artist is observation and taste.
I still don't use AIGC/LLMs in my own workflow because of the shortcomings you mentioned. I can buy a better computer to run it, but it's not that helpful for artists like me. The ethical issues associated with it can't be overlooked, as well.
But just as you said, if the models are ethic, the controls are more effective, the results are nice, and we can run this on integrated graphics -- I will use it in a heart-beat (we can dream) XD.
I would like to remind about the risk of living in an information cocoon.
Those who are vocal in front of you don't represent the silent majority. Those who have a fulltime art job don't usually have the time for politics. Their concern lies mostly on how to make their art better.
Not only that, because of the "AI witch-hunt" among the art community, many artists are fearful to state their voice in favor of AIGC/LLM.
In fact, I'm taking a huge risk by making my previous comment here. I did it because Krita means a lot to me. I want to make sure you made the decision without being misinformed.
Contrary to your observation, it is not "only the tech bros" who are excited about AIGC/LLM arts. Almost every artist I know are experimenting with AIGC/LLM. Those aren't just "amateur artists" either -- those are professional artists with notable skills, working in the industry on big titles. Concept artists in AAA studios have long been integrated AIGC in their quick concept iteration pipelines. They just don't talk about it because of the stain associated with it.
As such, I don't think it is wise to use the "tech bros" label so broad-strikingly on people.
Krita's ComfyUI plugin is unbelievably popular. Videos about Krita+ComfyUI have 10 times, 100 times more views and discussions than other videos. I watched people generating designs and poses on the go, and I totally got the appeal -- it is much faster to check whether a composition/design works without pouring in hours of effort only to see the result unused.
I even came across a very popular social media post saying, "The primary reason to use Krita these days is how easy it is to integrate with ComfyUI/SD. Ironically, the devs seemed to be anti-AI.". This post got liked and reposted like there is no tomorrow.
Too bad my PC isn't powerful enough to run AIGC/LLM (R7840U/RX7600) XD. So you might see this hardware requirement to be another issue of inequality :P.
What is the actual goal of this Commitment? Is this Commitment exists to say "Krita will never work on anything to help people use AIGC/LLM in Krita"? If that is the case, I have some thoughts I would like to share.
Who does Krita serve
In order to serve the artists better, Krita should commit to help artists to create better art.
What artists really need
Artists usually do arts for these reasons:
Since Krita labels itself as a professional art tool, it should help artists to get paid.
Why do professional artists struggle
As an art tool with many substitutions, it is not Krita's job to change the society's distribution issues, not does it have the power to do so.
What can this Commitment do, really
Can this Commitment help the artists pay the rent?
Or is Commitment, in the long run, works against the struggling artists to have better tools to stay competitive?
Ethics
I have a difficult time finding a clear stance around the ethics of copyright and AIGC/LLM. A point often contradicts with itself, even inside the same political spectrum.
Copyright did not exist until the Licensing of the Press Act 1662. It is not a "birth right" as people might believe today. It exists to protect the capital-driven publishers under the name of "fostering creativity".
However, before copyright, knowledge was freely distributed. Artists studied and even copied from each other's work. It was known as the Renaissance, an era bursting with creativity.
No creativity exists without basing itself on existing works. To grant someone "all rights" over their works -- as if it was created from vacuum -- just feels wrong to me.
We love to criticize Disney for holding on their copyright so dearly. We often use the argument of "free culture is good for creativity". But are we willing to apply the same standard to indie artists? Why the double standards?
If it is acceptable for human artists to learn from each other, on what base do we consider AIGC/LLM learning from the works of human to be unacceptable?
If protecting the rights of human artists is "Empowering Human Arists". Do you consider using new technology to help less talented, less trained people to create art "Empowering Human Artists"?
What is our philosophy, really?
Is our philosophy about "Free Culture" -- to encourage the liberal distribution of art and knowledge? Then why are we using Copyright of all things as an argument for ethics?
Is our philosophy about "Equality" -- to encourage the equality of rights? Then why are we granting more rights to some, but not the others?
Is our philosophy about "Solidarity" -- to unite people together and help each other for a cause? Then why are we so obsessed with alienating people by labeling "us 100% good", "them 100% bad"? Why can't we put down the difference and try to concentrate our strength on what really matters?
Is our philosophy about "Capitalism bad"? If that is the case, what can we provide to those who join our cause? Have we built an ecosystem in which people can get paid? Or are we encouraging people to fight with "community-driven sticks" against "capital-driven guns"? Are we serious about winning this war, or are we charging blindly into our certain demise?
Is our philosophy about "Anti-Elitism" -- to aid the grass-root, indie artists to survive, to freely express themselves?
By denying people from using more capable digital art tools, are we saying that -- only those who are talented, or entitled enough to invest the money and time to finish a traditional art course, only them, can have the right to be called artists?
Impressionism was deemed worthless because they were "low effort works". Photography was deemed to be the toys of the riches. Post Renaissance artists were deemed less capable because they bought their tools instead of making the tools themselves. Digital art was considered "not art" because "computer does everything for you". The list goes on and on.
I was an untalented, self-taught hobbyist. It took me more than 30 years to become a semi-capable artist. If it weren't for the digital art apps like Krita, I would never have become good enough to express myself freely. Needless to say, I was proud of my "hard-trained skills".
Then AIGC/LLM happened. Like many artists, I went from making fun of the nightmare fuels, to picking the wrong anatomy, to feel truly threatened by how appealing AIGC pictures can sometimes look, despite the incorrect elements.
People can only improve by learning humbly from what is better. Chess players don't deny themselves from using AI in training -- as long as they are not using AI to cheat in the competition.
For artists, AIGC/LLM is a very effective way to find inspiration and do focus study on certain topics. We don't have to use it for "cheating". Don't throw the baby out with the bathwater.
I think AIGC/LLM is actually helping us see what part of creativity is truly human-distinctive, and what part of it can be handled by automation.
I'm not saying you can't take pride in using your bare hands to plow the field. But if everybody kept doing that, we will never be freed from the busy works, and we won't have the highly-developed culture as what we are enjoying today.
Heck, if basic survival remained to be so difficult, art as we-know-it wouldn't happen at all.
I know there are many unsolved issues around this topic. But I also know that the world has moved on, and we can't stop the tide of the times. Making a statement to deny ourselves from competition benefits nobody.
Maybe we will have to figure it out on the way.
Do I need to manually port these 2 fixes to 5.2 branch, or you can help with it? Thanks!
Fixed two :guilabel: format mistakes in Render Animation page and Fill Tool page.
Tyson Tan (0ef8cb41) at 31 Jul 08:35
Fixed a mistake in the Current Selection label.
Tyson Tan (761a280f) at 31 Jul 08:33
Fixed a mistake in the Nearest Neighbour label.
Please port this to the 5.2 branch as well. Or do I need to make another MR for this on 5.2 branch? Thanks.
Disambiguated the string "Static" and "Dynamic" for Wide Gamut Color Selector.
Tyson Tan (815a70a2) at 31 Jul 08:26
Disambiguated the string "Static" and "Dynamic" for Wide Gamut Colo...
Thanks Niccolò, I think I can come up with something with Day/Night versions.
Hi Nate, a few questions:
Other Konqi artwork ideas I have so far:
No, I'm not expecting anything like a default treatment. I know how much some people hated the cartoon Konqi and the anime Kiki XD
The reason I thought about providing a wallpaer now, is that I provided so many Konqi artworks in the past, but Plasma never had a wallpaper made by me, which is...weird. XD
Well, I can provide them anyway, hopefully something useful or a sketch in this week. I guess if I'm doing this, I can perhaps match the art style of other Konqis better.
(My "default" art style also changed over the years, so it's not a blind match anymore XD)
Hi guys, I hope it is okay to put this discussion here.
Is it okay for me to provide a new Konqi wallpaper for KDE Plasma, and possibly some other artworks too (if needed)?
BTW, sorry for not being able to provide some Konqi artwork when requested a while prior to this. I wasn't in a good shape to draw things, due to the circumstances.
Now things have been eased up a bit for me, so I hope I can make up for it.