As some of you may know, I’ve been dabbling with Ai generated art for a little while now, including some non-trivial efforts at creating fluffies via said medium. Until recently, I had had little success, with very few images looking even remotely like fluffies, much less giving me artistic control over meaningful visual representations of them. For a while, I had mostly given up.
Then I did what I pretty much knew I should have done from the start, and I made a LORA using existing fluffy art (from long dead accounts or artists that have given their permission)
That work is still ongoing, but I figured I’d give you guys a taste of what it can do so far, so please, enjoy and don’t be angy just because it is Ai, please judge these works on their own merit.
Part of creating Ai images is that you really do need to use a firm hand guiding it. A lot of what stable diffusion was spitting out, even with the trained fluffy LORA made no sense- for each picture posted here, I discarded dozens of others.
Also, I am still refining the LORA to produce what I am looking for more reliably, that is to say, fewer ‘duds’ like these.
A huge part of creating Ai art is figuring out what models respond well to which prompts (all these images start with a text prompt, aka ‘fluffy pony on fire, burning, scared, crying’) and specifically which words work well and which words cause problems.
It’s kinda funny, because while the ‘training model’ is being trained on the source pictures, I’m being trained on how to talk to it!
“I for one, welcome our new fluffy AI overlords. I’d like to remind them that as one of the first humans here to post AI fluffies, I can be helpful in rounding up others, to toil in their underground skettie caves!”
I can’t, my hands shake badly due to a medical condition, and as stated above, i only used old work from inactive artists or those who I directly contacted and who gave me permission.
You make it sound so dirty- as if each human must recreate all of human progress all on their own from scratch, which surely you did based on your tone-- how long did it take you to figure out calculus? What about the transistors in the computer which you built from the resources you harvested from the ground?
If each human started with nothing and had no help, no guidance and no teaching, we’d be little better than wild animals- in point of fact, feral fluffies would tower above us in terms of morality and ethics.
That said, what is so very different between an ai learning from the training data presented to it and an art student learning from the artworks presented to them? Is it some special magic that humans have that makes our experience unique? Surely it would be the peak of hubris to decide that we experience things so differently that we must despise all non-human experiences?
Finally, Fluffus, I’m still left with a nagging question, what is it about me using a computer to help me make art that angers you so? It does not diminish the quality of your art (in point of fact, I have long enjoyed your work, and I hope to do so long into the future) by existing. I’m not in competition with your work, as I simply don’t post the kind of work that you do, and honestly, I’m not sure if I’ll ever even release the fluffy_pony LORA I’ve created as I’m ever self-conscious of my work and I doubt I’ll ever meaningfully ‘finish’ it. I hope it’s not that more people creating art means that you don’t feel as ‘special’ anymore, as that would be a very petty reason to stomp on someone else’s creativity. But Fluffus is an honorable man.
This isn’t you being creative, it’s feeding things into a machine that’s still spitting out mangled artist’s signatures.
There’s plenty of ways to create art with shaking hands. I have such a condition myself that makes holding pencils painful, so I just adapt. Digital art is amazing. You don’t want to learn or imagine or create for yourself— all this machine will do is iterate and iterate based on what you’ve fed it, it will never truly come up with anything new.
No matter your ideas, all it can do is approximate.
It’s disrespectful to the inactive artists to use their work in this way, instead of just enjoying what they left. It’s digging up a corpse instead of collaborating with the artists that are here or supporting them. It’s self-centred and anti-community to just straight up steal and mash things together. Your keywords aren’t transforming or creating anything, just sick chimeras without souls… fitting, eh?
I just gave “drawing with tremor” a Google and there are so many resources, references and advice from people!
What matters with art is how much you put into it and -try-. Is everyone here at the same level of skill? No. Does it matter? Not at all!
Drawing takes time, observation and lots of practise but you can develop your own voice and style and ideas and that’s what creation is about. I know it’s hard when you’re limited in some ways but there really are ways out there to help and adaptations. It’s all about the will to do it.
However, I do want to note that it does take quite a bit of work to produce meaningful images from the vacuous mind of the AI. An untalented individual such as myself could never hope to coax anything more than generic pop garbage imitations from it. I’ve actually tried to use the interface and it’s more like wrangling a mentally defective animal than simply vomiting artistry into existence.
Think of it like photography. You’re not “creating” anything (semantically); you’re just capturing and recontextualizing something that already exists. Yet we call this art anyway. Granted, photography does require much more realwork than AI, but it’s the same principle. This reminds me of how people bitched and moaned about how the Star Wars prequels were too light on non-CG special effects, not knowing that Lucas used more miniatures for TPM than for any of the films prior to it. While concerns about trends in art are understandable, most people care only for the end product.
I think what matters is that we should always make sure to clarify what is AI-generated. We should learn to recognize it. It’s certainly inevitable, so instead of flat-out denying and rejecting it, we should learn what it’s applicable for and what it isn’t, and learn to not let it creep up unnoticed until we have no way to moderate it at all.
And I do agree immensely on the point about using the work of artists who don’t have any say in the matter. Perhaps it should be standard practice to allow artists an opt-in/opt-out for AI-generated art. No idea how that would work, but there needs to be some kind of guideline.
Even photography takes a good eye for composition, understanding the tools and using them effectively. Everybody has a unique way of using them and point of view and that is what makes it transformative— three people will photograph the same subject but there will be nuances to them all.
No matter what you put into an AI, it is reusing the images fed to it. It cannot imagine any more than you have given it. All it can do is piece together images to try and get closer to being told ‘yes’. I’ve used several of these tools before, but when it runs off of art theft as a core training tool? There is no way it can be ethical.
These are weak imitations of people’s actual work.
I would, at the very least, appreciate an AI category so that people who don’t like it don’t have to look at it. This community is so small and artists are such a core part of it that doing such an anti-artist thing like using their work and calling it ‘my own’ is such a slap in the face to people who have made art here.
What matters subjectively is how much you enjoy what you’re doing. Grim clearly enjoys what he’s doing. Beyond that, art is objectively measurable in quality.
My terrible mspaint scribbles of Twilight Sparkle shoving a can of string beans up her ass are nowhere near the same level of quality as the Mona Lisa. Even if I spent the rest of my life meticulously making one single mspaint image of Twilight Sparkle shoving a can of string beans up her ass, it would still be inferior to the Mona Lisa, even though I would have tried harder on it than Leonardo Da Vinci did on his masterpiece.
“You tried” is not a valid metric for anything beyond about first grade. That’s not how art works. Either it’s good or it isn’t.
