Is AI Art "soulless"?
Thalidomide Vintage Ad Shirt $22.14 |
CRIME Shirt $21.68 |
Thalidomide Vintage Ad Shirt $22.14 |
Is AI Art "soulless"?
Thalidomide Vintage Ad Shirt $22.14 |
CRIME Shirt $21.68 |
Thalidomide Vintage Ad Shirt $22.14 |
Well, yeah, souls only exist in novels and science fiction, that's the typical artgay big COPE when dealing with ai instead of embracing and using new tools at their dispossal.
It's called sovless you massive newbies. There is a reason we write it different because it has a different meaning.
>newbie calling others newbies
sovlposting was always a tourist zoomer thing
I'm here since 2006 how about you?
Cool, what was the first thing classified as a "meme" by the general populace? I'll accept close answers.
Probably some cave painting of your mom 50000 BC
Why is it called sovl, oldgay?
I agree. Frick the haters. This is our world now and it's an exciting tool. If I ever get to it, I'll use SD to make textures for my game. I'd be stupid not to.
It's a nice novelty, but most of it looks samey and predictable. It has a definite "style," that makes it easy to tell apart from a decent artist's work. The interface you use to generate AI art is extremely black boxy and there's not much you can do to prevent this. As big tech starts clamping down on it and getting more restrictive, this problem will only get worse. Until there are decent FOSS alternatives, it will remain in the novelty/corporate realm.
>it will remain in the novelty/corporate realm.
Local image gen is infinitely better than on someone else's gimped server
You are searching false memories of a set of neural weights, a crude 1/10000000th approximation of the real process, as if you are activating the imagine neurons in your brain, it actually doesn't have any style because it is resolving 2d image states from the unconscious the dataset covered
>It has a definite "style," that makes it easy to tell apart from a decent artist's work.
It takes effort to stray from "default" poses, faces, and styling that AI models impose, but not difficult to varying extents. But most users are lazy, so they settle for standing 1girls and whatever "default face" of any model loaded by the software. This is why so much looks samey.
>novelty
All manner of artists are getting into its use, not just companies that benefit from AI graphic design. It is causing plenty of upheaval across the globe, we are only at the beginning of all the chaos.
>But most users are lazy
This wouldn't be a problem if so many of them weren't so impressed by their own slop. The most common (and bizarre) posting pattern is when they post several variations of the same idea/prompt that wasn't notable to begin with. Most of the time people don't even respond to them, so it's not like they're getting any (yous) or engagement, I can't imagine what's possessing them to spam their crap.
NTA but don't lie to yourself. Most people don't find "good ai" art interesting nor it gets (You)s neither. Most you get is people either replying to memes and shock value (eg coomers) or at best a style they haven't seen before that gets boring fast. More often than all of these and I can not emphasize this point enough, traction is circlejerking in every AI slop board.
And there's nothing wrong with this. People like different things, that's why you see people posting low quality that they think it's good, because they legitimately do, and then you see people criticizing every little pixel. Standards and taste varies wildly.
Nothing more obnoxious than an AI slop elitist kek. What you find impressive most others don't care, by and large.
>NTA
What? It says Anonymous in the name field.
It means "not that anon" (that he was replying to)
Why would he need to specify it if we are all Anonymous?
It's a newbie thing, let it rest.
I have to assume you haven't seen the average DALLE user's "contribution" to regular threads, because that's not what I was talking about.
AI art tourist detected
>it looks samey and predictable
Go to civitai or some more niche site, most normgays on twitter post just generic 1girl garbage, you can get any kind of style you want with AI, if you know how
>the interface
ComfyUI allows you to get as low level as it get without directly writing Python code, you can absolutely control the pipeline. And there are already many tools, loras, controlnet, etc. That give you much control over the process
>FOSS alternatives
...homie WHAT THE FRICK IS STABLE DIFFUSION
>Until there are decent FOSS alternatives
what rock have you been living under for the past year and a half?
Souls don't exist, christgays
Well sadly they do cause of wave function collapse.. AI by accident has soul because by using it to observe the images it has you are adding soul as much as witnessing a cool Minecraft world gen is adding soul to a computer
Makes you think
not really
Yeah they do
Proof? I can observe mine so I'm pretty sure they do
ye
it can be charmful but it gets boring eventually.
t. guy who accompanied stable diffusion since 2022
>but it gets boring eventually.
>t. guy who accompanied stable diffusion since 2022
your lack of creativity is why you got bored, we're far from done exploiting SD.
t. guy who has been in this for over a year, and if I run out of ideas or inspiration, there is always more foxgirls to gen in new outfits, settings, etc.
you're a delusional homosexual playing with your toys after everyone else got bored of them
>everyone
you only speak for yourself and presumably your discord troony buddies, but there's a whole lot of artists out there integrating into the movement little by little, don't you read or watch the news? there's stuff each week about some company or artist starting to use AI all the time, though also how AI is causing all manner of trouble
me? I'm a hobbyist with a lifelong passion for glorious 2D art and anime, lick my balls
>artists
Those who use AI in their “art” might as well have been AI generated from the beginning. This is the fascinating thing about AI “art” to me. It doesn’t actually change anything but allows one to filter the low quality “content creators” from real artists.
>Those who use AI in their “art” might as well have been AI generated from the beginning.
No. The software still has a lot of limitations, and for some folks, the workarounds are not worth their time, they are better off doing half the work by hand, but how much varies wildly, like everything else.
I mean in the sense of artistic worth. If you use AI “art” then whatever you wanted to create was already meaningless consumerist slop to begin with and there was nothing to subvert or make worse in the first place.
>If you use AI “art” then whatever you wanted to create was already meaningless consumerist slop to begin with and there was nothing to subvert or make worse in the first place.
You can wish all you want, you are clearly not an artist and don't get it. I hardly was before discovering SD and I get it. Believe whatever illusions float your boat.
>just generate more foxgirls bro
Very creative.
give someone a pencil and tell them to draw 10 images, most likely they'll end up stalling not knowing what to draw next
Yet non-AI art is saturated with 1girl pics, frick off
I've yet to see AI-generated anime art with different lighting than 7PM in the month of May.
I've been using SD since the leak thing and really started getting into it with NAI, but I have to agree with in that the novelty has largely worn out. ControlNet was a huge breath of fresh air, but stuff is still too random. Every time I get frustrated at it I open up blender and mess around a little bit because it's nice to have control again, even if it's at 1/10000th the speed. But then I lose motivation to actually work on stuff and go back to SD to generate a few hundred more images and then go back to blender when I get frustrated again and then I...
really started getting into it with NAI,
The NAI leak for SD, that is. Don't confuse me for someone that actually paid for that stuff
A lot of good AI art I've seen has a lot more 'sovl' than the majority of actual artists can do.
Your basic midjourney and dalle shite is extremely easy to spot.
I'd soon believe in machine spirits or computer chips containing demons than grifting artcels having anything more than the meager of souls.
If you gen an image that doesn't look as much as your common AI image, polish it up, and call it original, no one would know any better.
The "soullessness" is just nonsensical bias.
>the defining feature of AI art is that it is not made by humans
So wise! Time for 2.7 million upvotes!
First of all AI art is not the best but it isn't the worse neither. Anyone that claims either extreme is just a fanatic.
That out of the way and regarding soul or emotion. I daresay the argument is very nuanced. I find that for people it's easy to forget that for all intents and purposes AI is just learning from us or to be precise, the data we provide and curated for it. It's representations are human projections in essence.
So I'd say it begs the question, how much of our own projections are intricate to the soul and how much is learned through experience. For the western man a dog is a symbol of loyalty, for a Chinese is just another farm animal. So in this context, an argument could be said that AI outputs are the closest we get to see what the generalized interpretation of things are by mankind and ergo, the closest we get to visualize the collective soul of our species.
I asked the AI once about the future of art and to create a new art movement and it created something it called "neo sentience" were the artist provides the emotion and the ai provides the material and technical side of things.
it mentioned the idea of virtual landscape art.
this is most likely the future of art if artists embrace the AI
These posts are the closest to the truth.
"AI" is a predictor, not a creator.
If you ask an AI to draw you a girl, it will draw the most mundane representation of a girl you can think of, nothing memorable.
Even if you trained it on excellent "soulful" art of millions of perfectly crafted girls that are each and everyone brimming with "soul", the AI will just average these out, they will cancel themselves out when you ask for "A girl".
But that's not the AI's fault, it's 100% on the prompter.
When you ask a specific artist to draw you a girl, they will never do just that, they will never average out every girl they ever met; they have their style, sure, that can be copied, but they have their preferences in pose, expression, composition that are unique to them. Each of these preferences warps their perception of "a girl", that is maybe the essence of soul: how your unique perception (that is based in human cognition, ingrained at multiple biological and psychological levels) warps reality.
Replicating this level of attention to detail requires the prompter to intentionally elaborate on every single possible aspect of the image, in a way that prevents the AI from being "average" is certain aspects. At that point, when you've been toying for some time on the same image to generate "just the right" thing, this is when you've been adding "soul" to the thing. As shown on these timelapses. Think about it: what is the smallest amount of words you'd need to perfectly describe any image (made by a real artist) and how likely is an AI to generate something very close to it given this description? Well, not only it's extremely unlikely you'd get any good result, you'd most likely run into context size limits first.
Current AI can't into expression because they don't have agency and personal experiences. What is up to debate is how abstractly current and future systems can reason.
>Current AI can't into expression
You just explained why and how it can in the same post, you ignorant pseud. Or rather, you demonstrate that you use words you have no clue as to what they mean and how they are applied in the real world. You're as ignorant as everyone else on this board and this thread.
Anyone else despises AI “art” but wants to see it get better to filter out low quality human “art” that might as well be AI generated?
the only people worried about AI art are trannies making a living off furry diaper porn commissions
and graphic designers all over various industries
A man wakes up, spends hours working some menial job, goons to some AI generated kita-chan, listens to some AI generated music, then goes to sleep.
Answer truthfully. Are you _unironically_ okay with living like this every day?
I think it's a great life. Add ai waifu, animu, gaming n intimacy too.
AI generated slop : No
Man-made, AI-assisted art : Yes
What would you classify
as?
It's entirely AI-generated AFAIK, but it's not a concept the AI would be able to piece together in a single gen by itself.
There's
>single-gen, effortless AIslop
then there's
>well-prompted, hires-fixed AIslop with an appropriate artstyle, regenned over and over until all the little details like fingers and toes are right, then having the face, fingers, toes, etc. inpainted for even more detail, with any and all obvious flaws fixed via shooping and inpainting
A lot of effort and good decision making goes into non-shit AIslop.
If you can't appreciate the work and intelligent decision making that goes into making non-shit AIslop, then you simply haven't put in the time to thoroughly learn SD.
It is a tremendously powerful tool with unlimited potential for actual artists who wish to utilize it as a tool.
Also
soul
Sovl
kek this shit is 10/10
how did anon make it? I had thought about doing this shit with a surface or something and piping inpainting from a server
No. It's copyrightless.
>it different because it has a different meaning
Everyone who says that hasn't seen good AI art. Go to /trash/ see what furgays make on a daily basis. Literally a training issue combined with skill issue.
yes but so is most human art ,your point?
don't care about art, it's gay, i use AI to make celebrity porn
bulk of online human art is already soulless, just look at the ms paint bullshit that gets posted on deviantart daily
full on autismos in their 30s posting shit made exclusively using the circle tool
It is what it is trained on
If the outputs are soulless, the training data is equally as soulless
>no practice
Meanwhile millions of compute hours of training
Luddites really cannot even be called humans due to their lack of intelligence
>troony
>my dad said
>twitter screencap
>moronic question
Yep. I'm on nu-/g/
thats a youtube comment moron
Define soul
If the technology becomes significantly better you could make something with a soul, it's just that currently every single piece of AI artwork is shit.
The best you can do is make something visually busy, in an attempt to deter the viewer from paying attention to the countless horrible small details.
What kind of people comment on youtube videos? It's completely alien to me. I haven't made a comment on youtube since before google bought it.
The problem with AI is that it replicates so it has no real use. But if it at some point starts to create something new, then it might have some utility and can be adapted.
Another problem right now is that it has no value. If I can generate a infinite images myself why should I care about the infinite images you or anybody else has generated?
because it made concept art pointless
you can iterate quickly with inpainting
and then make models and textures based on the best concept art your AI generated after constraining the art style
it eliminates/compresses that part of the pipeline for commercial art
There's one subtle thing that raw AI outputs (without post-processing) lack, which is intent. It's subtle because you usually don't sit down and ask yourself "what was the artist trying to achieve here?". But the fact that there was conscious thought behind the creation process affects the end result in very distinct ways. An image drawn with intent will feel like it has something to communicate, whether it's a story, an emotion, a joke, or just the cuteness of the artist's waifu. With AI art, you get the visuals, but not the intent, so it often feels strangely empty. Details don't add up. Characters don't express any relatable emotions. There's no extra depth if you explore beyond the surface details.
That being said, people who make "AI will never..."-style statements are absolute imbeciles.
I don't know about AI images, but AI music as it is now, is super soulless
https://suno.com/song/14572e0f-a446-4625-90ff-3676a790a886
I liked SD and the other models at first, they are great tools to have.
But every homosexual that uses them happens to be either some insufferable tech-evangelist or EST who thinks he becomes Picasso.
One interesting thing to note: 99% of imagegens are garbage, and are treated as such, real art is similar though the ratio is at 80%
The traditional methodology is still king if you have a distinct style, you can make anything and if you've been doing it for a while, it takes 3-5 hours to make something that no model can offer, the overhead however is insane at 1000 hours to learn and develop a style.
It is a matter of time investment, but I'd still say learning to draw is worth, if anything it'll allow you to make use of the models.
All the homosexuals claiming they "can" touch up their drawing are delusional homosexuals.
>One interesting thing to note: 99% of imagegens are garbage, and are treated as such, real art is similar though the ratio is at 80%
you've never stepped foot into the art world, it is also 99% pure garbage, art these days is who you know and money laundering
not just AI art, all AI shit is soulless.
Most human art is completely soulless, so it's a moot point.
no
Art is generally soulless, craft is based thoughbeit
Soulless is just a cope, I can do a stickman art full of soul and yet it would be garbage compared to AI.
>AI art is soulless
>Dearest father says the main thing photographic '''art''' is lacking is just...humanity. No passion, no life, no skills and no practice behind it. it is simply soulless.
>Truly wise words. (dear god do my own farts smell good)
It's not even art but basically media player visualization.
It's honestly true. The brain is just too good at picking up on stuff like this. It's the reason body language is a thing.
>m-muh soul!!
it's all a big cope
Ultimately people use "soulless" as synonym for fake and Sonic looks like a genuine child's drawing while the usual AI slop looks disgustingly fake.
Try having sex.
Overrated. VR is better
SOVL
>AI morons are too autistic to understand the figurative meaning of "soul"
cringe
>pretentious pseudointellectual implies that there are people on this site who actually understand "sovl"
Inherently so.
Art is an exclusively human trait. AI cannot make art on a fundamental level. Even if a picture looks good, it isn't art if it wasn't made by man, because there is zero experience, skill, and expression behind a single stroke in the image. May as well say you're good at a game for using invincibility and infinite ammo cheatcodes; you aren't.
If something is not art but can still be called beautiful, what is it?
>zero experience, skill
Literally trained on the skills and experience of real artists. It's just as imperfect, too.
>expression
Mostly up to the user. Requires a creative mind with a good grasp of English. AI is a tool just alike a hammer and chisel, and takes practice to learn to world effectively.
It's also quite disingenuous to call AI "cheating".
yes, but not really in the way you expect
i had a pretty rough realization a few months back when i went on a ski trip and visited a local art gallery, and it was the first time i had ever been to an independent one
gallery was featuring an exceptionally good local artist - been painting people, landscapes, and everything in the town for the past 35 years
the prices on the pieces were all within the thousands of dollars, but the gallery itself was independent and there was no price to admission
every piece had a story about it, written by the artist, on a placard next to the paintings
i draw and paint as a hobby and really enjoyed the experience, but in the hour i was there, i was the only person that stopped by
i thought this was an anomaly because the town was small, so i committed to doing the same thing when i got home
i did, in a much bigger city
i was there for an hour, there were lots of paintings of aquatic life and beaches, in a similar vein of what i saw on vacation
there were multiple artists displaying work, and one was even there himself, and while most weren't as technically competent as what i saw on vacation, there were a lot of beautiful pieces there
4 people entered while i was there, it was two couples, neither stayed longer than 10 minutes
came to the basic realization: the majority of the population has never appreciated art, or craftsmanship, or pretty things made by people
then i found comfort: AI will never replace the relationship people can have with the person holding the brush
You've basically hit the nail in the head on why AI art is so devastating in that industry, barely anyone gives a shit and those that do can now do an amazing job at it for free or a small subscription.
Normies are scared of something new or sovlful, they want what they already know and whats popular, and AI is perfect for that.
I think using ComfyUI skillfully is like learning how to sculpt marble. It's a new art form and requires a lot of study to get good. The art comes from the human who envisions what he wants to make. Some humans can slooshy up better kino than others, but AI art (local only) is more of a tool like a hammer and chisel, than a creative brain