Stable Diffusion, NovelAI, Machine Learning Art - AI art generation discussion and image dump

  • 🐕 I am attempting to get the site runnning as fast as possible. If you are experiencing slow page load times, please report it.
This is nothing but derivative from the training set and limited to more than very gross descriptions:
1666099386312.png
AIs do not understand soft tissue manipulations, for instance, so hamburgers made from one's cock and balls will not be drawn from it.

Likewise, the context clues sometimes just go off entirely:
1666099436342.png
 
That's the nice thing about sticking to a performative artistic medium, it's not actually possible for a computer to replace a live, interactive performer.

Why try? I no longer want to learn new skills, because the moment I learn them to any competent level they'll just be automated and unnecessary.
That depends on why you're doing it. When you're engaging in an artistic pursuit, there's a meeting ground in the middle between clinical precision and expressiveness.

That's an overlooked part of artistic competency, is the adaptive human element, and it's what sets really good artists apart from the mediocre ones.
It's like being a good speaker, it isn't about using the correct words, it's a lot more intangible than that, and the more concerned a person is about being "right" the less effective they tend to be. It's about emotional conveyance, which computers can't replicate because they don't actually have emotions.

I for one support this whole thing because I think if you're an "artist" and a computer can reproduce what you do, then you're defeating the purpose of art.

I’m all for protecting kids from pedos that are so rampant on this here internet, but some of you motherfuckers are uncomfortably obsessed. is there a predator in the mirror too, buddy?
I genuinely have no idea how anyone can think about anything as much as some of these people do about the topic of attraction to children.
 
You know, I've been hearing in this sector for years about the risk of deep fakes to end the careers of businessmen, politicians, etc. I always blew it off as the tech won't be up to par for another 15-20 years, but damn, this ai generated art shit came out of no where in the past couple months.
 
Y'all saying this will make the art market crash know nothing about art. People who already have a presence in the industry, a following and are making good money out of it will continue doing it. The issue is for people who are new or want to start making art. Artists who are already skilled beat the AI and have the advantage of being able to create content they can patent and copyright. AI generated images can't be considered intelectual property of nobody and for this, no reputable art business can use them and call it their own.

It's a depressing scenario for starting artists or aspiring artists. But again, this is the modern world, gotta adapt or die.
 
Procedurally, generally writing is easy. Childrens movies already follow very few tropes. One day TV tropes will go from just being a den of autism to being a resource for ai to identify what media tropes stimulates humans. "ah yes so humans enjoyed THIS scene in The Count of Monte Cristo therefore I should recreate this too".

Procedural generation is already a thing for video games, with much of the 2008 game Spore and unreleased Star Citizen having procedurally generated environments. Just a matter of time.
"Grandad what was it like before all the globohomo movies were made by AI?"

It's a depressing scenario for starting artists or aspiring artists. But again, this is the modern world, gotta adapt or die.
AI music isn't at the level of AI art yet, but this is from a 2018. AI generated after being trained on a bunch of country music:

I think all of these examples are showing that the future will be in AI generated human curated art. The job will go from making a whole piece to tweaking AI outputs till you have the right sort of thing.
 
AI music isn't at the level of AI art yet, but this is from a 2018. AI generated after being trained on a bunch of country music:

Honestly there's so much music already, why do we have to generate more? It's mostly IP that people can't use already created music for their vision for a video game/movie. But I guarantee if you want an EDM track it's already been made, it's just IP stopping you from using it so you have to get an AI to create an original.

Actually, I am legitimately asking this, if I want to use some random guy's track in my indie video game or whatever, how is the contract usually if it's something not really popular? Like I'm not making something that's going to be popular, and your track isn't that popular either, how much to use it?

Though I do think if you like a guy's style, you usually hire him to make the music.
 
Last edited:
This is going to be a gigantic shitshow. I've barely spent a day fucking around with this and I can tell you right now you'll be hearing about this for years.
  • Artist outcry about being replaced by a $600 GPU.
  • IP concerns in that an artist's style can be trained into a hypernetwork and reproduced instantaneously.
  • Government, legislative, nonprofit, and public outcries regarding the triviality of producing near lifelike CSAM.
  • Privacy concerns in being able to make passing nude / otherwise offensive or humiliating images of real people.
I'm calling it now, the Information Era is dead and we are now in the Machine Era. Computer information processing is now more important to the direction of society than the general ubiquity of information which defined the information era, just as the WWW redirected the world from the Nuclear Era and the cold war standoff to a global interconnected community.

Governments using ML to memorize faces from public IDs to track people in public (China literally already does this), civil litigation over copyright defemation and nonconsensual pornography, criminal litigation over production

It's happening.
Yeah that's cool and all but do privacy and human rights get me the ability to have an Ai generate lolcows hate fucking while shaking their fist at a picture of you?

Didn't think so.
 
  • Like
Reactions: JJLiautaud
I honestly don't know if I prefer Stable Diffusion or MidJourney AI more for AI art. I haven't played enough with them both to decide, but thanks for the opportunity!
 
Y'all saying this will make the art market crash know nothing about art. People who already have a presence in the industry, a following and are making good money out of it will continue doing it. The issue is for people who are new or want to start making art. Artists who are already skilled beat the AI and have the advantage of being able to create content they can patent and copyright. AI generated images can't be considered intelectual property of nobody and for this, no reputable art business can use them and call it their own.

It's a depressing scenario for starting artists or aspiring artists. But again, this is the modern world, gotta adapt or die.

AIs only work off of an existing library of samples and context clues, so you might have it compose stuff for you ($600 AI photoshop) or make the woman/woman centaur from my post above.

I'll put it this way: a real artist could be given pictures of null, Allison Rapp, and such, and draw us "Alison Rapp dressed like Mario, pegs Null in Princess Peach's dress, while Keffals stands in a tub full of troonshine." An AI cannot possibly do that.

An AI can't even sketch that retarded scene out for composition before filling it in! An AI can't take a sketch, even with tags and metadata and context clues, and fill it in. AIs might help generate iterative bullshit and an AI tied to an artist's library could get better and better over time, sure, but it isn't creative or able to compose scenes at all.
 
The quiz
-snip-
36/45 fuugggggg I should have put more scrutiny on the mediocre work. All my wrong answers were me thinking it looked normal and not double checking, which is probably the level of attention people put on retweeting art on twitter. So as long as you don't show the hands, or redraw just redraw them, generic or mediocre anime art is indistinguishable from AI. And if you're also not planning on doing anything that actually requires scene composition.
Are there any examples of women without massive tits made by AI? Are we collectively training AI to always draw women with massive jugs?
You can get women with normal sized chests, it's just that coomers specify "(((((huge breasts)))))," and get results with exactly that.
 
Y'all saying this will make the art market crash know nothing about art. People who already have a presence in the industry, a following and are making good money out of it will continue doing it. The issue is for people who are new or want to start making art. Artists who are already skilled beat the AI and have the advantage of being able to create content they can patent and copyright. AI generated images can't be considered intelectual property of nobody and for this, no reputable art business can use them and call it their own.

It's a depressing scenario for starting artists or aspiring artists. But again, this is the modern world, gotta adapt or die.

Keyword : Reputable.
Even the reputable companies know how to steal without getting caught.

No one ever has to know what image you used as a reference - unless you're an idiot that uses the first thing pops up in google or you use the most popular image of the fucking week.

An photo of a woman holding a sword.
Person A - Traced it and turned it into Wonder Woman holding a sword.
Person B - Found a model making the same exact pose and used it as a reference image, to draw Arden Lyn holding a Lightsaber.
Person C- Cut up a magazine and frankenstien'd the same exact pose as A and B to make Joan of Arc holding a Sword.

There's a finite number of poses and compositions. Every position that ever will be drawn, has been drawn. It's up to you, as an artist to present this in a way that makes you look special.
 
Back