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

[Roland TB-303]

Just tried to sign up on the official site. It wants my fucking phone number? Fuck that, I guess I add it to my list of sites to maybe register on if I ever bother to drop two bucks or so on a burner SIM.

You asked for FBI crime statistics? Sorry the best I can do is get you vanned.

 

[MarvinTheParanoidAndroid]

I love how we haven't had this stupid AI whining stuff in this particular thread for a good while.

Thanks for fucking that up.

>Calls me a whiner
>Whines about a disagreement

 

[Boulevard of Broken Memes]

Asked ChatGPT to "Write an Anime about Joshua 'Null' Moon and his do Slobbermutt fighting an evil villain called Queen Clara"

Title: "Null and the Slobbermutt: The Battle Against Queen Clara"
Synopsis:
The anime follows Joshua "Null" Moon, a brilliant but socially awkward hacker who lives a solitary life in his small apartment. He spends his days creating and experimenting with new technologies, often neglecting his personal hygiene and relationships.
One day, while working on a new project, Null accidentally creates a strange creature that he names "Slobbermutt." The creature is a dog-like robot with advanced artificial intelligence, capable of understanding and communicating with Null. The two form an unlikely friendship, with Null teaching Slobbermutt about the world and Slobbermutt helping Null come out of his shell and learn to connect with other people.
As Null and Slobbermutt embark on a journey to improve and showcase their invention to the world, they encounter a powerful and mysterious villain known as Queen Clara. Queen Clara is a tech-genius and the leader of an underground organization that seeks to use technology to control and dominate the world.
As Null and Slobbermutt become embroiled in a battle against Queen Clara and her minions, they discover that Queen Clara has her own tragic past that led her to become the villain she is today. Null and Slobbermutt must use all of their skills and intelligence to stop Queen Clara's plans and bring her to justice, while also learning about the importance of empathy and understanding towards one's enemies.
The anime is a blend of technology, action and drama with a touch of comedy. The series showcases the bond between human and machine, the power of friendship, and the importance of empathy and understanding towards one's enemies, even if they are considered as the bad guy.

 

[Involuntary Celebrity]

Then why use the "inspiration" argument? How is it able to generate new images without sources to draw data from?

read up on how CNNs work

do even the most superficial amount of reading on a subject before making authoritative statements about a new technology

please

 

[Pocket Dragoon]

I found an awesome source for Midjourney styles, and each entry has examples w/ prompts. Most seem to work with SD.
Andrei Kovalev's Library

 

[purpleflurp]

I found an awesome source for Midjourney styles, and each entry has examples w/ prompts. Most seem to work with SD.
Andrei Kovalev's LibraryView attachment 4258547View attachment 4258552View attachment 4258555

Damn, I was hoping the dude posted the datasets Midjourney was trained off of. I'd give an arm and a leg for those resources.

 

[Scavenger]

Asked ChatGPT to "Write an Anime about Joshua 'Null' Moon and his do Slobbermutt fighting an evil villain called Queen Clara"

Title: "Null and the Slobbermutt: The Battle Against Queen Clara"
Synopsis:
The anime follows Joshua "Null" Moon, a brilliant but socially awkward hacker who lives a solitary life in his small apartment. He spends his days creating and experimenting with new technologies, often neglecting his personal hygiene and relationships.
One day, while working on a new project, Null accidentally creates a strange creature that he names "Slobbermutt." The creature is a dog-like robot with advanced artificial intelligence, capable of understanding and communicating with Null. The two form an unlikely friendship, with Null teaching Slobbermutt about the world and Slobbermutt helping Null come out of his shell and learn to connect with other people.
As Null and Slobbermutt embark on a journey to improve and showcase their invention to the world, they encounter a powerful and mysterious villain known as Queen Clara. Queen Clara is a tech-genius and the leader of an underground organization that seeks to use technology to control and dominate the world.
As Null and Slobbermutt become embroiled in a battle against Queen Clara and her minions, they discover that Queen Clara has her own tragic past that led her to become the villain she is today. Null and Slobbermutt must use all of their skills and intelligence to stop Queen Clara's plans and bring her to justice, while also learning about the importance of empathy and understanding towards one's enemies.
The anime is a blend of technology, action and drama with a touch of comedy. The series showcases the bond between human and machine, the power of friendship, and the importance of empathy and understanding towards one's enemies, even if they are considered as the bad guy.

Yes very kino but;
>her

 

[Monkey Pink]

MY FAVORITE TOY IS BACK

 

[Lemmingwiser]

I'm getting a better GPU just for this. I'm surprised how fucking far AI has come. How have I slept on this?

What can AI not generate well yet?

I'm trying out all the different AI's. The music one (ampermusic) seems to be the least varied so far. Any good music AI?

 

[Stephanie Bustcakes]

What can AI not generate well yet?

Beside the obvious stuff like bad anatomy, extra fingers and legs etc, it has a real problem with overlapping objects or scenery where the overlapped area won't have good continuity with itself. If, for instance, you have a fence in the background and a tree in the foreground, the fence on one side of the tree won't line up with the fence on the other side. It doesn't always mess this up, but it often does.

Here's an example I just made. Look at the space between the legs.

 

[Lemmingwiser]

Here's an example I just made. Look at the space between the legs.

Sure. It's not hard to fix with inpainting though.

It's funny how the low tier graphic designers will become gophers to fix AI mistakes.

 

[BirdUp]

class-action lawsuit being filed against stable diffusion (archive)

 

[Lemmingwiser]

>Calls me a whiner
>Whines about a disagreement

 

[Stephanie Bustcakes]

Sure. It's not hard to fix with inpainting though.

It's funny how the low tier graphic designers will become gophers to fix AI mistakes.

For a while they will. There is a saying in the ai community "two more papers down the line." This stuff is advancing so fast that every quirk or oddity we find will be fixed by the time the version after the next one comes out. It is likely the extra fingers issue, for instance, will be greatly reduced in near future versions. Shit is progressing extremely rapidly.

View attachment 4263210
class-action lawsuit being filed against stable diffusion

Oh, good, a lolsuit. These guys don't seem to understand what the AI does. It doesn't "remix" anything. Not a pixel of those copyrighted images are present in the generated ones.

 

[Roland TB-303]

Any good music AI?

I’ve been using Dance Diffusion.

 

[The Mass Shooter Ron Soye]

Beside the obvious stuff like bad anatomy, extra fingers and legs etc, it has a real problem with overlapping objects or scenery where the overlapped area won't have good continuity with itself. If, for instance, you have a fence in the background and a tree in the foreground, the fence on one side of the tree won't line up with the fence on the other side. It doesn't always mess this up, but it often does.

Here's an example I just made. Look at the space between the legs.
View attachment 4263167

It sounded like Stable Diffusion 2's "depth-to-image" was supposed to tackle that problem.

If they make a future version optionally generate layers, that would be even better. It could even be done automatically based on the prompt.

https://stable-diffusion-art.com/depth-to-image/

 

[seri0us]

It did work on the punks and DnD players eventually. Both of those groups are rife with leftardism.

View attachment 4263210
class-action lawsuit being filed against stable diffusion (archive)

Ah yeah, making it "fair and ethical for everyone" by attacking the only instance of it that a normal person can self-host.

 

[AmpleApricots]

Oh, good, a lolsuit. These guys don't seem to understand what the AI does. It doesn't "remix" anything. Not a pixel of those copyrighted images are present in the generated ones.

It really was bound to happen eventually anyways. Maybe best to get it over with in the pioneer days. The tears will be delicious.

The place where I realized AI has evolved past the novelty of things like deepdream was when a) AI services for normal people sprung up that also managed to actually survive on their prices b) these AI haters appeared. As much as these artists belittle the AI art, they wouldn't be so malding if they didn't feel threatened by it's quality. Where were they in earlier (and significantly less impressive) iterations of the technology? Exactly.

E: Reading that lawsuit it is incredibly clear they have absolutely no idea how SD works. Grifters gonna grift.

 

[Monkey Pink]

It's 100% a lolsuit. Just posted about it in the AI Art Seething Thread.

By the way, Karla Ortiz is the GoFundMe organizer for this lawsuit, and she's the comic book artist who stole Dr. Strange from his creator.

Sarah Andersen makes those cringe SarahScribbles webcomics you see everywhere.

I have no idea who Kelly McKernan is.

I'm throwing down this 4chan post made about Ortiz last month, it's very relevant.

 

[Pendulous Dong]

Last session my players hired an ogre and requested that I make a size comparison between him and the halfling, and I'm blown away at how AI combined with shitty bootleg photoshop blows every RPG character artist out of the water.