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

'Originality' is invariably drawn out from natural observation, though. Even the most fantastical worldbuilding is going to be informed in some way or another by one's experiences with reality. In theory it simply requires a sufficiently vast amount of observable information to draw reference from and, most critically, a talent for coercing the AI to meld that information in the ways you're picturing it in your mind.

It's like poetry, one might say.

but peoms can be come by anything to someone

this post is poemtry
 
  • Thunk-Provoking
Reactions: umami's milk
How much of the 'art' will truly be yours, though?
Who cares its quick and easy to do with some practice and more fine tuning industries can use it for proof of concept, place holder art and even final product with some help of people who can do touch ups. Imagine waiting a week for inferior commissioned works that can cost thousands of dollars when this AI can do 90% or more of the work for you. Brush to canvas work will still be a thing but as far as computer graphic design goes this technology will level that particular industry in a big way.
 
I like to get away from photorealism in order to skip that uncanny effect
art1.jpg art3.jpg art4.jpg art5.jpg art6.jpg art8.jpg

I made these a while ago and I tricked somewhat
art7.jpg

art9.png
 
Look I understand the concern but Null who said he had less then 3 days experience with this software took this original image and turned it into what is close to a Disney Princess in a short order of time. Imagine if this imperfect AI drawing was hit with some photoshop to fix some of the inconsistancies. This technology is pretty new to people and they are doing crazy things with it. At minimum this is a huge shortcut and time save.
v231231v321.PNG212v1v21.png

Like if you do not think this is an insanely useful tool for any fucking person you are in denial full stop. Instead of getting defensive about it I recommend you learn how to incorporate this into your skillset.
 
I am new at this and am just using an RX590 so it takes 9 minutes on the current settings.
The novelAI model is supposed to be better and I'll be getting that soon.


00000-2124125821-stalin and tr.png


turned it into what is close to a Disney Princess
The disney filter is already in an app, too. Being able to use the non-specific model on any image with img2img and getting a similar outcome is crazy.
 
I am new at this and am just using an RX590 so it takes 9 minutes on the current settings.
The novelAI model is supposed to be better and I'll be getting that soon.


View attachment 3749195
wheres the penis battle?!?!??!?!?!

Look I understand the concern but Null who said he had less then 3 days experience with this software took this original image and turned it into what is close to a Disney Princess in a short order of time.
You cannot tell the difference between a generic anime girl vs a Disney Princess.
Yes, a true wordsmith like you will have a blast using this software.
 
No matter how much good this kind of stuff can do for mass content creation and super niche stuff.. I can't help but fear a takeover by things like this.. even on a more practical level. And I don't even want to think about the potential for artistic.. "maturity" efforts, PC/SJ crap or censorship from a future where stuff like this is controlled corporately or "publicly" and artists are reduced in number.

Also let the drawn porn and hentai flow!
 
I specifically put down penis comparison and it didnt do it. I can try increasing the weight for penis
You can increase the weight for specific objects? Like, can I say "animal", and increase the weight for cats so that I can get a variety of animals but more cats than other animals?
 
Using AI will never make you an artist. You guys are to artists the same way trannies are to the sex they try to imitate.

Nah. You have to play with it to understand just how much human intervention is involved. It doesn't make you an artist in the traditional sense, it doesn't embue you with blending or shading skills or the ability to draw a curve but it is creative. You have to see the gigabytes of unusable junk I have in my stable diff output folder. Sometimes you get lucky but it's artistry in the same way that directing a movie is artistry, you have to make calls, decide which threads are worthy of pursuit, see the potential in an image/composition for iteration. The director has to get the performances out of others and form them into a cohesive vision and you have to do the same to get some of the really good stuff. It's also kind of similar to using a search engine and being skilled enough to get exactly what you want out of it.

Also I'm convinced that it trains your brain to see these potential morphs and visual similarities, just like how acid heavily influenced the art of the 60's by showing artists visuals they could not easily imagine before hand and that had prior to then only been present in the artwork of schizos. Using it is a form of learning and absorbing a style, just like viewing existing art or listening to a record. It just doesn't teach anything mechanical (your hands learn nothing).

On one hand it is going to help a lot of artists. It can be a composition tool for pen/paper artists and painters and such, it can generate textures for 3d modellers and game modders. It will have positive effects on the upstream creative pipeline. On the other hand will it absolutely destroy downstream artists and nobody will need to commission digital art unless it is very complex/specific or in an analog medium.

I think in the end we will see a splitting of conceptual visual art from 'performative/analog' visual art. The same thing is happening in music too. You have virtuosos and machines laying out crazy time adjusted/auto-corrected tracks but people will still go see a shitty punk band play live for the experience and the authenticity and the energy and the subculture will value tracks that have rawness about them. You will see a resurgence in traditional physical mediums (that can't be machine replicated without 100K+ industrial robots) and stuff like folk art, street art and tatoos.
 
How much of the 'art' will truly be yours, though?

Using AI will never make you an artist. You guys are to artists the same way trannies are to the sex they try to imitate.
If I have an exact composition in mind, and I replicate this mental image using the tools at my disposal - whether that is a graph pencil or AI + retouching - how is it not my art if the result corresponds to my exact mental composition and how does this then differentiate me from an artist?
You can increase the weight for specific objects?
You just put the word in ((( ))) in your prompt with more ( ) being more weight
 
I think a lot of people are scared of the unknown and they are feeling threatened by this new technology thinking they need to compete against it. I would just look at this as another tool you could use to elevate your work to the next level. Think about it you already have arts degrees which is the hard part, learning this new software is the easy part unless you are one of those purist types that do not want to stray away from the old way of doing things. It is like saying an 1800's butter maker makes better butter then a factory its just not true its a cope.
 
Back