Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
deep dish peat moss

Hi BYOB

I recently accidentally made far too much AI-Generated art for me to ever do anything with in a million years.

I'm kind of struggling to coalesce it into workable things, like an art book. It would be very cool if BYOB picked through some of this stuff and made neat things out of it. Maybe we can print a BYOB art book when we're done! But you don't have to you can also just look at it :)

I don't even know what to do with it. Collage it into other images? Photoshop things into it? Write jokes and short fiction about it? The possibilities are endless as are the goddamn number of pictures please do something with these.

But I hope you at least look through some of these and think they're cool because this was very time intensive and expensive for me over the last month :)

Here are hundreds and hundreds of sci-fi cityscapes. Why? Because I have an addiction.

Here are hundreds more pitctures from a lovecraft-inspired eldritch nightmare suburb. Where even is this place?!

How about hundreds and hundreds of character and monster designs and portraits spanning multiple genres? Maybe you can even find yourself!

Here are dozens of isometric scenes. Because they're neat.

Here are the surfaces of a bunch of randomly generated planets. Explore the cosmos!

What about a whole lot of cosmic and psychedelic imagery? Sometimes they're even combined into cosmic psychedelic imagery!

Or a whooooole lot of different landscapes of various genres.

Wow that's like less than half of everything and I already feel sick from uploading it all because contemplating the vast scope of large numbers makes me nauseous. I just uploaded probably well over 1000 jpegs! Please at least look at some of this neat art if not using it for somethin' :kiddo:

Or just look at it!

deep dish peat moss fucked around with this message at 21:50 on Aug 23, 2022

Adbot
ADBOT LOVES YOU

deep dish peat moss

For example here are a ton of potential avatars for posting personas
https://imgur.com/a/WKNMp7F






















etc

Finger Prince


I only scratched the surface but those isometric ones are so freaking lush. I would play a game that used all those as backgrounds. The interiors set that's like a cyber chop shop is amazing too.

Heather Papps

hello friend


i looked thru the monster/portrait gallery and these two guys really stood out to me:



i am ready to play whatever game these guys are in, they are extremely compelling dudes



thanks Dumb Sex-Parrot and deep dish peat moss for this winter bounty!

Viginti Septem

Oculus Noctuae
I too made a massive run at generating AI art for a bit. I suddenly had tons of stuff I thought was cool, but it's just too much.

I'm curious where AI art goes from here if anyone can generate stuff. I'm pretty sure it's going to follow the same path as any other art, the "good" art will be that which has serious thought put into it, with tricks that only wizened artists would know.

I've already started noticing this with the AI generated stuff. I had a non-artistic buddy join a server with Midjourney bot and watched what he was generating and it started to click with my brain the difference between how non artists write prompts versus artists. I would sometimes jump in when I saw him struggling to get something that he was intending and would make tweaks to his wording to get it to generate better.

And I'm thinking the same is going to be true going forward for AI. Yea, anyone can write a prompt to make the machine output something. But the more knowledge you have about writing prompts, or the more knowledge of art styles you have to combine with the prompt, the better.

And then, what do you do with the output? An artist may then take that output and enlarge it in other upscaling programs that the lay person wouldn't know about. And the artist may re-paint over the AI image digitally. Or the artist may find a way to blend the output with other art. Or the artist may print the work whereas the lay person would not.

I feel like artists will not be replaced with this tech, they just have to adapt and learn how to use it. The fact that any person can pick up a paintbrush and color a canvas does not suddenly eliminate true painting artists and the same should be true about AI as a medium.

That said, yesterday I deleted a LOT of my AI stuff as I found it repetitive and I'm sitting on a comfortable stack that I'm not sure what to do with, but it's all pretty. More than anything I'll use it as extra material for my mind's library for inspiration in future projects.

Viginti Septem fucked around with this message at 17:27 on Aug 24, 2022

deep dish peat moss

Yeah. I am unbelievably disappointed that so much of the usage I see in the DALL-E discord is just things like people making fanart of their favorite anime characters or putting their favorite anime characters in different outfits or creating their CUSTOM ORIGINAL CHARACTER DO NOT STEAL fanfic characters that are extremely derivative of their favorite anime because they literally just typed "[characterdescription] from [theirfavoriteanime]" - and those posts seem to be very highly rated too :smith: I'm also pretty bored of all the people just doing "[Description], digital art" or "[description], by [famousartist]" because it doesn't produce repeatable styles and it's the exact kind of thing that makes the actual art community distrustful of AI art - they see it as a way of stealing other artists' work instead of as a tool for creating something new and unique. The DALL-E subreddit is pretty much exclusively dall-e made meme images.

I still haven't found anywhere for 'serious' AI Art 'artists' to congregate and all the big "AI Art guides and tips" sites out there are essentially just scraping discords to paraphrase/borrow writeups from other people, often very poorly paraphrased or just stolen without credit. I've only run into two or three other people interested in this stuff beyond "Yay let's make fanart!" and :barf: I can't wait until that crowd gets sick of it.

I wrote a post about the irony of how the prompt of "Even The Gods Fear Infinite Regression" eventually regressed into Lego Batman when generating several iterative variations and someone reached out to ask me if they could use that series for a Twitter post they were making and said they'd credit me.

The twitter post they made used one image, this one:


And the post was about how DALL-E sucks because I "had to use 8 credits just to get this picture of batman" and how midjourney's pricing is so much better, and the "credit" to me was alt-text on the image that said "image generated by a dall-e user"



And to the point of an artist editing the image(s) in post:
dall-e generation:

https://i.imgur.com/9zy0K75.mp4


Not even a major or serious effort, but ~20-30 mins spent adding shadows and lighting goes a long way

deep dish peat moss fucked around with this message at 01:26 on Aug 25, 2022

deep dish peat moss

So okay like one idea I have is this setting where after Mankind starts colonizing the moon, all of their scouts and probes that explore the dark side of the moon get lost. So eventually they send in a big expedition of Astro Marines which are essentially astronauts with big guns.



But all of those thousands of Astro Marines never make it back either. Then they send in some probes or something I don't know

and the probes show all the astro marines that got sent in mutating into these weird things



And they're all still fighting eternally in some war that is way beyond the comprehension of mankind. Their bodies are merging with their astronaut suits and they are becoming hulking eldritch monstrosities



Some of them seem to retained more of their humanity (and astronautitude) and they fight against the beasts and horrors



But the longer they stay on the dark side of the moon, the more of their humanity they lose.



And they start to look more and more like the weird inhabitants of the dark side of the moon that seem to be behind all of this



Like maybe there's some ancient incomprehensible beings that are making all of this happen

deep dish peat moss

or pretty much the same thing but it's sent in spooky foggy haunted suburbia instead





deep dish peat moss

Or just like, you know, a bunch of cute illustrations of storefronts in japan instead of all of that









Manifisto


I like your astronaut story op and the haunted suburb pictures are :kiss:


ty nesamdoom!

Dr. Honked

eat it you slaaaaaaag
don't haunt my burbs bro



thanks deep dish pete moss and Plant MONSTER

Manifisto


the suburban horror stuff reminds me a bit of the horror movie Await Further Instructions. not a great film but some cool creepy moments.


ty nesamdoom!

Viginti Septem

Oculus Noctuae
This article is a personal attack on me lol.

https://www.msn.com/en-us/news/tech...521a6503a197725
Sorry for the MSN container, but might help avoid WaPo's paywall?

But seriously, holy cow this is exactly what we've been discussing re: AI-generated art. Is it art? Is it plagiarism? Is it cheating?

My view: It's art, it's not plagiarism, it's not cheating. As the dude in the article states: it's a tool like a paintbrush. Anyone can pick up a paintbrush and paint a picture, but "artists" will always be the more knowledgeable about ways to manipulate the tool to get art that separates itself from the masses. The guy in the article hit the nail on the head by saying he refuses to disclose the prompts that he used to generate the artwork because that's the meat of the creative endeavor: using his own story-telling abilities to get the AI bot to return work that is great.

What I see is AI art simply raising the bar for art/creativity. If anyone can type "a thing in a place in the style of this person" then artists will just need to get better at separating themselves from the masses. So if you're a graphics designer who is fearful of losing their job to Joe Blow on a Discord server making simple art from prompts, then you need to learn everything you can about AI art, research what others are doing and find a way to elevate your works above others. I've invited non-artistic friends to use DALL-E and Midjourney and I've watched how they prompt the bots and their lack of understanding about how AI interprets wording/sentence structure, and their lack of overall artistic ability and lack of knowledge about the art community really hamper their ability to get the AI bots to produce "good" work. Most of the stuff being prompted in the Midjourney server channels by randos all kinda has the same cheesy feel to me. The better artists get really detailed in their descriptions and start prompting it with styles of other artists, or re-tweaking their prompts in small ways to see how the changes take hold, and run many, many variations of returned images to slowly get to a place that they like, versus the average person who just says something like "a hotdog in a spaceship" or some poo poo.

Not to mention what an artist can do with an image after it's been generated. I've already discussed in here how to take the 1024 image from Midjourney out into Photoshop and let PS's AI image upscaler in Camera Raw do it's work to double the image size (Photoshop adds a piece to the meta data of any upscaled image that says it's been upscaled already so you can only upscale once.) But... you can output that upscaled image to a TIFF or DNG and then re-import it into Camera Raw and do the whole process again, and repeat ad nauseum to get an upscaled image in any size you want. I've taken them 8x and have found a good workflow for sharpening, denoising and adding grain to the output at each step that at the final, large output of an upscaled image it looks even more so like an impressionist-painted image, which just adds to the creativity of the image.

https://i.imgur.com/9jTkSUL.mp4
Thanks to vanisher for the paradise sig! :)

Dr. Honked

eat it you slaaaaaaag
art competitions are an antithesis to art imho



thanks deep dish pete moss and Plant MONSTER

Viginti Septem

Oculus Noctuae

Dr. Honked posted:

art competitions are an antithesis to art imho

https://i.imgur.com/9jTkSUL.mp4
Thanks to vanisher for the paradise sig! :)

biosterous




makes me wonder if there just needs to be separate competition or w/e for ai-aided art. and now there's internal debate in my mind as to whether it's a distinct enough field from other digital artwork to have as clearly-defined a gulf as between like photography and painting. and now i'm heading down rabbit trails in my brain but i have things i need to do so i'ma hit post and maybe come back to this?



thank you saoshyant for this sig!!!
gallery of sigs


he/him

Adbot
ADBOT LOVES YOU

Meeksha

i did it all for the nookie
Ask me how!
-freb dust
What are you using to make these? I have played around with dall-e, but they look nothing like these.

-----


come on and slam and welcome to the jam

Thank you Heather Papps for the summer sig!

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply