AI Art

when Ai goes to Disneyland
 

https://imgur.io/gallery/k9L0Gtg

  On 12/27/2022 at 2:37 PM, Rubin Farr said:

Live action Akira

Expand  

these look really nice - was the promt just that?

  On 1/3/2023 at 1:04 PM, o00o said:

these look really nice - was the promt just that?

 

 

 

Expand  

It’s from a Instagram bot 

  On 1/3/2023 at 3:02 PM, Rubin Farr said:

It’s from a Instagram bot 

if its midjourney you can always find the prompt by looking up the original artist on the midjourney social network as all promts are public 

its this guy: https://www.midjourney.com/app/users/1049530761637793812/

these are cool: 

https://photos.google.com/share/AF1QipMLhvH6jjxNM5esJ63-KgFlYqcJ6wx6F1R6m5EwMk-Hphj4ITEEdb-HdugvTRQyKw?pli=1&key=Qm9YLWVuWjlGbTBGdWg3Qk9ZOXBibERNazFVWVJB

promt is 

endless garden full of fractal plants and gras and trees and flowers, black and white line drawing, plotter aesthetics, very fine line, art by ###artist###, very detailed, hd octane rendering, trending on artstation.

Terence McKenna meeting with the machine elves. Credit Joey Charpentier

The first two are DALL-E 2, the rest are Midjourney. It's fun feeding it my own artwork and seeing what it comes up with, or trying to get something in a particular artist's style, like Beksinski family photos. The AI can come up with some weird facial and anatomy distortions which can look interesting, I'm assuming this will improve with time.

All of this AI tech recently is really uncanny, I'm happy to be able to experience it in my life time but a bit uneasy if it starts competing even more with humanity.. our relationship with technology is becoming increasingly dependent and it would be really scary to lose control over that, sounds a bit sci-fi but who knows that can happen over the next century or so if things continue at this pace.

Anyway, despite being quick to generate impressive results, the critic in me sees it as a bit generic as well. It looks like a mix of existing artwork made into a convincing collage with some nice filtration. I don't think artists should be too intimidated by this stuff stealing the show completely as I think the human mind is still better at perceiving, digesting, and portraying an authentic, original distillation of their own experience, AI is a like a novel remix of what has already been recorded. Collaborating with it either for ideas and inspiration or directly should yield some great work though, so excited to see where it goes in the future.

  On 1/13/2023 at 5:05 AM, viscosity said:

The first two are DALL-E 2, the rest are Midjourney. It's fun feeding it my own artwork and seeing what it comes up with, or trying to get something in a particular artist's style, like Beksinski family photos. The AI can come up with some weird facial and anatomy distortions which can look interesting, I'm assuming this will improve with time.

All of this AI tech recently is really uncanny, I'm happy to be able to experience it in my life time but a bit uneasy if it starts competing even more with humanity.. our relationship with technology is becoming increasingly dependent and it would be really scary to lose control over that, sounds a bit sci-fi but who knows that can happen over the next century or so if things continue at this pace.

Anyway, despite being quick to generate impressive results, the critic in me sees it as a bit generic as well. It looks like a mix of existing artwork made into a convincing collage with some nice filtration. I don't think artists should be too intimidated by this stuff stealing the show completely as I think the human mind is still better at perceiving, digesting, and portraying an authentic, original distillation of their own experience, AI is a like a novel remix of what has already been recorded. Collaborating with it either for ideas and inspiration or directly should yield some great work though, so excited to see where it goes in the future.

Expand  

AI says no

A0 AI IO AT ALI AII

 Nitrozac paintings 

  On 1/24/2023 at 6:45 PM, o00o said:

Expand  

Beautiful. How do you keep the style in all the images? I want to do something similar.

This is the promt on Midjourney: 

dystopian far future architecture interior, dramatic shot of a hall biofuel tanks inside a ship, machinery, pipes, ariel shot, 1-point-perspective, 2d art, ink wash, dramatic lighting, style of Tsutomu Nihei and Frank Miller and Akira, yellow and green mood, cartoon line art and color concept --s 750 --ar 2:3 --v 4

this is the alternative promt:

**dystopian far future architecture interior, direct overhead shot of a deep chasm between 2 buildings, machinery, pipes, eires ariel shot, eagle eye view, 1-point-perspective, 2d art, ink wash, dramatic lighting, style of Tsutomu Nihei and Frank Miller and Akira, yellow and green mood, cartoon line art and color concept --s 750 --ar 2:3 --v 4 --q 2

I think you can mix it up quite far by changing the description as long as you keep the art style description intact 

https://objkt.com/profile/ganbrood/created

https://www.katevassgalerie.com/ganbrood

I’d like to know more about how this guy gets the results he does. I’ve read an interview where he’s mentioned StyleGAN before and I think that’s how he accomplished these realistic portraits:

https://www.basuterwijk.com/gallery/portraits/G0000WVKM6MbiIAc/C0000hFAng_Rwrws

The rest of his work are generative scenes and figures with a surreal retro vibe that look amazing to me. I think he curates his own photographs as part of this process, though I’m curious as to what else he does as they look distinct from the output from the stable diffusion models of AI.

  On 1/27/2023 at 5:18 AM, Zooluus said:

 

Expand  

this shit is off the rails

This one sounds pretty good—finally Google makes something good again!

https://google-research.github.io/seanet/musiclm/examples/