4 ms·

If shown out of context, I would totally believe that #16 was generated by one of the recently-popular text-to-image AIs. "Baby C-3PO in the fetal position, fl

If shown out of context, I would totally believe that #16 was generated by one of the recently-popular text-to-image AIs. "Baby C-3PO in the fetal position, floating in outer space, wearing Beats By Dre headphones. Retro-futurism. Black and white photograph"

jcims 4y ago
Some attempts, started with your prompt then started chasing the theme. Image generation process in the captions. Stable Diffusion and DALL-E did the best in general (the uploaded examples cheated by going two stage, do the starfield first then erase the middle and infill). Midjourney thinks of stars like you're drawing a children's book: https://imgur.com/a/pO6JFWH https://imgur.com/a/pO6JFWH Makes me think AI image golf would be fun. Start with an image, then see who can get to something approximating the finished image with minimum prompts. Nearly impossible to accurately score but it's a fun process.
- mxuribe 4y ago
  > ...AI image golf would be fun. Start with an image, then see who can get to something approximating the finished image with minimum prompts.... That would be such a fun game indeed!!!
- TigeriusKirk 4y ago
  Beeple ran a little mini-contest on Twitter for people to recreate one of his daily works with AI. The thread has people's results as responses. https://twitter.com/beeple/status/1559182964587651072 https://twitter.com/beeple/status/1559182964587651072
aasasd 4y ago
I'm gonna bet that the pic is an interpretation of an image from ‘2001 Space Odyssey’. Basically that child plus Sorayama.
tantalor 4y ago
I'm still trying to figure out what "obviously inspired by the word of artist Hajime Sorayama" means.