Content
@
https://warpcast.com/~/channel/art-school
0 reply
0 recast
0 reaction
Sky Goodman š¾
@skygoodman
I've been thinking about forced obsolescense and the need to contunually upgrade systems and interfaces. This is true for many types of digital art technologies but nowhere is it as fast and as apparent as with using AI tools for art. I recently stumbled onto (after literally 2 years of daily iteration) what felt like a perfect workflow. But two things happened. Em props (which I was using as my way to access stable diffusion easily) and Magnific (as my upscaler) both upgraded. I never imaged that an upscaler upgrading would so wildly change the final outcome/look of a piece. But for some reason the same flow I used before now looks like and ai oil painting and the nuanced, wonderful look I had before, is gone. I have tried many new settings to get back to my old look, but nothing has worked. So I have surrendered to the fact, that, due to these upgrades, that are out of my control, my series is now "complete." There are 30 works in the series. I easily could have made 50 or 75 or 100.
3 replies
0 recast
8 reactions
Sky Goodman š¾
@skygoodman
But maybe we should secretly thank the forced upgrades? For, the series needs to stop somewhere and with AI, we seemingly could just go on interating at wildly fast speeds forever. So now these 30 works become something "rare." Something that I really can never get back to again. And while on the one hand, that is frustrating. There is also a kind of beauty to it. And something an experimental artist dabbling with AI in these times, has to maybe accept as part of circumstances for playing with these tools.
2 replies
0 recast
2 reactions
Travis LeRoy Southworth
@travisleroy
Yeah I was really disappointed when RunwayML sunsetted their GAN training, still havent found something to export my checkpoints to
1 reply
0 recast
1 reaction
Chris Follows š©šāØ
@chrisfollows
70 $degen āØļø
1 reply
0 recast
1 reaction
