As someone very active in both the programming and art communities, all logic and reasoning always goes out the window when it comes to artists discussing A.I.
What "objective truth" would there even be, here? That it's possible to train a large language model to generate images in a specific style? Is that where the conversation should end?
There's a certain irony to having an art history education and not reflecting on the origin of 'technology'.
I don't know how many times I've offered to train a text encoder on someone's particular project to show them how quickly they could pipeline storyboards into animatics, or create a style LORA, or generate 3D assets for reference from a sketch, or whatever, just so they can see.
I've been calling my 3D and VFX friends "the old masters" for a while to drive home the point, and now that they're using some controlnet/ipadapter tools professionally, they understand. The traditional and 2D artists, though, are still grappling with unilateral veto after SD 1.4, although some are coming around.
7 comments
[ 2.8 ms ] story [ 24.7 ms ] threadIn that regard, AI generated art must be very poignant, because it makes me feel miserable
I don't know how many times I've offered to train a text encoder on someone's particular project to show them how quickly they could pipeline storyboards into animatics, or create a style LORA, or generate 3D assets for reference from a sketch, or whatever, just so they can see.
I've been calling my 3D and VFX friends "the old masters" for a while to drive home the point, and now that they're using some controlnet/ipadapter tools professionally, they understand. The traditional and 2D artists, though, are still grappling with unilateral veto after SD 1.4, although some are coming around.