@davidrevoy re copyright issues and pessimism at these ever being addressed: that, i think is really the most tragic part of this all.

an artist friend yesterday said that for her, an artist with aphantasia (she doesn’t get ‘inner pictures’ in her imagination), generative imagery could’ve been such a useful tool. in an ideal world, she said, there would’ve been such a tool trained on everything that’s public domain or freely given to that cause. with restrictions on direct commercial usage of its results, and so on.

and maybe we could’ve gotten there. had this slowly grown from proofs of concept in academia, some ‘try at home’ enthusiasm without those huge pre-trained models, and so on. it would’ve taken so much longer. but there would’ve been time to tackle each of the really big problems it brought with it.

but we couldn’t have that because techbros and their insufferable ‘move fast and break things’ (as naomi wu pointed out: move fast and break things favours the financially stable, which artists tend not to be en masse) and ‘better to ask forgiveness than permission’.