Double Bruh. (Gaming)

by stabbim , Des Moines, IA, USA, Thursday, March 26, 2026, 23:46 (3 hours, 26 minutes ago) @ Bones

It's been pretty good at taking a texture I've fed it and outputting an infinitely-tiling version; that shit is tedious and time-consuming.

I'm aware that one can get the desired output, but for me that still doesn't negate the ethical problem with generative AI. Nearly all of this stuff is built on what I see as theft. One could train it on opted-in data, in theory, and that would be fine IMO (at least as far as the ethics of the training data goes - there's also the loss of the creative process which I don't love, tedium aside). But not many are doing that, and it's usually unknown to the potential user anyway, so how are they even supposed to make a judgement call? I guess what I'm saying is, for those of us on this side of the debate, saying there's no good use doesn't literally mean you can't ever get what you wanted out of it - just that there's no use case that's a net positive. Even when it does what you wanted, it's still bad, from that point of view.