Well, can we risk that everything that could go wrong as already done so, thus, from this point forward it can only get better?
On a side note: is it a sign too much anime/manga exists?
This isn’t an issue with stable diffusion, this is an issue with the wrong data set being used. Diffusion isn’t like a lot of AI in that it doesn’t actually think (in as much as any AI thinks) it’s just finding patterns, so it doesn’t actually look at the situation and go “hmm, that’s not right”.
your earnest explanation of an AI’s inability to understand context, in response to an obvious joke, is kinda funny
Facebook stickers
Why are you here soldier?
“I want to make a difference!”
Why are you REALLY here soldier!?
“…I want big ol tiddies.”
Don’t you EVER forget that.
Dude, uncle got some deployment on the Frontline.