emortalcoil

Continuing my experiments to see what comes of asking an AI engine to make an image in the style of someone else... Mona Lisa in the style of Piet Mondrian. Not quite what I was expecting; I was thinking it would render the general portrait in a bunch of squares/oblongs, sort of like a mosaic. Even this: falls rather short of what I was hoping would result. It's not nearly as abstract as it should be, IMO. But asking AI to make the Mona Lisa over in the style of Jackson Pollock was even more disappointing: the background may well resemble some of Pollock's work, but, again, the result was not nearly as abstract as it should be to actually fulfill the request. The synthesis between the original portrait and the requested style is, at best, incomplete. But then, maybe that's kinda the point. AI image generators don't really do abstractions, and they really don't "understand" the request; they synthesize an image based on whatever training set they had.

Here's a piece of weirdness for you: a guy seems to think it's appropriate to go up to a stranger from behind, put his hand on her shoulder, and then say "thank you, ma'am, for wearing your hair the way it should be". WTF?? Do I really have to shave my head again just so I can buy groceries in peace?