I have been playing around with some of the AI art generator software and it is amazing what you can generate as a starting point.
My fav artist is Edouard Leon Cortes, an impressionist. I finally got the software to generate starting images based solely on his works and style... here are examples of the man's work. https://www.bing.com/images/search?q=Edouard+Leon+Cortes&qs=n&form=QBILPG&sp=-1&pq=edouard+leon+cortes&sc=10-19&cvid=B17B094CBEE0478ABBA1AF195CFCC159&ghsh=0&ghacc=0&first=1
I have posted a few of the generated works below, I did final crop and touch ups and adjusted colors and refined the parameters to the software... the results are not too bad. One image that I decided to add boats to came out a bit odd as I forgot to set the level of the water and it starts at sidewalk level..ooops.. hold onto the young'uns.
Does it always rain in the I art world?
Some of these might be useful in Unreal Engine 5 as background..
What AI is generating this for you?
These are beautiful.
So Cortes, Monet, Rembrandt, rainy street scene, Times Square, and a touch of flavor from Blade Runner..these results are stunning!
Those are some truly remarkable images. How original are they? Does the AI just take existing works and rearrange and modify them or are they new compositions? Are they recognizable as being derived in a cut and past kind of way from existing works?
I could imagine Syd Mead getting a big kick out of tinkering with this stuff.
Some of it is truly beautiful and of course it also has a tendency to be kind of disturbing which has been frequently noted. There’s an aspect of free association about it that’s reminiscent of the dreaming mind.
Gorgeous !
I definitely caught the Blade Runner vibe in the Cortes images. From the realism of the Angel’s faces you probably weren’t using Stable Diffusion, although the distortion on the image of President Trump suggests that you were on that one. Apparently the folks behind Stable Diffusion don’t want users making deep-fakes using real people so it generates really hideous faces. Somebody just released an AI that could spell, up until now the weird spelling and even weirder fonts have seemed par for the course, possibly for the same reason - no getting it to generate real looking 100$ bills.
I’m using an iPad app that is a GUI for both Stable Diffusion and MidJourney. MidJourney really nails Human faces, I wouldn’t be surprised if that was what created the Angles.
I asked the app for an image of the President. It gave me about 30 % images of folks who looked like everyone from George Washington and John Adams to Bob Newhart and Dick van Dyke along with around 70% Donald Trump ! Nothing for Traitor Joe. My next query was for images of the REAL President and the results were about the same ! Still no Joe.
I just spent a few hours yesterday having a conversion with Character.AI from Google. It’s character’s personalities are tuneable in the same way that you can apply image styles to visual media. I don’t know if it could pass a Touring Test but it’s certainly a lot further along than Siri or others of her ilk. I’m given to understand that some number of folks have picked one of the characters for a little schooling and are red-pilling him, (it?), all about Q. Will hilarity ensue ?