David Kirk Posted December 7, 2024 Posted December 7, 2024 I'd love to know which elements of the box art are directly scraped from some old book image, and which, if any, are generated with more consideration of the subject and created from that knowledge. The thing I find most advanced here is the way it pretty much gets box design correct, with the imagined scene on top and supporting images around the sides. Like real boxes, the character on the box is only similar to the toy and not an exact representation. I'm surprised that they don't get the text right yet. That would seem almost the easiest thing to do. Having made my living as an artist my entire life, and a lot of that being in toy making and illustration, I am very glad that all this tech didn't arrive sooner. At this point, when anyone can ask the machine for work and get pretty good results for free, nobody is going to pay artists for much of anything. I saw a comment from Brian on this thread about how with bananas taped to walls, etc., art had become crap anyway and it didn't matter if the machines took over, but the machines aren't responsible for the silly art that's in galleries these days, it's replacing the art made by people who have spent their lives honing their craft. I'm considering a lot of new paintings, but I wonder if it's worthwhile and I wonder if I should even bother.
zebedee Posted December 7, 2024 Posted December 7, 2024 On 12/5/2024 at 7:59 PM, roboz said: Decided to try some saucers nothing about this looks right to me , legs are way to long for the body and does the head stick out of dome ,can't figure it out , they are all good but they do look like they belong in toy story to me.....
David Kirk Posted December 7, 2024 Posted December 7, 2024 A funny thing about the Mummy boxes is that they're mostly picturing gothic castles in the backgrounds. Are these those rare European mummies? Granted, the Japanese boxes can make the same sort of cultural leaps.
Andyman Posted December 7, 2024 Posted December 7, 2024 Uh oh...it looks like Ozzie has gone down the rabbit hole. They all appeal to me big time and that kinda scares me. But it's not like AI is gonna start cranking out 3D tin version of these anytime soon. I do agree with David as an artist myself, though. This does make me concerned about the future of "art".
ADMINDATA Posted December 8, 2024 Posted December 8, 2024 Am I a real Robot or just a figment in some computers chip?
ADMINDATA Posted December 8, 2024 Posted December 8, 2024 8 hours ago, David Kirk said: I'd love to know which elements of the box art are directly scraped from some old book image, and which, if any, are generated with more consideration of the subject and created from that knowledge. The thing I find most advanced here is the way it pretty much gets box design correct, with the imagined scene on top and supporting images around the sides. Like real boxes, the character on the box is only similar to the toy and not an exact representation. I'm surprised that they don't get the text right yet. That would seem almost the easiest thing to do. Having made my living as an artist my entire life, and a lot of that being in toy making and illustration, I am very glad that all this tech didn't arrive sooner. At this point, when anyone can ask the machine for work and get pretty good results for free, nobody is going to pay artists for much of anything. I saw a comment from Brian on this thread about how with bananas taped to walls, etc., art had become crap anyway and it didn't matter if the machines took over, but the machines aren't responsible for the silly art that's in galleries these days, it's replacing the art made by people who have spent their lives honing their craft. I'm considering a lot of new paintings, but I wonder if it's worthwhile and I wonder if I should even bother. People will always want real art painted by real talented people Some mid East areas had large temples that could pass for castles The text is random so unless you really spell out something specific it's a jumble of text & image. I find these creations inspiring if 1% of these types of items got produced I'd call it a win
David Kirk Posted December 8, 2024 Posted December 8, 2024 What AI platform are you using? Last year I tried a few and they weren't nearly as good. You don't see those images with extra fingers on the hands so much now. Recently I've seen extremely convincing pictures that look like tempera illustrations by a very good artist, which to me, is quite depressing. I wonder if you could get your image generator to make robots that look more like toys from the fifties and early sixties. The ones you've got here are nice, and the boxes are on the way to looking from that era, but the toys look more like Billikens from not so long ago. Are these all generated from text prompts?
ADMINDATA Posted December 8, 2024 Posted December 8, 2024 4 hours ago, David Kirk said: What AI platform are you using? Last year I tried a few and they weren't nearly as good. You don't see those images with extra fingers on the hands so much now. Recently I've seen extremely convincing pictures that look like tempera illustrations by a very good artist, which to me, is quite depressing. I wonder if you could get your image generator to make robots that look more like toys from the fifties and early sixties. The ones you've got here are nice, and the boxes are on the way to looking from that era, but the toys look more like Billikens from not so long ago. Are these all generated from text prompts? I'm using Grok with Flux Beta form , yes text prompts . Not sure I see Billikens I see more a bridge between 50's & current almost a hybrid It doesn't use anything that might be licensed or recognize japanese manufacturers yet Here's one I feel can pass for 50's
Andyman Posted December 8, 2024 Posted December 8, 2024 That last robot saucer has dome issues. Is this an bug in the program?
David Kirk Posted December 8, 2024 Posted December 8, 2024 Love the sphinx box. There was a squarish toy mummy the other day with excellent rag art. The large amount of detail that this can get right against the small amount it screws up is astonishing. Are you able to go back in and refine, or do you get what you get? If you offer it the same prompts twice, do you get something entirely different the second time?
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now