Took me by surprise that this was AI generated, although in hindsight it should have been obvious. For me, this is the moment I realized AI art had become something "useful", and the world isn't going to be the same.
I can see where this is going... The commercial implications are enormous. Speeding up the concept art process for movies, etc. As someone here mentioned, why not make entire movies this way? Once they figure out how to animate this stuff, it puts the movie industry out of business.
I can only imagine what my grandkids are going to be using this for.
Literally everything will be made this way. In 10 years anything you’ll watch will be custom-generated specifically for you in real time.
I'm also intrigued about the potential of AI-generated animation.
However, I don't think the industry would be "out of business", but rather they would simply evolve into a new phase. The established movie industry will most likely have access to the most expensive and performant AI models to make short and long form animations, which would be time and cost prohibiting for hobbyists.
Cause he would definitely shit on AI generated art. He was bored of mindless American shit long before anyone was complaining - https://www.youtube.com/watch?v=xNQZF0KF-zw
What I‘d gather about him over the decades of work, is that he deliberately uses provocation/expressive words for his art/communication.
I think in the context of that little clip he was hinting at the ruthless commercialisation of American cinema/comics; and indeed since then it had well developed to today‘s Marvelization of Hollywood[0], a multi-billion industry of its own.
He is 93 now, so I think he doesn’t give a shit about anything at this point, but I like to think that back then as a young artist he would have embraced/explored/experimented with AI generated good shit to enhance/help his artistic vision. In the end it seems that mattered to him most.
[0]https://www.paudal.com/2022/11/24/quentin-tarantino-marveliz...
First, because his most compelling stuff is heavily surrealist and I've found the audience for that is niche and has very little overlap with what modern Sci-Fi/Fantasy fans have in mind. Second, because at this point whatever they have in their minds for what it would have been is likely an impossible standard for any director, much less the one that made El Topo or The Dance of Reality (not a slight, just pointing out this guy makes art pieces - not blockbusters).
production still from 1976 of Jodorowsky's TRON, 20 ASA 35mm --v 4 --ar 3:2
production still in version 4 doing most of the heavy lifting, with the year and film propeties giving a bit more vibe, and of course the overarching theme of TRON in the style of Jodorowsky. Add some extra words to differentiate scenes, such as light cycle or disc or computer. Add wide aspect ratio, and that's it. So it's 99% Midjourney. Mind blowing.
I think there is some skill in creating these, but the output can be really amazing and it’s constantly getting better.
However while there is an awful lot of 'background symmetry' AI seems woefully incapable of 'human figure' symmetry that, to me anyway, it really sticks out. Eyes not quite the same size, breasts pointing in markedly different directions, bone structure on the face markedly different to each side.
To that end I am also amazed at how forgiving our brains are for this stuff (a bit like being able to listen to music over a fuzzy, in/out AM station without much bother).
Anyway, thanks to whoever made this. I think there could be a market in 'imaginary film posters'.
P.S. Did it really need the 'Disney' logo on it though?
Instead of selling the 4K version of an old movie, now they can re-imagine it in 10 different styles.
Even with the unavoidable AI artifacts, the aesthetic of these images is incredibly evocative.
E.g., here, I'm skeptical about the specular reflections, which show a blue light on the left, a yellow light on the right, and a (seemingly) physically-accurate transition down the middle. Plus the too-plausible DOF in the background.
https://www.djfood.org/wp-content/uploads/2022/11/Jodo-Tron-...
This image, with the patterns following the 3D contour of her head. And the glossy reflection of the lights on the tabletop. https://www.djfood.org/wp-content/uploads/2022/11/Jodo-Tron-...
My wager is this is possibly using some AI-generated imagery, but it is largely a human art project.
This has been edited as a video for the song Soft Core by Tamaryn:
https://youtube.com/watch?v=AnZDaK9Vy0Y
The Gang’s All Here, and yes that’s Carmen Miranda: https://youtube.com/watch?v=xZp-s4wIc_s
Out of boredom, I fed the prompt "Fight Club" into Stable Diffusion, and it generated some interesting images of a human that looked like a mixture of Brad Pitt and Edward Norton.
I've seen similar mixed up results when I type a prompt that includes two obscure animals, such as "an emu playing chess with a lemur". The AI spits out an animal that looks like a cross between an emu and a lemur, instead of the two separate animals playing chess.
I want to know more about how these algorithms work, but that level of detail is hard for me to comprehend. I remember struggling in my college algorithms class to write code to do something relatively simple: I think the assignment was to: given a XY plane with a set of points, I had to find a set of lines that would create an intersection of polygons that contained only one point each.
This is a clear context in which re-doing something in the style of some Author can be radically inappropriate:
the visual style is stunningly good, but an unintelligent operation - it is relatively easy to mock a style, while the real thing was the judgement underneath. The Author knows if and why he would have used a curve made in such way in that place - not the simple mocker.
Edit:
in case some wanted to advance the idea that they are contented with the visuals, AJ is a foremost example of visuals studied to serve a content. Not to mention his declared contempt against states of poor hallucination.
It caused me to wonder if Tron was a world inside a computer ... or our own future (and, sure, why not both).
What do you think are the odds that we'll see a feature length AI powered "Jodorowsky's Tron" film within say.. the next 20 years? Some shepherding by humans allowed of course.
How far can the singularity possibly be?
This outfit:
https://www.alamy.com/star-wars-return-of-the-jedi-1983-20th...
I suspect that something like this is running already on scripts and vfx and that explains some of the really mediocre content from Netflix, Disney, etc (ie, “they don’t make ‘em like they used to”).
https://goodinternet.substack.com/p/if-jodorowsky-directed-t...
It looks like nothing was found at this location.
El Topo makes it clear.