AI Cinema gonna be wild
AI Cinema gonna be wild
Postmodern motion picture cutup techniques for the masses starring "Tron 2" by Jodorowsky
The weekend in the official Midjourney group on Facebook was a sight to behold. Since the kickstart of the fake movie craze mid last week — I think it was GANweaving on Twitter who started this wave, but I’m most likely wrong and ofcourse, since Dall-E 2, fake movies and mashups always were part of the fascination with image synthesis — dozens and dozens of movies from an alternate reality scrolled through my feed, all of them awesome and mindblowing and bursting of possibilities. I so hard want to see a silent movie from 1919 made by Fritz Lang that plays in the Hellraiser universe.
(I posted a selection of fine AI Cinema-stills in three parts here, here and here, and here’s Jodorowsky’s Tron that brought AI Cinema to the attention of the wider public.)
There are predecessor to this, ofcourse. The 2000s was the age of the mashup where people used new possibilities of editing cultural artifacts to create amalgams of well known popculture items, and the trend spilled over from music production into movie editing, one classic of the genre being the infamous retelling of Steven Spielbergs Jaws as a romantic comedy. Soon, those movie mashups not only were simple recuts, but technically sophisticated collages of various movies featuring elaborate editing skills. While I loved whoisevans Premakes back then, which transformed iconic films into black and white trailers for 50s scifi-b-movies, Fabrice Mathieu produced some of the finest example to date, his Master of Suspense being a personal favorite of mine. Using 30 movies by Alfred Hitchcock and his schtick to appear in all of them, his mashups told a new story through cutup-techniques. (Other great cuts by Mathieu are Far Alamo which fuses all kinds of Western-heroes with Starship Troopers and Darth by Darthwest, transplanting Star Wars-shenanigans into Hitchcocks classic thriller.)
I sometimes think of social networks (both on- and offline as in corporations and organizations and digital social media) as of form of artificial intelligence. We even have established a legal category for their personhood, a step that is yet to come for digital AIs. Their formation is not biological, and they certainly show a form of collective intelligence, emergent behavior and maybe even a form of creativity that is unique to all of those organizational AI-systems, as i call them. The creativity on display through the democratization of editing capabilities, resulting in the emergence of remix culture since the 90s and a flood of mashups in the digital age, is what i call “lived postmodernism”, and it wouldn’t have been possible without the collective intelligence of those social networks.
The digitization of cultural artifacts and the democratization of the editing process has enabled us, as a collective, to play creative games with this cultural grey goo on an unprecedented scale. Sure, the beatniks in the 60s used cutup techniques to create unique artifacts that kickstarted the postmodern revolution in art, but these techniques were exclusive to avant garde artists on the edge of contemporary thinking, or too expensive, to be used by the collective on a societal level. Digital Tools and price drops in hardware lowered that barrier to near zero, all you needed is technical skills. Then Social Media enabled us to select the best of those creative works for virality, spreading a contagious memetic creativity-virus that led to the mashup-craze of the 2000s.
Enter AI.
Suddenly, even that technical barrier is gone and anyone can create dozens and dozens of images from Jim Hendersons Labyrinth with H.R. Giger as art director. Even I, with a cheap-o laptop for 200 bucks, can create hundreds and hundreds of good looking illustrations from another universe where Picasso produced cavepaintings of The Terminator. Soon these will be videos, and then these will be movies. AI enables postmodern production techniques on steroids at a cost of near-zero, requiring no techincal skills. Our cultural past is liquid now, and the library is a soda dispenser with a billion trillion parameters.
We are already seeing all kinds of niche CKPTs for illustrative styles and themes (there is even a CKPT for Cats the movie), and as I understood Shutterstock, this is where they are going with this too: Specialized niche-models for all kinds of photography and aesthetics. Same will be true for video-production and moviemaking. In a few years, AI and finetuned CKPTs are about to create whole new markets and creative processes in all digital media design, from art and illustration to FX-work and animation to sound and music, for all entertainment industries. And all of this will be democratized and networked. (Also huge: Black markets for CKPT-files which may yield better results than authorized ones. Fun!)
In 10 years CKPTs could be community computed theme-parks that you explore visually in a VR-environment (in 20 years maybe via BCI in a lucid dreaming like experience). The technologies are already here (horde for stable diffusion for instance is communal generative AI done by distributed computing) and is akin to Bittorrent: peer2peer image synthesis. This will grow.
I'm very sure that "Tron by Jodorowsky" will be an animated video latent space within a year, and soon a game-like exploration experience. The next steps would be making those explorable spaces tell a story and synch to sound, which would be the emergence of real AI-cinema: a completely democratized new form of fluid visual storytelling, computed on a distributed network.
These emerging creative techniques will enable not only new ways to tell a story, they will enable a new way of seeing.
All possible futures gonna be wild, and we will watch all of them. On a screen, in your pocket, on prompt.
I’ll follow up this post with a seperate one featuring some favorites from these early days of AI-cinema. For now, enjoy part 2 of Jodorowskys Tron by Johnny Darrell:
