Collective Cinema on X

Are you bored of the movies and shows available from homogenized content factories and streaming platforms? Imagine you could watch unlimited movies created by the most talented directors, artists, musicians and writers in the world. Well — now you can! Introducing Collective Cinema on X using tools like RunwayML, Stable Diffusion Video, Midjourney, Civitai, Scenario and Pika Labs.

Sounds interesting? Here is how it works.

tl;dr — It’s a chain of creators working together to produce a full length movie as a giant X thread using Generative AI tools

The process starts when one creator on X creates a 1–3 minute movie, tags another creator and introduces a plot or storyline plus details around which checkpoint, LoRa, negative prompts etc to use. It can also work by having a prominent writer or director suggest a topic and “Direct” the movie chain live online as people add to it to shape the direction over time. After several iterations, you arrive at a much longer movie.

Since we are on X, nothing stops the best directors in the world from suggestion or proposing ideas. Likewise, musicians can step in to provide sound tracks to the movie.

Why now

Tools have advanced to a point where individual creators are able to produce snippets of movies (1–3 minutes). In the future it may be possible to generate much longer movies individually, but in the near term — I see an opportunity for X creators to collaborate to generate much longer works by combining creativity and resources. Here is how it can happen today on X using existing tools.

Step 1: Start the movie: Initiate an X post with a list of tagged generative video creators

ex. “I want to generate a movie about RoboCop fighting with Aliens” (tag creators who I want to help) or something like: “What would happen if dogs could speak english tomorrow” You get the idea — It’s a common technique in improve comedy to provide an open ended prompt

— optional — — Include the checkpoint, LoRa, negative prompts, to set the cinematography of the user-generated film If you have a certain look and feel you want the movie to be in, tell people the parameters and tools to use to get the consistent look and feel you want the movie to have.

Step 2: Each creator generates a 1–3 minute movie and adds their own extension prompt

This creates a chain of movie scenes in an X thread. Creators go sequentially to add to the movie. Perhaps they add a prompt or go from an initial script describing a series of 10+ prompts for individual scenes. If done properly, this will produce a coherent movie where one thing leads to another

Step 3: Download and stitch together, add music and dialog

Once the movie reaches a certain length (it could go on forever), download the movie and stich it together. Edit with movie music etc.

Step 4: Release the movie

The original creator then tweets the final movie which combines the music, dialog etc. It’s a simple idea, but I think could have big results. Just thinking out loud here, I see huge potential. It’s time to think much bigger with what can be done.