If you expect to type in “comedy movie oscar bait five stars” and have it spit out a finished MP4, then sure, that’s not happening any time soon.

But movies are composed of shots. Most shots are shorter than one minute. Narratives are constructed in the edit. Actors talking to one another don’t have to be in the same room… or alive at the same time. One composite wide shot and a bunch of jump cuts will stop the audience from even thinking about it.

This is going to be used for short films before summer, the same way image generators were used for comics. Both generally terrible - but mostly because the people leaping into it are boring, impatient, and just want to go ‘look what I made!’ while pointing at the parts they absolutely did not make. It’s the fancy version of saying ‘my characters look so cool!’ when your webcomic is made from stolen Mega Man sprites.

But considering we’re about eighteen months removed from 256x256 blobs that vaguely resemble an avocado chair, and Sora slaps down a variety of pessimistic timelines, it seems incomprehensible to bet against using this for worthwhile storytelling. Sora spits out half-decent shots from text alone. Video-to-video style transfer has been in research papers for like five years now - and unless this is a completely novel form of generative network, that means you can probably insert your own footage halfway into the process.

Some of these networks are denoisers. They remove the parts of the input that don’t look like the prompt. Starting from random noise is only the laziest way to get a finished output. Any blurry approximation of what you want, any blob-colored animatics, any 1 FPS storyboard, should guide the network to produce matching results.

What that does for Tyler Perry, I have no friggin’ idea. I was under the impression most of his movies could have been filmed at his house. (Alright damn, A Jazzman’s Blues must have taken some money.) We are not decades away from twenty-minute OVAs of sci-fi bullshit that would otherwise cost a fortune. It will be a matter of months.

Narrative will come first and foremost, because this technology frees writers and directors from needing studios… not vice-versa.