Death at High Noon
Storyteller lives!
Writing about my previous efforts at video generation and the future of A.I.T.V. coincided with a couple of mildly noteworthy A.I. model releases:
OpenAI released image generation via GPT4o, blanketing the internet with a wave of Ghiblification
RunwayML released their newest video model, Gen-4
Runway also was accepting submissions for their annual AI Film Festival. I decided to dust off my API keys and gave it a shot. Here’s “Death at High Noon”:
Sergeant Ramirez finds himself in a grave situation when he lands in an Old West underworld. To get back his life, he'll have to do the danse macabre - survive Anubis' judgement and duel Death himself.
Overall, I’m impressed by how the models have improved in six months. It clarifies, with hindsight, what I got wrong:
Time spent on bespoke video and image fine tuning and local processing, which can just be abstracted with an API
and what I got right:
Building interfaces and tooling around those APIs, which are potentially even more useful now that that the models are better
I’ve been posting chapters of Copycats, my sci-fi novella, here and there ; my original aim with Storyteller was to animate the whole thing. I’m more optimistic than ever that it’s possible.
Storyteller lives!