i feel like this is 30% likely this year, depending on how fast SORA 2 is created. i’m not enormously confident though because this is very much lab prio dependent.

@firstuserhere will you accept systems that generate coherent 5-min films by splicing together many brief clips made with particular prompts (e.g., ChatGPT creates those prompts, which together have a reasonable narrative, but the videos themselves are of the type available today)? Or does it need to have normal length shots and scenes?