I mean, even besides the whole AI angle, why the hell is this “series” a bunch of 3 minute video clips? It seems like the design around the idea was trying to appeal to “those dumb Gen Z kids who doomscroll TikTok all day”, instead of trying to actually make art.
Video generation models can’t make very long individual scenes, and need special tools to maintain any kind of consistency between them (if they bother to do that at all). Shorter scenes are easier.
That, and I think it’s supposed to be more “feel good” depictions rather than something educational… otherwise you’d just watch the Ken Burns series.
I mean, even besides the whole AI angle, why the hell is this “series” a bunch of 3 minute video clips? It seems like the design around the idea was trying to appeal to “those dumb Gen Z kids who doomscroll TikTok all day”, instead of trying to actually make art.
This is marketing for GenAI video creation, and it can’t do more than 3 minutes without melting the data center.
I can do 20-30 seconds using a WAN video model on a home PC. The length isn’t the problem.
Besides, it’s not like a movie is just a two-hour scene. But, it needs coherence and the tech is not there yet.
Video generation models can’t make very long individual scenes, and need special tools to maintain any kind of consistency between them (if they bother to do that at all). Shorter scenes are easier.
That, and I think it’s supposed to be more “feel good” depictions rather than something educational… otherwise you’d just watch the Ken Burns series.