For that it needs a "director" to say: "turn the horse's head 90˚ the other way, trot 20 feet, and dismount the rider" and "give me additional camera angles" of the same scene. Otherwise this is mostly b-roll content.
I'm sure this is coming.
And when we want it to match exactly in an animatic or whatever, it needs to be far more precise than this, matching real locations etc.
I've worked with other developers that want to build high fidelity wire frames, sometimes in the actual UI framework, probably because they can (and it's "easy"). I always push back against that, in favor of using whiteboard or Sharpies. The low-fidelity brings better feedback and discussion: focused on layout and flow, not spacing and colors. Psychologically it also feels temporary, giving permission for others to suggest a completely different approach without thinking they're tossing out more than a few minutes of work.
I think in the artistic context it extends further, too: if you show something too detailed it can anchor it in people's minds and stifle their creativity. Most people experience this in an ironically similar way: consider how you picture the characters of a book differently depending on if you watched the movie first or not.
I could see this being really useful for exploring tone, movement, shot sequences or cut timing, etc..
Right now you scrape together "kinda close enough" stock footage for this kind of exploration, and this could get you "much closer enough" footage..
So if it’s an optional tool, great, but some people would be fine with it, some would not.
Maybe this works for ads for duner place or shisha bar in some developing country. I’ve seen generated images used for menus in such places.
But I doubt a serious filmography can be done this way. And if it can - it’d be again thanks to some smart concept on behalf of humans.
And let the robot tween?
Vs an imperative for "tween this by turning the horse's head left"
That said, I personally think the solution will not be coming that soon, but at the same time, we'll be seeing a LOT more content that can be done using current tools, even if that means a dip in quality (severely) due to the cost it might save.
Because camera angles/lighting/collision detection/etc. at that point would be almost trivial.
I guess with the "2D only" approach that is based on actual, acquired video you get way more impressive shots.
But the obvious application is for games. Content generation in the form of modeling and animation is actually one the biggest cost centers for most studios these days.
And there are a lot more degrees of freedom to get something wrong in film than in a single still image.