For a long time, generative video was judged through a single shot. Cinema, however, begins when one image has to enter into a relationship with the next.
The real challenge is continuity. A character must preserve face, costume and intention; a space needs a geography; light has to change according to a logic. Without control, the result remains a sequence of visual accidents.
That is why the work is moving from the prompt to the workflow. Reference images, storyboards, camera control, masks, compositing, editing and color become more important than the single sentence given to the model.
The decisive shift will come when we stop asking which model made a shot and return to asking why that shot is there. Technology becomes cinema only when a visual choice creates meaning.