Limit the Sequence to Four Legible Shots
A short AI-generated sequence should generally contain no more than four legible shots.
Capture
The original sequences attempted to show the complete chain of events: arrival, greeting, unpacking, instruction, correction, packing, departure, driving, reflection, and practice at home.
Every event seemed relevant. Very few remained legible.
Shots ended before the location, relationship, and action had settled in the viewer's mind. The sequence contained the story but failed to communicate it.
Why
Completeness appeared to protect meaning.
Removing an event felt like omitting evidence. Short clips also made every second appear scarce, so the temptation was to increase information density and represent every transition explicitly.
Why-Not
The viewer needed time to recognize each visual idea. The generator also needed time to establish the environment, locate the subjects, and complete a physical action.
Compressing more events into the same runtime did not preserve more meaning. It converted recognizable beats into fragments.
The second scar was learning that coverage and comprehension move in opposite directions after a certain point.
Commit
Decision: Treat approximately four primary story beats as the practical upper bound for a fifteen-second sequence under current conditions.
Use a simple progression:
- establish;
- develop;
- transition;
- resolve.
If the story requires more beats, create another clip rather than reducing each beat below recognizability.
Once the number of shots is bounded, the next problem becomes visible: each remaining shot can still be overloaded.
Timestamp
2026-07-24