Watching is not the only verb

There is a wall in entertainment so old that it reads as a law of nature: some people are in the story, and everyone else watches. It is not a law. It is an accounting fact. Putting a person on screen required a shoot, and a shoot required money, so the number of people who could ever appear in a story was capped by the number of people someone would pay to film.

That cap is what just moved. If a scene can be generated rather than captured, then the question of who is in it stops being a budget question — and once it stops being a budget question, there is no principled reason the answer cannot be the person watching.

What this is not

It is not a filter, and it is not a novelty face-swap on a clip. A filter puts your face on a video. What is worth building is narrower and harder: a series with a real cast, where one of the roles can be yours, across scenes chosen by the creator, with the story unchanged and the performance intact. The difference between a gimmick and a medium is whether the creator is still directing.

It is also not something a viewer should have done to them. A likeness is the most personal thing a person can hand a platform, and the only version of this that deserves to exist is one the viewer starts, controls and can end — with the creator deciding which roles and which scenes are open at all.

What it has to be to be worth building

• Consent that is real. A likeness belongs to the person in it: lent deliberately, used only where they said, and withdrawable — with withdrawal actually withdrawing it.
• Creator control. Which roles are open and which scenes are castable is a decision about somebody's own work, never a platform default applied across everyone's.
• Quality that does not damage the story. A version that breaks the performance makes the series worse, and being first is not a reason to accept that.
• Disclosure that survives leaving the site. A generated likeness travels, and it should carry what it is with it.

None of those is primarily a technical problem, which is why this is the part of the mission that deserves the most patience.

And then the harder one

Starring in a story is the legible version of a larger idea. The larger one is that a story built out of generated images does not have to be the same story for everybody — it can answer who is watching, in what it shows, in whose face it wears, in where it goes next. That is not a better distribution model for the medium we have. It is a different medium, and almost none of it exists yet.

That is the horizon this is aimed at. Everything else here is the ground it has to be built from.