Story First

The Temptation

Opening a tool before the story is finished. Starting to generate images before the ending is written. Beginning character design before knowing what the character actually says. It's an easy trap to fall into because the tools make it feel productive. You're generating. Something is happening. It has the shape of work.

But what you're doing when you generate before the story is written isn't filmmaking — it's exploring. Exploring is valuable. It isn't production. And the moment you start treating exploration as production, you're building on sand.

You'll generate a beautiful image and fall in love with it. Then you'll write the story around the image. Then the next image will contradict the first one. Then you'll spend three sessions trying to reconcile a story built backwards from visuals instead of forwards from intent. I've done this. It costs more time than writing the story first would have, and the result is weaker.

What Complete Actually Means

A complete story is not a finished screenplay in professional format. It doesn't need sluglines or action blocks or proper scene headings. It needs to answer every question a production decision might require. What does each character want — and what are they willing to do to get it? What do they say, specifically, in their own voice? What happens in each scene and in what order? What is the ending, exactly? What is the emotional arc?

Length is not the measure. Completeness is. If a production decision requires information that isn't in the document, the document isn't done.

What It Makes Possible

The payoff is concrete. When you reach the shot that requires a character to say something specific, the line already exists. You're not inventing dialogue under production pressure with credits running — you're transcribing from a document written when you had time to think.

Intercutting becomes grounded. A trailer cuts between moments from different parts of the story. You can only make those decisions well if you know what all the moments are. Without the full story, you're guessing at what might exist in acts you haven't written yet.

You know what to withhold. A good trailer creates curiosity by showing enough and not too much. That calculation is only possible if you know the full story. Without it, you don't know what "too much" is.

Most importantly: building the trailer becomes curation, not invention. Which scenes are strongest? Which line delivers the most in the fewest words? What is the one image that contains the whole film? These are good problems to have. They're much easier to answer than "what should happen next?"

The Last Analog

The Arctic thriller had a complete five-chapter screenplay before a single production shot was attempted. When it came time to build the trailer, every decision — which scenes to include, which dialogue lines to use, how to structure the intercutting across six movements, which moment to end on — was grounded in something already written.

The trailer was curation. The story was the raw material. The editorial choices were difficult in the right way — choosing between good options, not inventing something from nothing under pressure.

Eight shots were reused directly from the teaser. That was only possible because both the teaser and the trailer were built from the same source document. They knew they were part of the same world because the world was written down before either of them existed.

A Note on Trailers for Films That Don't Exist

Machine Dreams makes trailers for feature films that will never be produced. It's a specific format with a specific logic — and story-first matters even more here, not less.

A conventional trailer withholds the ending because the audience can go and watch the film. A trailer for a film that doesn't exist has no such constraint. It can contain the best scenes from every part of the story. It can show the ending. It can include the single best line from each chapter — because there is no film to spoil.

But this only works if you wrote the complete story first. You cannot include the best scene from Chapter Four if you never wrote Chapter Four. You cannot end on the most emotionally resonant moment if you don't know what that moment is.

Without the complete story, you are inventing the trailer and the story simultaneously, under production pressure. That is too many problems at once.