Where this is already the right tool

Six places it is genuinely production-ready

1. 1B-roll and cutaways. Shots nobody studies frame by frame. A city at dawn, a road through fields, a corridor, an abstract texture under a voiceover. Cheaper than stock and specific to your script in a way stock never is, because stock is somebody else's idea of a generic road.
2. 2Impossible or unaffordable shots. An aerial over a place you cannot fly a drone. A street as it looked in 1970. The inside of a machine. Anything where the alternative is a location budget that does not exist.
3. 3Concept boards and pre-visualisation. The clearest honest win, and the one working professionals adopted first. A director can show a client the shot before the shoot. A moving storyboard now costs an hour rather than an animatic team. The output does not need to be broadcast quality — it needs to communicate a shot, and it does that well.
4. 4Product motion with the product composited. Environment, light and camera generated; the real object tracked in. See the previous lesson for why that split is not a compromise but the correct division of labour.
5. 5Short-form vertical social. High volume, low continuity demands, and an audience with genuine tolerance for stylisation. It is also the one format where a phone-only workflow is realistic from generation through to publishing.
6. 6Music video and mood pieces. Where dreamlike inconsistency is a feature rather than a defect, this tool is not a substitute for something else. It is the thing itself.

Where it is still the wrong tool

Say this out loud in week one rather than week four:

• A character-driven narrative with dialogue across many shots.
• Anything where a real person or a real product is the subject.
• Precise physical demonstration — a how-to with hands.
• Anything a client will scrutinise frame by frame.

For all of those, a small live shoot, 2D animation or motion graphics is faster, cheaper and better, and saying so early is a professional service rather than an admission.

What has actually changed in the work

Not "video editors are obsolete", and not "nothing has changed". What changed is specific:

• The cost of a bad first version fell to nearly zero, so far more ideas get tested before anyone commits money.
• Small teams can show a moving version of an idea in a pitch. That used to be available only to productions that were already funded, which is a genuine shift in who gets to compete.
• Stock footage lost part of its market. The generic establishing shot is now generated.
• The scarce skill moved. It was never "can you operate the generator". It is shot design, judgement about which take works, editing, sound and knowing where the tool stops. Those were always the skills. They simply became the entire job faster than anyone expected.

Whether that is good news depends mostly on which half of the job you enjoyed.

A delivery checklist

Before anything leaves your machine:

• Every clip trimmed of its settle and its drift.
• One grade and one grain pass across the whole timeline.
• Real sound under every shot.
• Frame rate conformed to one project rate; upscale once at the end, if at all.
• A visible label on generated content, at the start, in frame — not only in metadata, not only in the caption. Several countries require this and none of them accept a label a viewer will not see.
• Written consent on file for any real face or voice, specific to this use.
• The prompt, seed, first frame and model name saved beside the project for every shot that shipped.

The loop, which is the whole course

Pick one shot from something you are actually making — not a test. Write it as a shot with a duration. Make the still and fix one thing in it by hand. Generate three takes. Trim them. Put the best one on a timeline with real sound under it. Label it.

That loop is everything in the preceding nine lessons compressed into an afternoon. Everything after this is repetition, better taste, and re-testing the limits list every few months to see which items have quietly moved.