Hey there, Riley here. This is a sample issue: it shows exactly what lands in your inbox every Friday. Same sections, same length, same honesty about what worked and what didn't.
Welcome to the Dispatch.
The Breakdown
Every prompt that works has the same five parts
When a shot comes back wrong, it's almost never the tool. It's the prompt leaving a part out, so the model guesses. These are the five parts we write every single time:
- Subject: who or what, specifically. "a woman in a yellow raincoat", not "a person".
- Action: one clear motion. "turns toward the camera".
- Camera: the move and the lens. "slow push-in, 35mm".
- Light: time and source. "overcast morning, soft window light".
- Style: the look. "film grain, muted colors".
Put together:
A woman in a yellow raincoat turns toward the camera, slow push-in on a 35mm lens, overcast morning with soft window light, film grain and muted colors.
Now change one part. Swap the camera to "handheld close-up" and you get a different shot from the same idea. Change one thing at a time and you always know what did what.
Bottom line: five parts, one change at a time. It works on images and video alike.
Camera Language
The slow push-in
A push-in moves the camera toward the subject. Slowly, it says "pay attention, this matters". It's the first move worth learning because AI video models understand it well and it instantly makes a still idea feel like cinema.
- Write slow push-in (or slow dolly in), not "cinematic". "Cinematic" tells the model nothing about movement.
- Keep the subject's action small. A big action plus a camera move often turns to mush.
- Ask for one move per shot. Push-in and pan and tilt is three shots fighting each other.
Rule of thumb: one subject, one action, one camera move.
Prompt of the Week
A ceramic coffee cup on a wooden table, steam rising slowly, slow push-in on a 50mm lens, early morning side light through a window, warm film look with soft grain.
Run it, then change only the light to "neon sign at night, blue and pink". Same cup, completely different mood.
Your Turn
Write one prompt with all five parts and run it in any free tool. Then change only the camera line and run it again. Hit reply and tell me what changed. I read every one.
From the Bootcamp
Want the full version, with camera moves, continuity and editing step by step? That's what AI Motion Bootcamp is for. $19/month on Skool.
