How to Make AI Videos That Don't Feel Like AI Slop

Join my Community: https://www.skool.com/cliefnotes/about This whole video came out of one folder on my computer. The voice, the little pixel guy, every screen you see, all of it. So instead of just telling you how it works, I opened that folder and let the video show its own making, from the brief Claude and I wrote before a single word of script, all the way to the render you're watching. The shape is borrowed from how animation studios have worked since about 1930. Record the voices first, break every syllable down on an exposure sheet, time a storyboard to the track, screen the rough animation in the sweatbox, and only then ink and paint. Here that's five folders (voice, words, cues, scene and render), with one file coming out of each one and me checking that file before the next one runs. And if you want one for your own work, a newsletter, client reports, onboarding, anything you do over and over, the last part walks through turning your steps into folders, one short file in each, and one file at the top that tells any AI how to walk through them. CHAPTERS 0:00 One folder made this whole video 0:27 The secret software is already on your computer 1:10 Quick honesty: this is my cloned voice 1:25 How animation studios have worked since about 1930 2:32 The actual folder: CLAUDE.md, CONTEXT.md, stages, videos 3:09 Stage 1: the voice 3:45 Stage 2: the words 4:11 Stage 3: the storyboard 4:46 Stage 4: the animation 5:26 Stage 5: the render 5:49 Build one for your own work 6:29 The premium vault 6:42 Wrapping up MENTIONED IN THIS VIDEO Voices first at Disney, from about 1930 (Michael Barrier's interview with Carl Stalling): http://web.archive.org/web/2020110809... Audio and animation, AWN: https://www.awn.com/animationworld/au... The exposure sheet, one line per frame: https://en.wikipedia.org/wiki/Traditi... The Leica reel: https://en.wikipedia.org/wiki/Leica_reel Michael Barrier on Dumbo and the Leica reel: http://web.archive.org/web/2020111202... Pencil tests, The Walt Disney Family Museum: https://www.waltdisney.org/blog/ten-t... The sweatbox, The Walt Disney Family Museum: https://www.waltdisney.org/blog/found... The sweatbox notes, Cartoon Brew: https://www.cartoonbrew.com/classic/j... Sweat box: https://en.wikipedia.org/wiki/Sweat_box Ink and paint, The Walt Disney Family Museum: https://www.waltdisney.org/blog/look-... Claude Code and CLAUDE.md: https://code.claude.com/docs/en/memory ElevenLabs (the voice): https://elevenlabs.io Whisper (the word timings): https://github.com/openai/whisper Playwright (the frame-by-frame render): https://playwright.dev FFmpeg (the mix and the video): https://ffmpeg.org Metricool (the scheduling): https://metricool.com The pixel guy is Doug, a nod to Doug Engelbart. The narration is my own cloned voice reading a script I approved, and every screen in the video is the real file it made. Either way, happy learning everyone.