A 60-second 3D motion film with no keyframes

I made a 60-second 3D motion graphics film without keyframing a single layer by hand. Claude Opus 5.5 did the heavy lifting: 🎬 Wrote the story, built the storyboard, and generated the 3D assets in Canva ⚙️ Wrote an ExtendScript that assembles the whole film in After Effects: 6 scenes, 3D cameras, lights, kinetic typography, and transitions 🎧 Composed a synced sound design track from scratch One script. One click. The project built itself. But here's the part nobody talks about 👇 The first render wasn't good. The titles fell back to the wrong font. Two scenes were blurred from start to finish. The cards ignored the 3D camera. One transition made elements vanish and reappear the wrong way. The AI didn't catch any of that on its own. I did. By watching the render. So we iterated: v1 → v2 → v3 → v4. Each time I said what felt wrong, Claude analyzed the render frame by frame, found the cause, and rebuilt the script. That's the real shift. The technical barrier is almost gone. Knowing expressions, ExtendScript or the right 3D renderer is no longer what separates a good motion designer from a great one. What's left is what was always the hardest part: → Knowing what story is worth telling → Feeling when a transition is off by half a second → Seeing that a scene is "too static" before anyone can explain why → Having taste After 7+ years in motion design, I'm not worried about AI replacing us. I'm excited that it finally lets us spend our time on the part that was always ours: the ideas. The tools got faster. The eye still has to be human. What would you create if execution was no longer the bottleneck? #AfterEffects #Claude #MotionGraphics