How to use Claude to iterate on a physics-led video
A creator used Claude and Fable with a custom production harness to combine game telemetry, physics modeling, camera solving, and organized feedback rounds for a nearly finished level.
news.ycombinator.com
Suggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Use [production harness] to develop [physics-driven video sequence]. Combine [game telemetry], [narrative beats], and [camera constraints]. Track my notes, produce A/B versions for [lanes], and organize each round's deliverables and revision record in separate folders.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source poster / publisher
- mikeyla85
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- After two weeks, the author reported a nearly finished level produced through physics modeling, camera solving, and repeated feedback. The final style pass was still pending.
- Human involvement
- The creator gave repeated notes and conducted adversarial testing. A final Cinema 4D style pass remained planned.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- Built a harness to control After Effects and Cinema 4D with Claude.
- Improved the harness with Fable for the production task.
- Used Claude to gather game telemetry and construct a physics model.
mikeyla85 described a custom harness for controlling After Effects and Cinema 4D with Claude, subsequently improved using Fable. The project addressed a video-game-like level that needed realistic physics while placing videos to deliver particular narrative beats. This required reconciling motion-graphics framing with behavior driven by game rules.
Claude pulled real telemetry from games, built a physics model, and created a framework that tracked the author's notes. A camera solver adjusted weights for different constraints to produce alternate video versions. The framework included a player model responding to threats, while the author could emphasize particular narrative beats within the solver.
For each feedback round, the system created folders for the lanes being tracked, containing deliverables, A/B versions, and a Markdown file. The author supplied notes and repeated the cycle. They also conducted many rounds of adversarial testing; one report included 40 recommendations. After two weeks, the level was nearly finished, with import back into Cinema 4D for a final style pass still planned.
Who it's for: Motion designers coordinating physics-driven animation, camera constraints, and repeated creative feedback.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: Hacker News · mikeyla85
Frequently asked questions
How did the system handle feedback rounds?
The author supplied notes, which the framework tracked. Each round produced folders for the tracked lanes, with deliverables, A/B versions, and a Markdown file.
Was the final level complete?
The author said it was almost finished after two weeks. Importing it back into Cinema 4D for a final style pass was still planned.
More details
Remaining steps
- Generated alternate videos through weighted camera constraints.
- Repeated notes and adversarial testing rounds with organized deliverables.
About this account
The system modeled motion, solved camera constraints, and organized alternate deliverables while the creator supplied narrative priorities and revision notes.
Platform: Hacker News. Attribution: https://news.ycombinator.com/item?id=48953007. Prompt: Suggested (written by Agents In Use, not the author's). Sourced by the Agents In Use pipeline.
Credit: mikeyla85 on Hacker News
Source published
Added to Agents In Use
Source checked
Tools and topics
Claude · Fable · After Effects · Cinema 4D · Custom harness
Motion graphics · Cinema 4D · Physics · Creative iteration
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.