Aviation's ASD-STE100 controlled language as an anti-AI-slop prompt hack, and where it fails
Paimaamu · x · 2026-10-02
Karpathy suggested prompting LLMs to write in ASD-STE100, the controlled English built for aircraft maintenance manuals, as a way to avoid AI slop. The backstory: in the late 1970s, European airlines needed ambiguity-free manuals for non-native-English mechanics, so the spec limits writers to roughly 900 approved words, each with exactly one meaning and one part of speech.
One user who tested this months ago reports it doesn't work well for scientific papers: papers aren't instruction manuals, and the controlled language fits IKEA-style documentation better.
More from coding & agent
- WebMCP lets pages expose tools to AI agents directly, replacing screenshot-based clicking — TejasKumar_ · 2026-10-03
- City Farmers joins OpenAI's first Codex Physical Builds batch with a Raspberry Pi setup — broodsugar · 2026-10-02
- SWE-sweep benchmark tests whether AI agents can find bugs before users hit them — klieret · 2026-10-02
- Turn Claude's Dot into a live DJ with browser use and your calendar — pvncher · 2026-10-02
- Indie dev builds 'AdSense for AI agents' with free search and 70% ad revenue split — Keats0206 · 2026-10-02
- One agent doing everything is a recipe for chaos: role-specialized agent teams work better — InfoTechRG · 2026-10-02