Karpathy: we'll spend far more time understanding LLM outputs — and shares ASD-STE100 writing trick
DavideCrapis · x · 2026-10-02
Andrej Karpathy posted that we'll be spending a lot more time trying to understand the outputs of language models, sharing tips including one concrete technique: ask your LLM to write or explain in ASD-STE100, a controlled language spec originally developed for aerospace maintenance documentation. LLMs are well-versed in it and its heavy constraints produce much cleaner writing.
A replier added that monitoring will become the key human function in engineering and governing these systems.
Related event: Karpathy shares techniques for understanding LLM outputs(3 posts)→
More from coding & agent
- Matt Pocock: GitHub Actions is the easiest way to start with software factories — mattpocockuk · 2026-10-02
- Daily brief: Grok 4.7 rolls out as base model everywhere, Claude Code ships Mods — testingcatalog · 2026-10-02
- Hjarni notes app ships MCP server so ChatGPT can read your notes — tehmadnezz · 2026-10-02
- Screen Studio trained on 5,000 fake macOS apps to build a UI-specific upscaler — CurieuxExplorer · 2026-10-02
- Codemode is just a subagent if you squint, argues developer Lucas Meijer — lucasmeijer · 2026-10-02
- Reddit user finds one prompt line makes Claude communicate visually — ZgemboAdislic_ · 2026-10-02