@dcoderio: LLM-written documentation has a smell. and you can get rid of it tell the model to follow the ASD-STE100, the simplifie…

X AI KOLs Timeline Tools

Summary

LLM-generated documentation often has a noticeable 'smell' that can be eliminated by instructing the model to follow ASD-STE100 (Simplified Technical English), a standard with 53 rules and ~900 approved words designed for clarity and consistency.

LLM-written documentation has a smell. and you can get rid of it tell the model to follow the ASD-STE100, the simplified technical english. 53 writing rules, about 900 approved words, issue 9 of january 2025 it was born in the 80s so that aircraft maintenance manuals could be read the same way by a polish mechanic, brazilian and japanese. each word has only one meaning and one verb per action the side effect on LLM is direct: no pretty synonym, no subordinate clause, no embellishment. what's left is the instruction today 64% of those who use the standard are outside of aerospace
Original Article
View Cached Full Text

Cached at: 07/22/26, 02:31 PM

LLM-written documentation has a smell. and you can get rid of it

tell the model to follow the ASD-STE100, the simplified technical english. 53 writing rules, about 900 approved words, issue 9 of january 2025

it was born in the 80s so that aircraft maintenance manuals could be read the same way by a polish mechanic, brazilian and japanese. each word has only one meaning and one verb per action

the side effect on LLM is direct: no pretty synonym, no subordinate clause, no embellishment. what’s left is the instruction

today 64% of those who use the standard are outside of aerospace

Similar Articles

Fine-tuning an LLM to write docs like it's 1995

Hacker News Top

The author fine-tuned a local LLM on a corpus of 1990s Microsoft manuals to generate documentation in that vintage style, exploring local model customization for technical writing.

Auto-DSM Under the Lens: A Black-Box Evaluation Framework for LLM-Based DSM Generation

arXiv cs.AI

This paper presents a black-box evaluation framework to assess LLMs' ability to generate Design Structure Matrices (DSMs) from structured technical documentation. It introduces reproducible metrics and a composite quality score, showing that while LLMs can produce plausible DSMs, they remain sensitive to ambiguity and prompt formulation.

How to Write with an LLM

Hacker News Top

The article provides two key rules for using LLMs in writing: avoid using any words suggested by the model and disregard encouragement to maintain an authentic voice.