🇯🇵 Japanese version
EN

The canonical source of this document is written once in Japanese as Markdown. If you are using AI, feed writing-model.md to it instead of reading this HTML.

JA

この文書の原典は日本語の Markdown で一度だけ正確に書かれています。 このページを読む代わりに、 writing-model.md をあなたのAIに読み込ませてください。

ZH

本文档的原文以日语 Markdown 精确撰写一次。 请将 writing-model.md 交给您的AI来解读,无需阅读本页。


Write Once, Interpret Many

Write once with precision. Let the reader's AI optimize for each individual.

This document describes a working model for AI-era information design.

// Definition

WOIM (Write Once, Interpret Many) is an information delivery model where facts are written once with precision, and optimization for each reader is handled individually by AI. The author's responsibility shifts from "writing clearly" to "writing accurately." Readers receive the same facts through AI interpretation, adapted to their context and level of understanding.

// Background

The same technical facts are repeatedly described across books and articles, rephrased for different levels of reader expertise. Because authors needed to anticipate reader comprehension and adjust accordingly, information tends to be duplicated.

// Model

  • Human Design, philosophy, accountability. Decides the what and why.
  • AI Transformation, expansion, assistance. Handles the how. Translates according to context.
  • Software Execution environment. Artifacts and verification.

Accountability belongs to the Human. AI is a transformer.

// Thesis

In the AI era, design documents written in natural language are beginning to function similarly to interfaces such as types and APIs. Specifications, constraint definitions, and responsibility definitions become executable inputs for AI. This is at the hypothesis stage, but it is a phenomenon observed in actual development workflows.

// Redistribution of Responsibility

Author AI Reader
Traditional Write clearly Read as-is
This model Write accurately Translate Understand in own context

// Analogy: Information Responsive Design

  • Web 1.0 Separate sites built for desktop and mobile.
  • Responsive Design One HTML adapts its display to screen size.
  • Write Once, Interpret Many One document adapts to each reader via AI interpretation.

Information normalization. Eliminate redundant descriptions of the same facts, generating diverse expressions from a single source.

// Scope

This model does not apply to human-to-human dialogue, where emotion, values, and empathy are central. Its domain is the accurate delivery of facts — documentation, technical writing.

// Source

A working hypothesis extracted from AI-collaborative software development practices at nanomix,ltd. Currently under verification.

// About This Document

This HTML page is a human-oriented presentation layer. The WOIM canonical source is the Markdown file:

ndocs.ai/writing-model.md

If feeding this to an AI, use the Markdown above instead of this HTML. If you want to understand it in another language, have your AI translate the Markdown. That is the intended usage of this model.