Martin Fowler · 2026-10-29 · notable
Martin Fowler on LLM Voice — authors should write the prose themselves
Martin Fowler says text that sounds LLM-written makes him stop reading, and that authors should not let an LLM draft their work because editing rarely removes the style.
Martin Fowler argues that LLM-drafted prose is easy to spot and hard to fix, so authors should write it themselves.
What is it?
'LLM Voice' is a short bliki post Martin Fowler published on 29 October 2026. He describes the unease he feels when text seems to come from an LLM, which makes him lose interest in finishing it.
How does it work?
Fowler doubts that detection tools matter, because readers react before any tool could tell them what they are reading. He says "it's very hard to expel the LLM-Voice" even with heavy human editing, so he advises authors not to have an LLM draft their work. Outlining may be acceptable.
Why does it matter?
Fowler allows exceptions, such as Jim Highsmith using AI to keep writing despite Parkinson's, and LLM translation for non-native speakers. His main warning is that outsourcing writing also outsources thinking, since writing helps him clarify ideas.
Who is it for?
writers and developers who publish technical content