LLM Voice
Web page · https://martinfowler.com/bliki/LLM-Voice.html
First seen by Oriel on .
AI summary
Martin Fowler argues that recognizable LLM-Voice creates an uncanny valley, so authors communicating with humans should write their own prose rather than delegate drafting and thinking to LLMs.
- LLM-detection tools miss the human reaction involved, because readers may disengage before verifying whether prose was machine-generated.
- Fowler warns that human editing rarely removes LLM-Voice from an LLM-generated draft, while using LLMs only for outlines may be acceptable.
- Writing is presented as a thinking practice: delegating prose can also delegate the development of understanding and comprehension.
- Exceptions include accessibility, translation, form letters, and reports intended for machines, where LLM assistance may provide sufficient value.
- Fowler's policy requires AI-assisted writing on his site to be disclosed, and he says he used no AI for this post.
#llm writing #writing #ai ethics #communication
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Points
Score reported by the source.
- Hacker News: 597, latest observation ; +581 over observed period since
Comments
Comment count reported by the source.
- Hacker News: 329, latest observation ; +328 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| Hacker News | 50290177 | https://martinfowler.com/bliki/LLM-Voice.html Read discussion |