Bryan Cantrill's "revolt of the reader": 78% of surveyed developers stop reading the moment they detect LLM prose, and 71% avoid the author afterward

Bryan Cantrill, The Observation Deck·high signalXBlueskyLinkedInCopy link

Cantrill's 2026-09-05 essay (403 points on Hacker News) argues that readers can reliably detect LLM-authored writing and that publishing it is becoming self-defeating. He cites Cynthia Dunlop's survey of 668 developers: 78% "stop reading immediately" on detection, 71% "avoid the author in the future," and 98% prefer an author's own imperfect writing over an LLM-polished version. His framing is economic rather than moral — like email spam, LLM prose will lose to filtering, so using it will simply stop working.