Bryan Cantrill — Oxide co-founder and longtime systems engineer — writes as a reader who is fed up. Too many people he otherwise respects are putting their name on pieces that are clearly written by an LLM, and he has two questions for them: do you think readers can’t tell, or do you think readers don’t care?
His answer to both: readers can absolutely tell, and they emphatically care.
- In a survey of 668 developers, 78% say they stop reading immediately once they detect an LLM’s hand in a piece
- 71% go further and avoid that author in the future
- 98% say they prefer an author’s own imperfect writing over an LLM-polished version
The objection isn’t style — it’s authenticity. When an LLM writes the prose, readers can’t tell what is real anymore, and they’re being asked to work through sentences the writer never worked to create. Cantrill calls this a breach of the social contract between writer and reader.
He sees LLM writing following the arc of email spam: unstoppable at first, then neutralized once detection at scale made it economically pointless — and reputationally damaging to be labeled with. The missing piece was reliable identification. LLMs make poor detectors, and superficial style-scanners miss the point, but Cantrill reports that Pangram Labs’ latest detection model has an astonishingly low false-positive rate (it rarely flags genuine human writing) and high accuracy. Oxide has gone so far as to mandate that all its public writing be verified human-authored by Pangram.
The practical advice lands cleanly: if an LLM wrote your piece from your prompt, just publish the prompt and spare everyone. Otherwise, use the model as an editor — LLMs are superlative at that — and write the words yourself. This is the same trust calculus as the meat-proxy problem in engineering: when you put your name on machine output you never shaped, readers (and colleagues) eventually stop trusting your voice entirely.