Rick Manelius coins the acronym “AI;DR (AI; Didn’t Read)” for a growing frustration: people pasting walls of raw Claude output into Slack, newsletters, and social posts without reviewing or editing it. The post is short but the reaction is visceral — he describes physically flinching when respected colleagues send unfiltered AI copy.
The rule is simple:
- If you’re not bothered enough to review and edit it, the recipient shouldn’t be bothered to read it.
- Customer support and template copy are fine for unedited AI. Slack discussions, newsletters, and social posts are not.
- “It’s your name on it; are you proud of the prose and weird AI-isms sprinkled throughout it?”
The comments surface a deeper observation (from user Danny): AI creates “borrowed competence” — the gap between what someone can produce and what they actually understand widens. Outputs look sophisticated and use all the right terminology, but the sender may lack the underlying mental model. A vague thought like “we should probably collect the right data” becomes a highly specific 15-point technical requirement — with all the healthy hesitation and uncertainty erased.
The piece is a companion to Niklas Gruhn’s “Don’t be a meat proxy” — both diagnose the same anti-pattern from different angles. Gruhn focuses on code review and PRs; Manelius extends the critique to everyday professional communication. The “AI;DR” acronym is catchy enough to stick.