Claude Opus 5.0's Incoherent Output Sparks User Backlash

Opus 5.0 drives incoherence into the stratosphere

A GitHub issue details user frustration with Claude Opus 4.8 and 5.0's verbose, jargon-heavy language. Users report that the AI's default writing style is toxic and hard to read, forcing them to re-read sentences and even pipe output through other models for clarity. The issue highlights recurring patterns like invented metaphors and argumentative framing, and calls for a plainer, more concise default register.