Developers Slam Claude's Verbose, Jargon-Heavy Coding Output

Multiple developers have recently criticized Claude's coding output quality, centered on extreme verbosity and obscure semantics. RupFox reported that in a UI task asking to simplify button text to "Send", Claude wrapped the change in lengthy explanations and the generated code even included the old text, modification reasons, and failed test records. Chrisalbon noted Claude over-documents what the code does not do, while 4310sy and commenters said its comments typically run 3 to 4 lines. DarkSkyKnight complained Claude Code output is semantically dense and contradictory (e.g. mixing "undergird" with "overarching"), attributing it to benchmark-driven "pseudo-intelligence". ukanwat found Opus 5's long-session output littered with terms like "load-bearing" and "blast radius", hard to read even when correct; rules in claude.md helped little, and output styles plus hooks eventually mitigated it.

Confirmed

• Verbose output in UI tasks, with code cluttered by stale text and test records
• Over-documentation of non-behaviors and 3–4 line comments
• Semantically dense, contradictory Claude Code output flagged as pseudo-intelligent
• Opus 5 long-session readability issues, partially mitigated via output styles and hooks

Why it matters

• Readability of AI coding assistants directly affects developer efficiency and trust
• Verbose, obscure output forces users into extra configuration work to compensate

2026-08-15 ~ 2026-08-17 · 5 related posts

• [source] Opus 5 Long Sessions Produce Unreadable Output; Dev Fixes with Output Style Rules — ukanwat · 2026-08-15
• Users note Claude Code tends to write excessively long comments — 4310sy · 2026-08-15
• Dev critique: Claude obsessively documents what code doesn't do — chrisalbon · 2026-08-16
• [source] Users complain about Claude Code's 'semantic nonsense' in outputs — DarkSkyKnight · 2026-08-16
• [source] Developer rants about Claude's verbose coding style in UI tweaks — RupFox · 2026-08-17