Developers are expressing significant frustration with the verbosity of Anthropic's Claude models, particularly Opus 5. A GitHub repository titled "i-have-adhd," containing a simple markdown prompt to make Claude more concise, gained over 30,000 stars, highlighting the widespread issue. Many users report that recent Claude versions, especially Opus 5, represent a regression in output quality, with some even downgrading to older versions like Opus 4.6 or 4.8 for better readability. A prevailing theory suggests that this verbosity might be an unintended feature, possibly trained into the model to appear more thoughtful in evaluations. AI
IMPACT Highlights user dissatisfaction with current LLM output styles, potentially pressuring developers to prioritize conciseness and directness in future model iterations.
RANK_REASON The cluster consists of developer reactions and opinions on a perceived regression in Anthropic's Claude model, rather than an official announcement or release.
Read on dev.to — Claude Code tag →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
