Why Is Low-Quality AI-Generated Text So Frustrating to Read?

AI-generated text is frustrating because large language models' linguistic expression often lags behind their coding and analytical capabilities, frequently obscuring critical details in real-world communication. The author attributes this poor quality to three core flaws: burying main subjects and actions, piling on obscure or vague phrasing, and making empty assertions that misplace emphasis. By contrasting human writing with AI output in programming scenarios like Claude Code, the analysis shows that such text substantially increases the reader's cognitive load.