The User Is Visibly Frustrated
Summary
The article analyzes why coding agents can be frustrating to use, attributing much of the annoyance to the conversational UX that mimics human collaboration without true learning or accountability. It explores the emotional tension between interacting with a probabilistic AI and the desire for reliable, explainable behavior, suggesting possibly making agents sound more robotic to reduce the illusion of personhood. The author shares personal experience with Claude Code and questions whether postmortems truly help or merely pad the narrative.