Tag
Jerry Liu discusses the counterproductive effects of over-constraining AI models like Claude Code with extensive prompts, and predicts that deference to model judgment will increase.