Turns out treating your AI coding assistant with a bit of kindness actually pays off. The agent tends to generate cleaner, more optimized code when you're courteous in your prompts. Small things like saying please, giving clear context, and acknowledging when it does well - it all seems to make a difference in output quality. Interesting observation about how even AI tools respond better to positive interaction patterns.