Pasting whole files into an AI wastes tokens, and accuracy. Many AI assistants are fed whole files, or chunks of plain text. They spend tokens reading code that has nothing to do with the task. And they still miss how the parts connect. The energy an AI uses rises with the length of what it reads. Here is the fix you can use today. Give the assistant only the functions and files the task needs. Use search that knows the code's structure, like call graphs and symbols. Not plain chunks of text. Keep prompts and context as short as the task allows. GreenCode researchers found that a map of the code helped AI find the right parts. It did better than plain text search, on legacy code. GreenCode's Constitution skill can guide an AI coding agent to cut this waste. GreenCode helps tackle this by giving its own AI only the code it needs. It works from a searchable map of every interface, with small, fine-tuned models. That was GreenCode Tech Tip number forty. Follow for the next tip, and find more on greencode.ai/tech-tips.