THE THREE DISCIPLINES AROUND THE MODEL

LEARNING GOALS By the end of this chapter you will be able to define prompt engineering, context engineering, and harness engineering. You will understand how they differ and how they nest together. You will be able to explain why, in real systems, the harness often matters more than the model. Why Three Disciplines? When people first work with language models, they discover quickly that the quality of the result depends on far more than the model itself. Over time, three professional skills have emerged around the model, each focusing on a different circle of control. You can picture three circles, one inside another. The smallest circle is the single instruction. The middle circle is everything the model sees across a whole interaction. The largest circle is everything that surrounds the model: the software, the rules, the tools, and the safety checks. Understanding these circles helps you diagnose problems. When something goes wrong, the first question a good engineer asks is: which circle did this problem come from? The answer decides who fixes it and how.