Chapter 10. Agent Reasoning Patterns
Chapter 9 installed the vocabulary of decision making. Decision graphs, hypothesis weights, grounding, calibration, trade-off vectors, decision budgets, Why Traces. The vocabulary is precise enough to describe any reasoning the architecture does, but precision alone does not tell a team how their agents should be reasoning. What is missing, and what this chapter installs, is the shapes that reasoning takes in practice: the recurring patterns experienced engineers recognize and reach for when faced with characteristic classes of decision. Patterns are to Chapter 9’s vocabulary what design patterns are to object-oriented programming: a smaller, more memorable catalog of recognized structures that emerge again and again when the underlying language is applied to real work.
The reason this chapter is organized around patterns rather than algorithms is structural. Algorithms describe specific implementations; they live and die with the technology stack they run on. A reasoning pattern, by contrast, describes the shape of ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access