A harnessed LLM agent, clearly explained!

This page discusses the architecture of a harnessed LLM agent, emphasizing a model design that inverts traditional relationships by making the model itself thin while pushing intelligence outward. It aims to clarify the composition of intelligence at runtime.