THE TOPOLOGY OF AN LLM AGENT LOOP WITH TOOL USE

Summary from the TLDR newsletter

An agent is three things: a model, an agentic loop, and a set of tools. Where these run has implications for cost, capability, effectiveness, and security. If you focus on topology, you can break down any agentic product and reason about what a product can and cannot do. There isn't one perfect loop - good UX for a general-purpose chatbot is not the same as good UX for a coding agent.