When a pipeline beats an agent: three shapes that don't need a loop
Originally published on Loop & Retry — field notes on building LLM agents that survive production. When not to build an agent made the case in the abstract: an agent is an LLM that controls its own control flow, and that control costs you quadratic tokens, serial latency, and a failure surface no unit test can cover, on every single run. What that post didn't give you is the thing you reach for…
The article titled "When a pipeline beats an agent: three shapes that don't need a loop" discusses three alternative methods to building LLM agents that control their own control flow. These methods involve fixed sequences of calls with no model-decided branching, resulting in a flowchart that can be drawn before executing a single step.
This is the key distinction between these shapes and traditional agents, which often require quadratic tokens, serial latency, and a failure surface that cannot be covered by unit tests. The three pipeline shapes presented are: the linear chain, the router plus fixed handlers, and the fixed sequence with optional branching.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
