insideout Get Pro

12 · Several agents on one task

Pro

Agents, lesson 12

Several agents on one task

When splitting work across agents helps, and what it costs.

You know the single agent loop, and the orchestrator-workers workflow from "Fixed workflow or open loop".

Part 1. Start with one agent

Splitting work across several agents sounds like a team, and teams sound good. But each extra agent is another loop with its own prompt, its own context and its own ways to go wrong, plus the work of passing results between them. Most tasks are better done by one agent with good tools.

A few situations do push past one agent. The material won't fit in one context window (reading 50 files to answer a question). The work splits into independent parts that can run at the same time. Or parts of the job need very different tools and instructions that would confuse one prompt.

Your turn

Which task is the best case for several agents?

Six independent research jobs, each reading far more than fits comfortably in one context, run well in parallel and only need short summaries back. The other three are small, single-thread jobs where a second agent adds cost and nothing else.

Part 2. A lead and its helpers

Keep going with Pro

You've done the first part. Pro opens the rest of this lesson, and every other one.

Still to come
  1. A lead and its helpers
  2. What it costs
  3. How they fail together
See Pro plansI have Pro: sign in
Next lessonAn agent that runs experimentsAn agent edits code, runs a fixed-budget experiment and keeps what helps. Start lesson 13

Agents overviewSpotted a mistake? Tell us.