A key approach for agent learning: accumulating insights in a LEARNINGS.md.

  • A key approach for agent learning: accumulating insights in a LEARNINGS.md.[ir]
    • After it does a task, have it distill insights that it gained that would have made this run easier.
    • This makes the next run faster and higher quality.
    • This is where the feedback loop closes and becomes a meta, compounding loop.[is]
    • That LEARNINGS.md is a file that a human can help curate and steer.
    • That's what gives you compounding quality even with a fixed model.

More on this topic

From other episodes