Definition
Plain language
A method that lets an AI agent improve by keeping a running list of useful facts it can look back at, without retraining.
As stated in the literature
A label-free self-improvement approach that accumulates a memory of useful facts and notes across tasks, used as a memory-only baseline against harness-rewriting methods like RHO.
Why it matters: It lets an AI agent get better over time by remembering useful facts, with no costly retraining required.
For example, an agent solving a series of problems might jot down a handy formula it discovered and consult that note when a later problem needs it.
Heard on the show
“Because the label-free self-improvement methods that came before this — things like Dynamic Cheatsheet, ReasoningBank — they basically accumulate memory.”Episode 120 — How an AI Agent Rewrites Its Own Tools, Without an Answer Key