Skip to slide
086 / 170

CH 09 · Memory Outside the Weights · 02/06

Read as one page

Why it matters

Memory is how an agent gets better at your project specifically without anyone retraining the model. It is the difference between an agent that re-learns your test command every session and one that just knows it. It is also where reliability comes from: an anti-hallucination constraint written into shared rules ("never invent API names; grep the codebase first") applies to every session forever. And because it is all plain files in the repo, it is portable, reviewable, and version-controlled, which is exactly what you want for something that shapes how an agent behaves on your code.

Finished this idea?

Your learning progress stays in this browser.