This post discusses two methods an agent can use to process its memory: 'recall', which retrieves relevant facts instantly without the need for a language model, and 'reflect', which synthesizes insights from those facts. It explains when to use each method effectively.