When working with a large language model (LLM) AI, there is typically an exchange of information. I provide some criteria and information via a prompt to have the AI do something on my behalf. This could be to research something, make something, or provide a counterpoint to challenge my thinking. The AI then, hopefully, responds appropriately. As the conversation grows, more context is provided and so on, but what happens when we get to a new conversation? How can gathered knowledge be shared and maintained?