دانلود کتاب حافظه عامل – ساخت عامل‌های هوش مصنوعی حالت‌مند که به یاد می‌آورند، سازگار می‌شوند و در طول زمان کار می‌کنند

Do you remember the first time you used ChatGPT? I can recall talking with my colleagues about how Promethean the technology was. I can remember how scary yet oddly normal interacting with it felt. But I can’t remember the specific moment I started using it, what I asked it, or even if the answer I was given was particularly novel. Like the millions of input and output tokens that I have passed through LLMs since, my first interactions with ChatGPT were almost certainly unremarkable. What we want from these AI tools has expanded as fast as their capabilities have. We now expect them to use tools themselves, to carry context from one interaction to the next, to build on what they know about us and our work. But how does that actually happen? The answer is memory. Not of the biological sort that holds our dearest, most ephemeral experiences, but something more mechanical and pragmatic. Agent memory is a system for storing, retrieving, updating, and forgetting the knowledge an agent carries forward, so it can act on the right information at the right time. That’s what separates an agent that is generically helpful from one that is specifically helpful to you, your team, and your organization. This chapter shows this difference in practice by examining the scenario of a single, fixed twice: First by an agent working from a blank slate, and then by an agent wired into its organization’s existing knowledge. From there you’ll learn to delineate the boundary between the underlying model and the agent harness around it, distinguish the context an agent holds right now from the memory it carries across time.