An Analogy Between LLM Agents and the Human Brain

Nice framework. Makes me wonder if the future is better external memory, or models that just handle context natively without the retrieval layer.