What is a "cognitive architecture"?

The term "cognitive architecture" refers to the structure of a system's thinking process, including the flow of code, prompts, and LLM calls that transform user input into actions or responses. This concept is crucial in designing autonomous systems powered by large language models (LLMs). There are different levels of autonomy in LLM applications, ranging from hard-coded systems to fully autonomous agents. The choice of cognitive architecture depends on the specific task and desired level of control over the system's actions. LangChain and LangGraph are tools designed to enable developers to experiment with various cognitive architectures while building LLM applications.