Llm-agents

GPT Function Calls come with native support for planning tasks in the presence of multiple available tasks, tools, and actions. This is an emergent behavior and greatly simplifies building LLM agents, over existing frameworks such as Langchain.

We do an honest review of the new function calls feature from OpenAI and analyze it from the perspective of building LLM agents. We explore the benefits of function calls, and how they can be used. Lastly, we mention some of the current limitations of function calls.

We explore why most ChatGPT Plugins and LLM agent projects feel underwhelming, and uncover some challenges developers face in building LLM agents. The challenges are mainly in four categories: 1) latency, 2) reliability, 3) UX, and 4) trust / security. We also argue that many LLM agent projects are doomed to fail because natural language is an inferior user interface to GUIs in many situations.