LLMs have demonstrated an amazing ability to interact with humans using natural language, which has opened the door to many new applications. But how can LLMs also interact with our existing software infrastructure? For example, by letting it decide when to make a function call to a different program to get more information call to a different program to get more information or to take an action. LLMs were originally designed to generate text for humans, but some LLMs have now been trained to output formatted data, such as values stored as JSON, to make it easy to let the LLM decide when to call other code as a subroutine. This significantly expands what you can do with LMs, such as let them extract information from structured or tabular data which LMs have typically struggled with. Here to tell you more about this is Harrison Chase, co-founder and CEO of Landchain, who has also taught two previous short courses. Welcome back, Harrison. Thanks, Andrew. So great to be back. You're right, this new capability which OpenAI has named function Harrison, you've described Langchain in other courses, but maybe you could describe what has changed and what you'd be covering in this course. Absolutely. As you know, Langchain is an open source library that helps developers bridge the gap between traditional software and LLMs. It allows developers to support any number of different LLMs and provides over 500 integrations to different language models, vector stores, and tools, as well as supporting memory chains and agents. There are two significant changes that you will be getting into during the course. The second is changes to take advantage of the new function calling capability. You'll learn how to use that directly and we'll also show how it can be used to do tasks like tagging or extracting data. Function Calling makes building tools for LLMs simpler and more reliable. You'll build some tools, and then, use them to build a conversational agent. You get to use all of those elements in the course and in the final project. That sounds great, Harrison. And I think, Many people have worked to make this course possible. We're grateful to Lance Martin and Nuno Kampas from Landchain. And on the deeplearning.ai side, Jeff Lodwick and Eshmel Gagari also contributed to this course.