As some of you may already know, over the summer I wrote a book/report for O'Reilly Media on AI agents and large language models. I am now pleased to announce that as of this month (November 2023) it has been published—you can find it here! As the book's title suggests, What Are AI Agents? When and How to Use LLM Agents is about when and how to use AI agents.

What Are AI Agents? (or WAAA? as I affectionately call it) is an introductory report for those who have not worked with AI Agents before. For some of us, it may be hard to recall a world without GenAI, but that world is only actually about a year old. Stable Diffusion, one of the first high-quality text-to-image open-source models, was only released in August 2022. ChatGPT was a fast-follow, released in November of that same year. In the year since, we saw the software industry change from approaching these tools as revolutionary but flawed, to a staple of the development process.

Still, most people haven't used AI agents, and far fewer still actually know what an AI agent is, let alone a large language model. That's what this report is about—introducing the uninitiated to what AI agents are and where they can be useful. WAAA? is not meant to be technically comprehensive. Instead, its purpose is to break down concepts in a high-level, approachable manner, focusing on ways people are using AI agents: whether it be as document retrieval mechanisms or coding assistants. It also addresses critical questions such as what AI agents are and how they are different than the large language models that power them.

By the end of the report, I hope I'll have convinced readers why AI agents are likely here to stay, and not some fad like other tech crazes of the past. Readers are also encouraged to look ahead and anticipate some problems to come with AI agents, including problems of bias, law, and economics.

If this sounds like a book/report you could benefit from, then I encourage you to read it. It's short enough to digest in an afternoon! You can find What are AI Agents? When and How to Use LLM Agents here and the Github repo associated with the report here.