Hi there, this is the first of a few posts trying to help folks who are trying to understand a little more about AI so that some of the jargon makes more sense. Specifically, this is targeted at folks who aren’t quite so technical (or perhaps not quite so technical as they used to be).
Let’s define a few things first. We’ve got an underlying technology today (that wasn’t particularly widely available or useful just a few years ago) that we can think of as the “brain” of any given AI Agent or Chatbot, and that’s a Large Language Model (LLM). The larger, and more advanced the LLM, the “smarter,” and more capable the chatbot or agent. We call the largest and most advanced LLMs “Frontier” models and there aren’t many companies at the forefront of this technology, OpenAI, Anthropic, Google, Meta to name a few.
Why bother to differentiate the LLM from the Chatbot or Agent that it runs on? Chatbots are the applications we’re interacting with on our phones or in our web browsers (examples include ChatGPT.com or the Gemini app on your smartphone) that use LLMs to figure out how to respond to your prompts (we’ll cover prompts in a later post). What this also means is that other Chatbots you interact with may also be using the same LLMs (for example, the support bot in an app from another company).
So what’s an Agent? According to Anthropic, “Agents…are systems where LLMs dynamically direct their own processes and tool usage, maintaining control over how they accomplish tasks.” This feels a little technical, so I’ll try to simplify it. An agent is an application that can accomplish one or more tasks, using an LLM to figure out how to accomplish those tasks.
I hope you’ve enjoyed this short post written by a human. Next time we’ll tackle Open Source vs Open Weight vs Proprietary LLMs!
You may also like
