The Ops behind ChatGPT

ChatGPT – in the time that I have been writing this book – has gone from being a hot topic to just being a fact of life, sometimes almost second nature as a tool for information. The way it handles data and the very nature of it have been topics that have brought on a lot of controversy. But one of the things that I get asked very often by my friends who aren’t in the industry is, how does it work? They see that it delivers information nearly seamlessly on whatever the whim of the user is and then retains that information historically in that chat for future questions. It also does so very quickly. So, one does wonder how it all works.

Let’s start with what ChatGPT is: it is a large language model (LLM), which is a very large neural network that verges on general language understanding (i.e., the ability to understand questions or queries and deliver back responses that would be appropriate to the solution). While ChatGPT is the...