In the second of our five-part series, I’m going to explain how the technology actually works.

The artificial intelligences that powers ChatGPT, Microsoft’s Bing chatbot and Google’s Bard can carry out humanlike conversations and write natural, fluid prose on an endless variety of topics. They can also perform complex tasks, from writing code to planning a kid’s birthday party.

But how does it all work? To answer that, we need to peek under the hood of something called a large language model — the type of A.I. that drives these systems.

Large language models, or L.L.M.s, are relatively new on the A.I. scene. The first ones appeared only about five years ago, and they weren’t very good. But today they can draft emails, presentations and memos and tutor you in a foreign language. Even more capabilities are sure to surface in the coming months and years, as the technology improves and Silicon Valley scrambles to cash in.

I’m going to walk you through setting up a large language model from scratch, simplifying things and leaving out a lot of hard math. Let’s pretend that we’re trying to build an L.L.M. to help you with replying to your emails. We’ll call it MailBot.

Step 1: Set a goal

Every A.I. system needs a goal. Researchers call this an objective function. It can be simple — for example, “win as many chess games as possible” — or complicated, like “predict the three-dimensional shapes of proteins, using only their amino acid sequences.”

Most large language models have the same basic objective function: Given a sequence of text, guess what comes next. We’ll give MailBot more specific goals later on, but let’s stick to that one for now.