Episode 2- How large language models (LLMs) actually work (at a high level)

LLMs feel intelligent. But under the hood, they’re not thinking. They’re predicting.
In this episode, we explain in plain English how large language models like ChatGPT actually work. What they’re trained on, what a “token” really is, how probabilities drive every word they generate, and why they sometimes sound confident but still get things wrong.
We break down the core mechanics step by step: training on massive text data, learning patterns, predicting the next word, and using context to stay coherent. No math. No jargon. Just the mental model you need to reason clearly about what these systems can and cannot do.
By the end, you’ll understand their strengths, limits, and how to use them wisely in real products and workflows.