What AI and LLMs actually are

Let's demystify this. The AI everyone's talking about — ChatGPT, Claude, Gemini — is powered by something called a large language model (LLM). The single best way to understand it: it's an extraordinarily sophisticated autocomplete.

You've seen your phone suggest the next word as you text. An LLM does the same thing, but trained on a massive portion of the internet, books, and writing — so it can continue almost any piece of text convincingly. Here's the whole mechanism, no math:

• Text is broken into small pieces (roughly word-fragments).
• The model has one core skill: predict the next piece, over and over. It writes a bit, re-reads everything including what it just wrote, predicts the next bit, and repeats. A whole paragraph is hundreds of these tiny predictions chained together.
• It learned this by reading enormous amounts of text and guessing the next word billions of times, getting corrected each time. Along the way it absorbed grammar, facts, styles, and patterns that look like reasoning.

The one idea to hold onto: it's predicting plausible text, not looking up true text. That single fact explains nearly everything about how it behaves — why it's a brilliant writer, why it sometimes confidently makes things up, and why it can fumble simple math. It's not consulting a database of facts; it's generating what sounds right based on patterns.

"Generative AI" just means AI that creates content — text, images, audio, video — rather than only sorting or labeling. An LLM is the text-and-code kind.

Why this metaphor matters for you as a user: once you think of AI as a supercharged autocomplete that produces plausible output, its strengths and weaknesses stop being surprising. It's fantastic at anything language-shaped (writing, explaining, summarizing) and unreliable at anything requiring it to know something is true. Keep the "sophisticated autocomplete predicting the next word" picture in mind through the whole course — it's the foundation everything else builds on, and it's genuinely most of what you need to understand to use AI well.