Session promise

By the end of this lesson, AI should feel less like magic and more like a system you can reason about. You do not need to know every technical detail before you start using it well. You need a clear mental model for what is happening.

The simple picture

An AI model is a pattern engine trained on huge amounts of examples. When you give it input, it predicts useful next tokens, pixels, code, or structure based on patterns it has learned. The application around the model turns that raw capability into something you can use, such as ChatGPT, Gemini, Cursor, NotebookLM, or an API.

Model: the reasoning engine that has learned patterns.
Application: the product interface around the model.
Prompt: the task, context, examples, and constraints you give it.
Output: a draft that still needs human judgment.

Why the tool feels confusing

The tool arrived before most people received a practical explanation. That creates two common reactions: overtrusting the output because it sounds confident, or avoiding the tool because it feels too technical. Both reactions slow you down.

The goal is not to become a machine. The goal is to understand the machine well enough to direct it.

Practice prompt

Ask an AI assistant: "Explain how large language models work to a beginner developer using an analogy, then list three things the analogy gets wrong." Compare the answer to the mental model above.