Skip to content

Understanding AI

Before you dive into tools and techniques, it’s worth spending 30 minutes here.

Most people’s frustration with AI comes from a gap in mental model — they expect it to work like a search engine, or like a person, and it’s neither. Once you understand what’s actually happening under the hood (at a conceptual level, not a technical one), the results improve dramatically.

PageWhat You’ll Learn
How AI Actually WorksWhat an LLM is, how it generates text, tokens, context windows, what AI is good and bad at, why hallucinations happen
Talking to AI Is a SkillHow to write prompts that get real results, from basic to advanced — with examples you can copy
Which AI Should I Use?Honest comparison of Claude, ChatGPT, and Gemini — when to use each, free vs. paid, open-source options

The single most useful thing you can take from this section:

AI output quality is mostly determined by input quality. The model is capable of much more than most people get from it — the gap is almost always in how the question was asked.

That’s not a criticism of new users. Nobody teaches this. But once you learn it, you’ll get 10x more value from the same tools.

Read How AI Actually Works first — it gives you the vocabulary and mental model. Then go to Talking to AI Is a Skill for the practical technique. If you want to compare your options before committing to a tool, Which AI Should I Use? covers that clearly.

!!! tip “You don’t need to memorize this” Skim it once. Come back when something isn’t working the way you expect. The mental model will make a lot more sense once you’ve used these tools for a week or two.