Read this first

1. What this course is

This is a short course on writing prompts that work — prompts you can predict the output of before you press enter, and verify after.

It is free, which means a few things. First: there's no upsell at the end. Second: nothing here is rationed; you can read all of it in a weekend. Third: it exists because most prompt advice is either marketing fluff or specific to whichever model was hot the month it was written, and that's worth being loud about even when no one is paying.

2. Who it is for

You, if you have ever:

• Typed make this better into an AI and watched it confidently produce something worse.
• Pasted two hundred lines of code with why doesn't this work? and gotten a generic answer that missed the real bug.
• Felt your prompts working sometimes and failing other times, with no clear reason why.
• Tried "prompt engineering" tips that worked once and never again.
• Watched a senior engineer get a useful answer from the same model that gave you garbage — and couldn't tell what they did differently.

If none of that applies to you, you already prompt the way this course is going to teach. Read it for entertainment.

3. What it is not

This is not a list of "magic phrases" to paste at the start of every prompt. It is not about jailbreaking, role-play tricks, or prompt-injection security. It does not name specific models, because the underlying skill outlasts every version number.

What it has: the mental model behind a prompt that works, the anatomy of one (in five parts), and worked side-by-side examples of vague-vs-precise prompts on real software engineering tasks — code review, debugging, refactoring, writing tests, writing PRs.

4. The next page

The next page is the whole course in five examples — five before/after prompt pairs that summarize everything that follows. If you read them and the difference between each pair feels obvious to you and you already write the second column by default, you can leave.