Skip to content
Hi, Bot
All lessons

Unit 5 · Lesson 3

Ship and test your prototype

Build the smallest version, then hunt for where it breaks.

~20 min

You have a problem and a stack. Time to build the thing — and here's the rule that separates people who ship from people who plan forever:

Build the smallest version that could possibly work. Then try to break it.

That first rough version has a name: a prototype. It's not supposed to be perfect. It's supposed to exist, so you can test it on reality instead of arguing with yourself.

A rough thing you tested beats a perfect thing you imagined. Ship small, then break it on purpose.

One idea: prototype, then stress-test

You already know how to test — you did it to the cat-detector in Unit 2. Same instinct, now aimed at your own creation.

PhaseWhat you do
1. Build tinyMake the simplest version of your stack. Skip the fancy parts.
2. Run it realTry it on 3 actual cases from your life — not made-up perfect ones.
3. Break itFeed it the weird case, the edge case, the thing you expect to fail.
4. Fix onePick the biggest break and fix just that. Then test again.

The loop — build, run, break, fix — is the entire secret of how real things get made. Not one giant perfect launch. Many small rounds, each a little less broken than the last.

Do the thing

Build the tiny version of your project. If your stack used an AI step, that might just be a really good prompt (you have a Prompt Card now — use it).

Then test it on three real cases, and deliberately throw one edge case at it. Fill this in:

  • Case 1 (normal): Did it work? ______
  • Case 2 (normal): Did it work? ______
  • Case 3 (normal): Did it work? ______
  • Edge case (weird on purpose): How did it break? ______
  • The one fix I'd make next: ______

Quick check. If nothing broke, your test cases were too easy — a prototype that passes every test hasn't really been tested. Push harder: the messiest, longest, weirdest real input you can find. Finding the break isn't failure. It's the information that tells you what to build next.

Why this matters

You finished Unit 5. You didn't just use AI — you designed with it.

  • Shipping small beats planning big. The prototype turns opinions into evidence.
  • Build, run, break, fix — forever. Every app, game, and tool you love got there this way, one round at a time.

One unit left, and it's the one that keeps all this power pointed in a good direction: privacy, power, and you — what to share, what never to, and the personal rules you'll write for yourself.

Reflect & continue

One last thing.

The reflection sticks the lesson. One sentence is plenty.

Skip for now