🧪 Stop & play-test
After the AI adds something, stop and play. Try the new thing.
- Does the new thing work?
- Can I actually use it?
- Did anything else break?
Don’t just say “it works”. Actually test it.
You bring the ideas. The AI helps build them. Your most important job is to actually play the game after changes are made.
These are not homework rules. They are small habits that stop an AI-built project from turning into 3,000 mysterious lines before anyone notices.
After the AI adds something, stop and play. Try the new thing.
Don’t just say “it works”. Actually test it.
Before a big change, save a copy of the version that works.
Example: PoorFighter-working.html
If the new version goes crazy, you have something safe to return to. You do not need a new copy after every tiny change.
You do not need fancy prompts.
“Add a shop where I can buy weapons.”
If something breaks, say what you saw. If you are confused, ask “Where do I put this?” or “What does this part do?”
Try not to add 15 major features at once.
Small steps make bugs much easier to find.
You do not need to know the solution. Tell the AI what you did, what you expected and what actually happened.
Observable behaviour is enough to start debugging.
Stop adding new features for a moment.
“Let’s stop and make sure the current game is organised and working before adding anything else.”
The AI can help clean things up.
You do not need perfect prompts. You do not need to understand every line of code.
You do not need special programming vocabulary. Give the AI three pieces of evidence and let it help investigate.
That’s enough to begin. You are reporting evidence, not sitting an exam.
The useful skill is not sounding like a programmer. It is being clear about what you want, what you saw and what you do not understand.
Add a shop where I can buy weapons.
I want a boss that shoots fireballs.
The boss appears, but I can’t damage it.
Where do I put this? What does this part do?
Have ideas. Make things. Play them. Fix them. Keep building. 🚀