Two things at once: learning to run a project that doesn’t fit in a single file, and finishing your capstone.
Your capstone has to be done by the end of today.
Session 4 is 60 minutes and it is entirely demos. There is no build time. Whatever runs when you leave today is what you demo.
Plan around that. If you’re behind at the halfway mark, cut a feature — don’t hope.
| Block | |
|---|---|
| 1 | Why one file stops working |
| 2 | Plan before you build — revising your PLAN.md |
| 3 | Claude Code for real projects |
| 4 | Break |
| 5 | Save points — copy the folder |
| 6 | Capstone build (the big block) |
| 7 | Debugging clinic |
| 8 | Demo prep + wrap |
cd $HOME\Documents\Projects
cd <Name>
cd capstone
claude
<Name> is your name — the folder you made in Session 1.
Run it: python main.py
Ctrl+C quits · /clear resets the conversation without touching your code · one
command per line, no &&
Once an app passes ~200 lines, two things go wrong — and only one of them is the AI’s problem.
1. You can’t find anything. That’s yours. You cannot verify what you cannot locate.
2. The AI has to hold the entire file in its head to change one line. Every AI has a limited amount it can attend to at once — its context window. A 500-line file eats a lot of it. A 2000-line file eats most of it, and output quality visibly drops.
It’s not that the AI gets dumber. It’s reading a novel to answer a question about one paragraph. Give it the paragraph.
This also explains something you’ve probably already noticed: long sessions get worse.
/clear fixes it.
quiz/
main.py ← starts the app, wires things together
questions.py ← the question data
scoring.py ← how points work
display.py ← the window and buttons
Now “change the scoring” means the AI reads scoring.py — faster, more accurate, and
you know exactly where to look to check it. That second part is the real payoff.
The highest-leverage move in this entire course:
I want to build [your app].
Don't write any code yet. Tell me what files you'd create, what goes
in each one, and what order you'd build them in.
You’ll get a file breakdown and a build order in seconds. Now the important part:
You just made five architecture decisions in ninety seconds without writing a line of code. Changing a plan is free. Changing code is not.
Let's build step 1 only: questions.py with 10 multiple choice questions.
Nothing else yet.
Run it. Check it. Then step 2. This is Session 2’s one-change-at-a-time rule, scaled up from changes to components.
PLAN.md is your checklistYou wrote one last session. Open it and pressure-test it against what you now know about splitting files:
Given this plan, what files should this project have? Should any of
what you suggested be split or combined?
Change at least one thing. Save it. This is the last planning you do today — the rest of the session is building.
Look at scoring.py and tell me what it does.
In display.py, make the score label bigger.
The bug is in main.py, not the other files.
Naming the file focuses the AI and saves it reading everything. Faster and more accurate.
/clear — reset the conversation/clear
Fresh conversation, same folder, same files. Your code is untouched.
Rule of thumb: when you switch to a genuinely different task,
/clearfirst.
CLAUDE.md — notes the AI reads every timeMake a file called CLAUDE.md in your project folder. Claude Code reads it automatically
at the start of every session.
## Quiz Game
A multiple-choice quiz in Python + tkinter.
## Files
- main.py — starts the app
- questions.py — question data, list of dicts
- scoring.py — 1 point right, -1 wrong, no negative totals
- display.py — the tkinter window
## Rules
- Python only, tkinter only. No extra libraries.
- Keep every file under 100 lines.
- Answer checking is case-insensitive.
Now you never re-explain your project. Every new session already knows.
The difference between an intern you brief every morning and one who read the handbook.
Write one for your capstone before the build block. Two minutes, immediate payoff.
What would it take to add a timer to each question?
Is there anything in this project that's going to cause problems later?
Free consulting, no code changes. Use it constantly.
After a change, scroll back through the output. It shows you the edits. Skipping past that is throwing away your best verification tool.
SAVE — every time it works:
cd $HOME\Documents\Projects
cd <Name>
Copy-Item -Recurse capstone capstone-working
RESTORE:
cd $HOME\Documents\Projects
cd <Name>
Remove-Item -Recurse capstone
Copy-Item -Recurse capstone-working capstone
Name copies meaningfully — capstone-scoring-works, not capstone2.
“Being used by another process” → your app is still running or the AI is open in that
folder. Close the window, Ctrl+C, retry.
Argument order is Copy-Item -Recurse SOURCE DESTINATION. Reversing it copies the
broken version over your good one.
Don’t copy into the project folder. cd $HOME\Documents\Projects, then
cd <Name>, so copies sit beside your project, not inside it. A copy inside the
folder gets read by the AI as part of your project and causes genuinely bizarre problems.
This is check #3 from last session — “did it change anything I didn’t ask for?” — made concrete.
Before a risky change, copy the folder. Afterward, open the old file and the new file side by side and compare. Two windows, and you have a diff done by hand.
You can also ask:
Compare the current version of scoring.py to what it was before my
last request. What actually changed?
You now have the ability to answer “what did it actually change?” with certainty instead of trust. Use it before anything that scares you.
CLAUDE.md describing the projectPLAN.mdbuild → run → check → copy
README.md explaining what it does and how to run it20 minutes on the same problem = restore your last working copy and take a different approach. Sunk cost is real. Back out.
In scoring.py, when the answer is right the score goes down instead of up. Terminal output: [paste]. What’s happening?
Step 5 is the one everybody skips.
/clear and describe the problem fresh — the conversation may have a bad theory stuck
in itSession 4 is 60 minutes of demos with no build time. Whatever runs now is what you show.
cd $HOME\Documents\Projects
cd <Name>
Copy-Item -Recurse capstone capstone-demo
If you tinker at home* and break something, this one survives.
* Tinkering at home needs an AI coding tool. Claude Code, the tool you used in class, is a paid product that needs an account. Some tools have free options with no account needed — see the OpenCode Zen how-to.
1. What it is, in one sentence.
2. Show it working. (Practice the exact clicks. Twice.)
3. One thing that broke and how you fixed it.
4. One thing the AI got wrong that you caught.
5. What you'd add next.
Practice the clicks. The most common demo failure is clicking around live, hitting a bug you’ve never seen, and losing your nerve. Rehearsing the path twice fixes it.
If it crashes on stage: say what it does and keep going.
Plan before you build. Split by job. Copy every time it works.
Your project is on this computer. Save a copy to your Google Drive folder:
This Vibe Coding Curriculum by Brian Bird, created in with AI assistance, is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.