Build something by describing it
Outcome
By the end, you'll have used Claude Code to turn a plain-language description of a small program into working code you ran yourself.
Concept
You don't need to know how to code to start building with Claude Code — you need to know how to describe what you want clearly (the same skill from Tutorials 2 and 4), read what comes back, and ask follow-up questions when something is unclear. This is the same conversation loop from Track 1, just pointed at a tool that can actually build things. And the verify-and-question habits from Track 2 are exactly what keep you from running code you don't understand.
Weak approach vs. strong approach
The ask
make me a program
What comes back
I'd be happy to help! Here's a Python program — it's a simple to-do list manager with add, remove, and list commands, persistent file storage, a command-line menu, color-coded output, and an undo feature. The code is 180 lines. Want me to add anything? (Note: you'll need to install the 'colorama' package first.)
Try it
In your terminal, with Claude Code running (`claude`), try this task:
Steps
- Describe this program in your own words first: "Build me a temperature converter that asks for a number in Fahrenheit, converts it to Celsius, and prints the result like '72°F is 22.2°C.'"
- Read every line of the code Claude Code writes. If anything is unclear, ask it to explain that specific part before moving on.
- Run the program the way Claude Code suggests, and try a few different input values.
- Ask for one change — a different unit, a friendlier message, an extra check — and watch Claude Code update the file directly.
Takeaway
Describing what you want clearly, reading what comes back, and asking for one more change — that's the whole loop, and you just used it to build something real. But it worked once, on the one example you tried: that's a demo, not yet a tool. Closing the gap between the two is where the real work starts — and it's what comes next.