Stage 6 of 10 · 6 lessons · ~2 weeks

Debugging with AI

Use AI to find causes rather than hide symptoms: reproduce first, shrink the problem, give full evidence, test hypotheses, fix the root cause with a regression test and keep using the real debugger.

Start: Reproduce, then shrink the problem

Lessons· one lesson for each topic, in this order

Do this

Reproduce a bug reliably before asking for help.
Shrink one bug to a minimal example.
Ask for ranked hypotheses and test at least two before changing code.
Reject one symptom patch and find the root cause instead.
Fix a bug using breakpoints in your editor's debugger.
Keep a regression test and a two-line note for every bug you fix.

Start with

CourseFree
The Missing Semester of Your CS Education
MIT · Beginner
MIT's free lecture on debugging and profiling covers logging, debuggers and tools you will use for years

Keep handy

How to Ask Good Questions (Julia Evans)
How to ask a precise question with the right evidence, which works for people and AI tools alike
AI in VS Code Documentation
Fixing failing tests, /fix and attaching terminal output in VS Code

Also worth your time

A debugging manifesto (opens in a new tab)Read
Short, practical principles: trust the evidence, investigate before fixing, learn from each bug
Debugging: The 9 Indispensable RulesGo deeperPaid
David Agans' nine rules, from "make it fail" to "if you didn't fix it, it ain't fixed"

Before moving on: Three real bugs fixed with your tool where you reproduced first, tested hypotheses, found the root cause with evidence, used the debugger at least once, and left a regression test and a short note for each.

Tauseef's advice

Use AI to widen your list of suspects, not to skip the investigation. Ask "what could cause this and how would I check?" far more often than "fix this". The engineers people trust are the ones who can explain why something broke.