About

These are moments from real working sessions with an AI coding assistant, where its first answer was reasonable but human expertise changed the outcome. They are written up as teaching material for students learning to work with these tools.

How to read a moment

Every moment follows the same structure:

• Context: the project or task being worked on.
• The Initial Ask: the prompt that started the exchange.
• Claude’s Initial Response: what the assistant produced first.
• The Exchange: every turn of correction and revision, including the false starts.
• The Final Solution: what was actually used.
• The Lesson: what the assistant got right, what took human expertise, why the assistant missed it, and one takeaway to remember.

The why is the most useful part. It names the structural reason the gap happened, so you can spot the same gap in your own sessions.

How they are made

Each moment is drafted with the capture-learning-moment Claude Code skill, from the conversation it describes. Hostnames, IP addresses, usernames and other identifying values are replaced with placeholders.