2 Chat: "Explain this code" is the worst prompt in AI
This chapter covers
- How to ask sharper follow-ups instead of "explain this code"
- Serializing a whole repo into one file you can paste into chat
- Constraining a prompt to get a story, numbers, or a toy example
- The forgetting curve: why you lose two-thirds of it by tomorrow
- Chat vs workflow vs agent: when to reach for which
When developers first try using AI with code, they usually paste a file and ask, "What does this do?" The AI then returns generic text that could describe any code. They try again, get a similar answer, and conclude the tool is overhyped. The problem isn't the model; it's the question.
2.1 The first answer is always generic. That's the point.
2.1.1 Which model should I pick?
2.2 One file can't answer cross-file questions. 20 lines fix that.
2.2.1 What you cut matters as much as what you keep
2.3 A concrete prompt beats a vague one, every time
2.3.1 Ask for a story, not a category
2.3.2 Constrain by numbers, not adjectives
2.3.3 Ask for toy examples, not prose
2.4 You'll forget two-thirds by tomorrow
2.4.1 Let AI write the first draft. It's disposable.
2.4.2 Follow up on the line you think you understand
2.4.3 The step everyone skips is the only one that works
2.4.4 Re-reading feels like learning. It isn't.
2.5 Chat can't scale past five files. That's not a bug.
2.5.1 What's a workflow?
2.5.2 What's an agent?
2.5.3 Chained together, each tool covers the next one's weakness
2.6 Summary