Always Ask What Changed

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-vibecoder-ask-for-the-diff

1. A developer asks an AI agent to refactor their entire codebase in one prompt. What is the primary risk of this approach?

   1. The AI might accidentally rewrite code that was already working correctly across many files
   2. The AI might refuse to work on large projects
   3. The AI might add too few comments to the code
   4. The AI might use deprecated programming languages

2. What does 'write the smallest useful scope' mean when working with an AI coding assistant?

   1. Ask the AI to complete one discrete, testable task at a time before moving to larger changes
   2. Write the minimum amount of code possible to pass tests
   3. Write code that uses the smallest amount of memory
   4. Use the smallest number of files in your project

3. Before sharing AI-generated code with a team, what three things should you inspect?

   1. The diff, data access patterns, and the failure path
   2. The file size, creation date, and author
   3. The AI's confidence level, token count, and version number
   4. The color scheme, variable names, and comments

4. Why should you run the result as a user rather than as a fan of the tool?

   1. To verify the code actually works in real usage scenarios, not just that the tool claims it works
   2. Because fans of the tool will be biased toward positive results
   3. Because AI tools are always correct and need no testing
   4. Because running as a user is faster than running as a developer

5. What is a regression in the context of code changes?

   1. A backward compatibility mode
   2. A function that calculates mathematical values
   3. A type of AI model used for coding
   4. A new feature that breaks existing functionality that previously worked

6. What does the diff summary show you about AI-generated changes?

   1. The AI's confidence score for each change
   2. Which specific lines were added, removed, or modified
   3. How much the code will cost to run
   4. Exactly how long the AI spent processing the request

7. What makes code 'reversible' in this AI-assisted coding framework?

   1. Code that automatically fixes itself when errors happen
   2. Code that can be compiled on any operating system
   3. Code that can be easily undone or rolled back if problems occur
   4. Code that runs faster than previous versions

8. Why is 'naming the job before naming the tool' recommended?

   1. Because jobs are more important than tools in all contexts
   2. Because tool names are copyrighted and need permission
   3. Because clearly defining what you want to accomplish leads to better results than jumping straight to choosing a tool
   4. Because AI tools only work with specific names

9. What does the failure path refer to in code inspection?

   1. What happens when the code doesn't work—how it fails and what errors occur
   2. The path where code is stored on the computer
   3. How the code handles successful execution
   4. The folder structure of the project

10. What distinguishes a working demo from production-ready code in this approach?

    1. Production code is always shorter
    2. Demo code runs faster
    3. Production code is observable, reversible, and safe enough for others to use
    4. Demo code uses more comments

11. What should a test prove about a code change?

    1. That the AI wrote the code correctly
    2. That the change actually works and solves the intended problem
    3. That the code uses the latest programming patterns
    4. That the code looks clean and follows style guidelines

12. Why do practitioners on X judge AI-assisted builders by what happens between prompt and commit?

    1. Because commits are no longer relevant
    2. Because commits are the only thing that matters
    3. Because prompts are more important than code
    4. Because the quality of inspection, testing, and change rejection shows real skill versus just blindly accepting AI output

13. What does it mean for code to be 'observable' in this framework?

    1. That the code uses observation patterns
    2. That the code is visible to everyone
    3. That the code's behavior can be monitored and understood through logging, testing, and debugging
    4. That the code can be watched by users

14. When reviewing AI-generated changes, what does inspecting data access reveal?

    1. What data the code reads from or writes to, and whether it accesses anything it shouldn't
    2. What programming language the AI chose
    3. How fast the database queries run
    4. How much memory the code uses

15. What is the purpose of having a rollback path?

    1. To compare old and new versions of code
    2. To permanently delete failed code
    3. To be able to revert to a previous working state if the new code causes problems
    4. To speed up the development process