AI coding: large migrations with checkpoint commits

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-ai-coding-migrations-with-checkpoints-r7a1-creators

1. Why do large, single-step AI migrations often fail when moving between major framework versions?

   1. AI cannot access the internet to download new dependencies
   2. Single-step migrations exceed GitHub's file size limits
   3. Errors compound across multiple files and become difficult to trace
   4. The AI forgets the migration goal halfway through

2. What is the primary purpose of creating a 'checkpoint' during an AI-assisted migration?

   1. To generate documentation for the changes made
   2. To mark where the migration is officially complete
   3. To create a compilable, testable state that can be rolled back to if needed
   4. To save the AI's context window for faster processing

3. When prompting an AI to perform a migration checkpoint, which element is most likely to cause problems if omitted?

   1. A list of all files in the project
   2. The current date and time
   3. A polite greeting
   4. Clear boundaries specifying what NOT to change

4. Which task is an AI LEAST capable of handling reliably during a multi-file migration?

   1. Replacing string literals with constants
   2. Deciding whether a behavior change is acceptable for the project
   3. Finding all instances of a deprecated function call
   4. Updating import statements consistently across 20 files

5. A developer runs an AI migration prompt and the code fails to compile. What should they do before retrying?

   1. Commit the broken state to preserve the AI's work
   2. Delete all files and start over from scratch
   3. Revert to the last checkpoint and analyze what went wrong
   4. Ask the AI to fix both the compilation error and continue the migration

6. What information should be included in a well-structured migration prompt for an AI?

   1. Every file that will ever need changing during the entire migration
   2. The migration direction, specific scope, and what must compile/pass after
   3. A detailed history of why the migration is needed
   4. Only the target framework version number

7. Why is it important for each checkpoint to pass tests before proceeding to the next prompt?

   1. The CI/CD pipeline requires passing tests to continue
   2. Tests verify that the AI is following instructions exactly
   3. Tests make the migration run faster
   4. It ensures the migration is reversible and each step is verified

8. What limitation of AI models directly motivates the checkpoint approach?

   1. AI cannot hold the entire context of a large migration in working memory
   2. AI can only edit one file at a time
   3. AI models cannot read files larger than 1MB
   4. AI always produces syntactically correct code

9. When specifying scope for a migration checkpoint, what does saying 'do not touch X' accomplish?

   1. X will be automatically deleted to reduce migration complexity
   2. The migration will skip X and do Y instead
   3. It creates a boundary that helps isolate changes and simplifies debugging
   4. The AI will automatically skip X and finish faster

10. Which type of change is an AI most reliable at performing across many files?

    1. Changing business logic that affects user behavior
    2. Refactoring code to improve performance
    3. Applying consistent mechanical transformations like import updates
    4. Deciding which features to keep or remove

11. What happens if you skip committing before running an AI migration prompt?

    1. The migration will run faster
    2. The AI will refuse to make changes
    3. Nothing significant—it works the same either way
    4. You lose the ability to easily roll back to a known good state

12. A developer wants to migrate from React class components to functional components across 40 files. What is the best approach?

    1. Ask the AI to convert all components in a single file first
    2. Write a script to convert all files manually instead of using AI
    3. Break into checkpoints by converting a few related components at a time with testing between
    4. One prompt to convert all 40 files at once

13. Why might an AI miss logic regressions during a migration even if the code compiles?

    1. The AI intentionally changes logic to improve code
    2. The AI cannot run the code to observe behavior
    3. Compiling successfully guarantees logic is correct
    4. The AI doesn't have access to the test suite

14. What is the relationship between the scope of a checkpoint and the likelihood of migration success?

    1. Smaller, focused scopes reduce error complexity and make debugging easier
    2. Scope size doesn't affect success rate
    3. The scope should cover the entire migration to ensure consistency
    4. Larger scopes are better because they get more done in one prompt

15. When migrating a codebase, what should happen before labeling a checkpoint as 'complete'?

    1. All team members review and approve the changes
    2. All code compiles and relevant tests pass
    3. The developer writes new tests for the changed code
    4. The AI confirms it has finished all future checkpoints too