How AI Coding Assistants Actually Work

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-ai-foundations-coding-assistants-final1-creators

1. What is the primary factor that determines the quality of an AI code completion?

   1. The quality of context gathering from the editor
   2. The theme settings in your IDE
   3. The size of the language model
   4. The speed of your internet connection

2. Which capability is NOT listed in the lesson as something AI coding assistants do well?

   1. Answering questions about code you have given the model access to
   2. Suggesting completions that match your codebase's idioms
   3. Drafting tests, docs, and small functions from clear specifications
   4. Refactoring across many files without supervision

3. What does the term 'RAG over code' refer to?

   1. Retrieval-Augmented Generation applied to codebases
   2. Recursive Auto-Generated code blocks
   3. Runtime Application Generation for code
   4. Rapid API Gateway for code snippets

4. Who bears responsibility for the correctness, security, and maintenance of code you accept from an AI coding assistant?

   1. The open-source community
   2. The AI model provider
   3. The IDE developer
   4. The engineer who accepts the suggestion

5. What is the term for the process of collecting surrounding code, file names, and recent edits to send to the AI model?

   1. Syntax highlighting
   2. Cache warming
   3. Code compilation
   4. Context gathering

6. What is the main reason AI coding assistants cannot understand your entire codebase at once?

   1. Context windows still bind the model's ability
   2. The code is stored in different repositories
   3. The AI lacks access to private files
   4. The IDE prevents access to large files

7. Why does the lesson advise against auto-accepting AI code suggestions?

   1. It creates technical debt that compounds quickly
   2. The AI will stop suggesting completions
   3. It violates open-source licenses
   4. Auto-acceptance uses too much memory

8. Which of the following is explicitly listed as a key term in this lesson?

   1. Machine learning inference
   2. Transformer models
   3. Neural network architecture
   4. RAG over code

9. What should you watch in your editor to understand how much context an AI assistant is using?

   1. The network requests when you accept a completion
   2. The memory usage graph
   3. The file save history
   4. The CPU usage meter

10. What happens to code responsibility when you accept an AI-generated suggestion?

    1. The responsibility remains with the engineer
    2. There is no responsibility for AI-generated code
    3. The responsibility transfers to the AI company
    4. The responsibility is shared equally

11. Which component is NOT part of the three-part system described in the lesson?

    1. Prompt scaffolding
    2. A graphics processing unit
    3. Careful context gathering from the editor
    4. A code-trained model

12. What kind of output can AI coding assistants reliably produce from clear specifications?

    1. Large-scale enterprise applications
    2. Tests, docs, and small functions
    3. Compiled binary executables
    4. Entire operating systems

13. What does the lesson say about AI coding assistants compared to magic?

    1. They are not magic — they combine specific technical components
    2. They are magical once configured properly
    3. They have achieved true artificial general intelligence
    4. They only appear magical in movies

14. What fundamental limitation prevents AI assistants from understanding large codebases?

    1. The context window size limit
    2. Lack of internet connectivity
    3. Inability to read certain file formats
    4. Security restrictions in the IDE

15. Why should you read every line of code before accepting an AI suggestion?

    1. The model will learn from your review
    2. The IDE requires it for copyright reasons
    3. You are responsible for the code's correctness and security
    4. The AI never makes mistakes