AI Tools: Reduce AI Vendor Lock-In Without Adding Useless Abstraction

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-tools-vendor-lockin-and-portability-r8a1-creators

1. What is vendor lock-in in the context of AI LLM providers?

   1. When switching to another provider requires significant time, money, or code changes
   2. When a provider forces you to sign exclusive contracts with penalties
   3. When an AI provider owns the data you process through their API
   4. When a provider prevents you from using competing services

2. According to the core premise, what do most 'portable LLM' wrappers typically cost developers?

   1. Only the time to configure initial settings
   2. A small monthly fee for premium features
   3. Nothing extra — they are truly portable by design
   4. Real complexity in code for low real portability gains

3. Which of these abstractions would actually pay off if you needed to switch AI providers?

   1. A custom wrapper library that reimplements all API client methods
   2. A proprietary prompt template language only your team understands
   3. A debugging tool that only works with one specific provider
   4. Message format standardization across providers

4. The lesson recommends performing which recurring activity to ensure portability actually works?

   1. Monthly price comparisons between all major providers
   2. Yearly rewrites of your entire prompt library
   3. Daily backups of all API responses
   4. Quarterly dry-runs switching one workload to a backup provider

5. What can AI tools NOT help you predict regarding AI providers?

   1. Provider price changes over time
   2. Which model will release next year
   3. Which provider will have the best accuracy for your use case
   4. How competitors will respond to market changes

6. What is the main argument the lesson makes about abstraction layers for AI provider switching?

   1. You should wait until you need to switch before building any abstraction
   2. Only narrow, specific abstractions actually pay off when switching
   3. Abstraction is always bad because it adds unnecessary complexity
   4. You should abstract everything to maximize future flexibility

7. Which of these would be an example of an abstraction that the lesson suggests skipping?

   1. A testing framework that measures quality across different providers
   2. A wrapper library that adds extra error handling around API calls
   3. A system for tracking and versioning your prompts over time
   4. Standardizing how you format conversation messages between providers

8. What does the lesson define as 'switch cost'?

   1. The difference in price between the cheapest and most expensive provider
   2. The monthly subscription fee for any AI provider
   3. The penalty fees charged by providers for early contract termination
   4. The total time and resources required to move from one provider to another

9. What does the lesson say about portability that has never been exercised?

   1. It becomes stronger the longer you wait to use it
   2. It requires legal documentation to maintain validity
   3. It automatically transfers to new providers
   4. It is portability you do not actually have

10. Which three specific abstractions does the lesson identify as paying off when switching vendors?

    1. Error handling, rate limiting, and caching layers
    2. Cost tracking, model selection, and response parsing
    3. Message format, prompt versioning, and eval contracts
    4. Authentication, logging, and monitoring systems

11. Before switching providers, the lesson suggests you should identify what?

    1. Your current stack and which providers you might switch to
    2. All possible providers even in unrelated industries
    3. The cheapest provider in your budget category
    4. Legal requirements for data handling in different countries

12. Why does the lesson say most wrapper abstractions 'cost real complexity'?

    1. Because they need constant updates as models change
    2. Because they require expensive cloud infrastructure to run
    3. Because they add extra code, maintenance, and learning overhead without proportional portability benefit
    4. Because providers intentionally break compatibility with third-party tools

13. What is an 'eval contract' as referenced in the lesson?

    1. An evaluation tool provided by the AI company for testing their models
    2. A legal agreement between you and a provider about service levels
    3. A standardized way to measure and compare quality across different AI providers
    4. A contract clause that locks you into pricing for a set period

14. The lesson states that AI cannot eliminate switch cost — it can only do what?

    1. Calculate exactly what it will be upfront
    2. Reduce it through strategic abstraction
    3. Increase it to encourage provider loyalty
    4. Delay it until a more convenient time

15. The lesson argues that you should skip abstractions that do what?

    1. Require less than one week to implement
    2. Add meaningful value to your application logic
    3. Are open source and freely available
    4. Just add layers between you and the model without helping switching