Ai Coding And Product Systems medium
Your AI review system receives a 4,000-line pull request spanning 60 files. Sending the whole diff in one prompt yields shallow, generic comments, and real bugs in files near the middle of the diff are missed. How should you structure the review?
- A Send the entire diff in a single prompt with a larger context window so the model sees every cross-file relationship at once.
- B Split the diff into per-file or per-function hunks, add surrounding code and related definitions, and review chunks in parallel.
- C Review only the first few files of the diff in full detail and skip the rest, since early files usually contain the core logic.
- D Strip the diff down to added lines only, removing context lines and deleted code so that more changes fit in the same token budget.
Answer: B. Split the diff into per-file or per-function hunks, add surrounding code and related definitions, and review chunks in parallel.
**Diff chunking** trades some cross-file visibility for focused attention and parallelism. You recover the lost context by retrieving related definitions for each chunk. A larger window (A) is tempting, but long prompts dilute attention, which is exactly why mid-diff bugs get missed. A single prompt is the right call for small PRs that fit comfortably within the budget.
One system design question a day, with spaced repetition so the ones you miss come back.
Get the appRelated questions
- Your assistant uses a classifier to route queries to a small or large model, which saves 60% of GPU spend. Analysis shows about 20% of hard queries are misrouted to the small model, which then gives confident wrong answers, and regenerate clicks on those threads have doubled. What should you do?
- Your workspace AI retrieves the top-k chunks from a shared vector index and passes them to the model with ACL metadata and a system prompt saying 'only use documents this user can access.' A user just received an answer quoting a private HR page they cannot open. What is the correct fix?
- Your support bot answers billing questions. One user has asked the same refund question three times, and retrieval confidence for that topic is low. What should the bot do next?
- In your ChatGPT-style assistant, power users keep single threads running for weeks, reaching about 150K tokens. Cost per turn keeps climbing, latency is rising, and the model increasingly misses facts from early in the thread. What should you do?
- Your workspace AI filters retrieval using ACLs stored in the vector index, which is rebuilt nightly. An admin removes an employee from a confidential project page, but for the rest of the day the employee can still get AI answers drawn from it. What is the right fix?
- Free users consume 70% of your assistant's GPU hours but bring in about 5% of revenue through conversions. Leadership wants to keep the free tier useful enough to drive upgrades. Which policy fits?