korrents
Lee Robinson

What Lee Robinson thinks about software quality

@lee-robinson · 58 positions · On ppll ↗

Works on ML at SpaceX, as of 2026; formerly at Cursor, and five years at Vercel before that, ending as VP and working on Next.js. Writes at leerob.com.

Lee Robinson did not write this page.

We collected these quotes from things they published elsewhere, and every quote links to where it was said. They have no account here and have not endorsed this site. Quotes are word for word; the short line under each one is our own restatement, not their wording. Their own site. Is this you? Claim it or ask us to remove it. Or tell us what is wrong here.

10 dated positions, 2025 to 2026, in their own words. Our reading of what Lee Robinson has said — not written or endorsed by them.

  1. And sometimes that might work, but it's really important to follow the existing patterns in your code base, otherwise you can create a lot of tech debt.

    ↗Cursor: coding agents tutorial (2026)youtube.com 6th of 18 in this recording

  2. But by far, my biggest tip is to make sure you really understand the proposed solution. If you don't understand the fix, it's going to be very hard for you to validate whether the code is actually correct.

    ↗Cursor: coding agents tutorial (2026)youtube.com 10th of 18 in this recording

  3. coding agents

    Coding agents can write a lot of code, which means they can also produce tech debt or introduce an accidental bug. Your standards for what gets merged should be the same whether it was written by an agent or written by hand.

    ↗Cursor: coding agents tutorial (2026)youtube.com 11th of 18 in this recording

  4. code generation

    This AI-generated code can look correct, but just be slightly wrong. It might follow all your existing patterns, compile, pass test, yet still have an issue somewhere in your code base. This is why code review is so, so important.

    ↗Cursor: coding agents tutorial (2026)youtube.com 12th of 18 in this recording

  5. I find that when you break it down into more semantic commits, it's much easier for a human reviewer to step through the commits and understand things if they have more questions.

    ↗Cursor: coding agents tutorial (2026)youtube.com 13th of 18 in this recording

  6. But, I found that agents make writing tests much more approachable when you can ask the agent not only to write the test for a new feature, but also verify that they're correct.

    ↗Cursor: coding agents tutorial (2026)youtube.com 14th of 18 in this recording

  7. coding agents

    agents, you can only move as fast as the slowest part of your system. And that's often waiting for a test suite to finish or running type or lint checks across your entire code base or other steps in your CI pipeline.

    ↗Cursor: coding agents tutorial (2026)youtube.com 15th of 18 in this recording

  8. 8 months earlier
  9. The formatting is like low uh low risk versus running potential autofixes on the linting rules is a little bit more potentially problematic

    ↗Refactoring a Next.js & Tailwind app with Cursoryoutube.com 4th of 9 in this recording

  10. Prefer having more descriptive variable or function names. Like I really like this comment doesn't do anything. It's just slop.

    ↗Refactoring a Next.js & Tailwind app with Cursoryoutube.com 7th of 9 in this recording

  11. It's very much a pair programmer. It's not like I'm just delegating all control to the AI. I'm still making very surgical changes and using the model to help me in this journey of updating a bunch of stuff and cleaning up tech debt and moving things around.

    ↗Refactoring a Next.js & Tailwind app with Cursoryoutube.com 8th of 9 in this recording