At Regeneron, I learned how easy it is to get pulled into new ideas and lose sight of the main goal. I've become much more deliberate about asking whether what I'm building actually moves the project forward.

That matters even more now that faster coding tools make it easy to build more than is actually useful.