[DelaneyM]

I cringe at the idea that beyond a certain level one stops writing code.

You never stop writing code, unless you truly just don’t enjoy writing code, and if that’s the case please don’t manage engineers.

You do eventually stop shipping critical code, because at a certain point your managerial leverage means there are many other more productive ways to spend your time.

But the day you “stop writing code” is the beginning of the end of your effectiveness, because when your job is augmenting the impact of others, you need to have a deep familiarity with their daily experience.

Consider that any manager who “stopped writing code” five years ago has (probably) never had the daily experience of being a fully remote developer in a fully remote organization, with all the collaboration challenges and focus opportunities that entails.

And a manager who “stopped writing code” two years ago probably doesn’t get how the process changes with the assistance of LLMs, what they can dramatically accelerate and where their limitations introduce risk.

Never stop writing code. Never stop building a local version of your code base. Never stop using your internal build tools and dev environment. Never stop watching production logs and chasing down the occasional bug.

_Do_ stop assuming your code is production-ready. Do stop commenting on code reviews. Do stop assuming that the code you write has meaningful value (except as a way for you to keep your situational awareness.)

Most importantly, stop letting “writing code” be a source of stress or an obligation, and start having fun again doing it. Your organization will thank you for it.

[nine_k]

Additionally, code is an excellent tool; leverage it even if you don't ship any code.

One of my friends, a senior director in a reasonably large corporation, still manages his various financial data pipelines for analytical purposes using code, writing stuff in Python / Pandas / SQL. I bet it gives him a much better visibility into what he wants to discern in the data.