How good engineers write bad code at big companies

Big tech firms move engineers around so often that most code changes are made by people who have been on a team or language for less than six months. The average tenure on a single codebase is far shorter than the lifespan of the service, so expertise never fully matures. This structural churn means even competent engineers are regularly writing in unfamiliar territory, which breeds the kind of hacky solutions the article calls "bad code".

The few "old hands" who retain deep knowledge do so informally and are constantly overloaded. They may glance at a change in a spare half-hour, veto the worst parts, and push a quick fix forward. Their limited time and the incentive to stay productive mean many defects slip through, especially when senior engineers are juggling multiple deadlines. The article cites a concrete scenario where a junior engineer patches a bug, receives a brief review, and the change ships without further scrutiny, only to be noticed years later as a glaringly sloppy piece of code.

The underlying trade-off is intentional: companies prioritize internal legibility-being able to reassign engineers at will-over sustained code quality. Leaders should recognize that blaming individual engineers misses the systemic cause. By fostering longer-term ownership of services, reducing forced rotations, and giving "old hands" formal authority and time to mentor, technical leaders can mitigate the churn that drives technical debt and improve overall team performance.