Kent Beck to junior engineers: nobody is counting your closed tasks

Kent Beck’s advice to new engineers inverts the obvious assumption: the count of tasks you finish barely registers with the senior engineers around you. They could clear your backlog faster themselves, so raw output was never the point of hiring you. Your salary is an option premium on the engineer you’ll become — the company is betting that investing in your growth now yields a stronger next generation rather than seniors still grinding the same work a decade later.

Beck frames the real evaluation as sorting juniors into three buckets — game-changers (A), solid performers (B), and people who won’t last the year (C). Clearing the B-from-C bar is mostly about reliability: your code works, you communicate what you’re doing, you finish in a sane timeframe, and you don’t dump avoidable work on reviewers, on-call, or devops. Faking progress is an instant C signal. Crossing from B to A is about the first derivative — how much you learn per task rather than how many you close. Strong signals include arguing a task shouldn’t be done at all, isolating the 10 percent of work that delivers 90 percent of the value, shipping small daily diffs that also simplify surrounding code, building tools that ease similar work, writing up what you learned, and reviewing others well.

The throughline is that every A-grade behavior costs more time than the minimum needed to close a ticket — but not unbounded time. The discipline is finishing reliably while spending the margin on work that compounds for the team. It’s a pointed reframing of early-career performance: optimize for trajectory and leverage, not throughput.