In this episode of Dead Code, Jared and guest Mond explore how James C. Scott’s concepts of legibility, metis, and episteme apply to tech interviews, arguing that standardized hiring processes prioritize what’s easy to measure over what truly reflects engineering skill. They critique the over-reliance on algorithmic interviews, noting that senior engineering work often depends on tacit knowledge—metis—that can’t be captured in rubrics or LeetCode problems. The conversation touches on Goodhart’s Law, the risks of over-optimization, and how attempts to make human processes more legible through metrics can backfire. Jared shares how his company experiments with more realistic code assessments, though both acknowledge the challenges of scaling less standardized approaches. They conclude by warning that technological standardization of social systems, like hiring, can entrench flawed norms and obscure what actually matters.
Links:
Mond’s Blog – Here Comes the Moon
Dead Code Podcast Links:
Jared’s Links: