Hiring in tech is mostly bullshit

Tejas Kumar argues that many fashionable corporate practices - bar-raiser hiring loops, OKRs, the so-called Spotify org chart, and modern Agile rituals - are largely theatrical copy‑jobs that produce conformity, not better outcomes. He describes how bar-raiser panels (originating at Amazon) turn hiring into scored STAR storytelling even though Google’s own analysis found interview scores unrelated to on‑the‑job performance, and more recent meta‑analysis shows structured interviews (asking the same job‑focused questions) predict performance far better (correlation ~0.42) than elaborate loops. A personal hiring anecdote underscores the point: a small startup hired him for demonstrated technical skill and initiative, not ritualized values interviews.

He links these practices to leader filtration theory - formal filters reliably produce middling, predictable leaders and exclude risky outsiders who can be transformative - and shows how companies misapply Spotify’s original 2012 whitepaper, OKRs, and Agile into processes that strip accountability, encourage gaming, and narrow behavior. Research on goal setting documents predictable harms from poorly dosed OKRs, and Agile has been hollowed into ceremonies. AI makes the problem urgent because these systems select for people who follow scripts - precisely what large language models can substitute - evidenced by early employment declines for young workers in AI‑exposed roles and examples of interview automation. The practical alternative Kumar proposes is simpler: job‑knowledge tests, work samples, structured interviews by people who actually do the work, and fewer, focused evaluators.