How AI is quietly making your team fragile — Writing<br>“If we all reacted the same way, we’d be predictable, and there’s always more than one way to view a situation. What’s true for the group is also true for the individual. It’s simple: overspecialize, and you breed in weakness. It’s slow death.”<br>Major Motoko Kusanagi, Ghost in the Shell (1995)p]:mb-6 [&>p]:max-w-[65ch] [&_a]:text-accent [&_a:hover]:text-accent-hover [&_a]:underline [&_a]:underline-offset-2">The models are good enough now that the surface looks the same whether or not anything real is underneath. You can’t read a candidate the way you used to. This is about what that’s doing to your team, one hire at a time.
Everyone optimized. Nobody gained.
p]:mb-6 [&>p]:max-w-[65ch] [&_a]:text-accent [&_a:hover]:text-accent-hover [&_a]:underline [&_a]:underline-offset-2">Whichever side of the table you sit on, you’re already caught in this. Three things are happening at once, and each one is reasonable on its own.<br>Candidates automate the search. AI writes the CV, tailors the cover letter to the posting, and applies at a volume no human would attempt by hand. Why wouldn’t they? Everyone else is.<br>Companies add AI pre-screening, because the volume coming in is now unmanageable by hand. Why wouldn’t they? They have to.<br>And interviewers, quietly, start judging candidates on how fluently they use AI, regardless of whether any policy tells them to. Why wouldn’t they? It's important, right?<br>None of these choices is wrong in isolation. Put them in a loop: everyone adopted AI for an edge, so nobody has one. The signal you were all competing on just got noisier.<br>You can already watch this happen, and people are starting to name it. Companies repost the same role month after month, plainly unable to fill it, while experienced candidates fire off application after application and can’t get so much as a first reply. Gergely Orosz calls it a Catch-22: the hiring managers who can’t find senior people and the senior people who can’t get answered are playing Marco Polo in a nightclub: both calling out, neither able to hear a damn thing over the noise.
01The candidate automates the search: AI-tailored CV and cover letter, auto-apply at volume.<br>02The company adds AI pre-screening, because the volume is now unmanageable by hand.<br>03The interviewer quietly judges candidates on AI fluency, before any policy says to.
↩ …and the interviewer's bar feeds back into how the next candidate prepares. Everyone adopted AI for an edge, so nobody has one. The signal just got noisier. It's a feedback loop.
What you can measure<br>Performance isn’t a number you can see
p]:mb-6 [&>p]:max-w-[65ch] [&_a]:text-accent [&_a:hover]:text-accent-hover [&_a]:underline [&_a]:underline-offset-2">Line everyone up on a single number, best to worst, and call it performance. That’s the assumption behind every stack-rank, and it doesn’t hold. There are two different axes here: what you can predict about someone at hiring time, and how they actually do on the job. They’re not the same axis, and they correlate less than you’d hope.<br>The best-validated selection methods we have correlate with job performance at around r ≈ 0.51 , an R² of about 26%. Roughly three-quarters of what determines how someone performs is invisible at the moment you decide to hire them (Schmidt & Hunter 1998; Sackett et al. 2021 revise even that downward).<br>To their credit, hiring tools have mostly absorbed what a century of personnel psychology learned. Structured interviews and work samples predict performance best; unstructured chats, years of experience, and school prestige predict it worst, and the scorecards most modern applicant-tracking systems ship are built on exactly that finding (Hunter & Hunter 1984; Sackett et al.). The problem was always that even the best-validated method still leaves most of what drives performance unmeasured, and no scorecard, however good, can score what it can’t see.<br>The mistake is to treat that missing three-quarters as a measurement error, or noise to be cleaned up with a better filter. It’s the territory, not error. The spread is real, and most of it is unknowable in advance.
good hire, missedgood hire, caughtcorrect rejectbad hire, passedpredicted at hiring →actual on the job →Predicted at hiring against actual on the job. The correlation is real but loose (r ≈ 0.51). Two of the four quadrants are errors you never see: the good hire you rejected, the bad hire you passed.Why we hire<br>You’re not hiring an individual. You’re changing a system.
p]:mb-6 [&>p]:max-w-[65ch] [&_a]:text-accent [&_a:hover]:text-accent-hover [&_a]:underline [&_a]:underline-offset-2">When you make a hire, you’re not bolting on an isolated unit of output. You’re adding a node to a network, with edges: who it talks to, who it unblocks, what it changes about how the people already there work.<br>That means local optimization can degrade global throughput. You can add the individually strongest candidate and make the whole graph...