More Friction Won’t Fix Recruiting’s Real Problem

Andrew Stockwell used to run hiring like a science. As head of people at Vendr, he posted a role, waited a few days, and reviewed a few dozen applications, up to 100 if he was lucky, before passing sterling candidates to hiring managers. Within the last year, that routine broke. He now gets flooded within a day or two, sometimes topping a thousand applications for a single role, many fake, many more written by AI in ways that make candidates indistinguishable. His fix was not better judgment. It was paying his highest-paid talent-acquisition staff to read through the pile all day, according to Stockwell’s account to WIRED.

That detail captures the whole shift. The industry’s answer to an overwhelmed pipeline is to make it smaller again. That solves volume. It does not touch the judgment problem sitting underneath, and that problem, not application volume, is recruiting’s real failure point.

The Friction Recruiters Say They Want

Ophir Samson, who leads voice AI at Greenhouse, told WIRED that a year ago every recruiter wanted a “seamless experience” for applicants. What they got, he says, was 2,000 applicants in 24 hours for a job. Now recruiters tell him the opposite: “Actually, we kind of want friction. The friction is good. We want to make it harder.”

That reversal is an admission, not a strategy. A frictionless front door was never the real constraint on quality hiring. Recruiters made applying effortless, then discovered that ease simply pushed the question of who is actually good further down a process built without a reliable way to answer it. Friction now just delays the same reckoning.

What LinkedIn’s Own Numbers Actually Show

LinkedIn told WIRED that submissions per applicant on its platform are up 46 percent compared with February 2020. Separately, applications overall are up 22 percent since the launch of ChatGPT. Those are LinkedIn’s own figures, describing behavior on its own platform.

They sit alongside separate Bureau of Labor Statistics data. Job openings peaked at 12.3 million in March 2022, then fell for two years before settling near 7 million since mid-2024. Applying got easier in the same window that roles got scarcer. WIRED does not claim one caused the other, and neither should you. Either way, more candidates are chasing fewer roles through a process built for the opposite ratio.

LinkedIn’s own response is instructive. The platform is capping automated and low-quality applications, and this month began telling likely-underqualified applicants they probably are not a fit. That is filtering at the gate, executed by the platform instead of the recruiter. It screens out obvious mismatches without telling anyone how to judge two plausible candidates against each other.

The Diagnosis Buried in Greenhouse’s Own Pitch

Samson’s own framing matters here: a recruiting-tech vendor is naming the exact problem it sells a fix for. “We frequently hear from recruiters: they have 1,000 applications, but they know that only 30 of them are serious,” he says. Recruiters already know roughly which fraction of a bloated pool matters. The problem was never finding that fraction. It is what happens to those 30 once found.

Samson insists these AI interviews serve a purpose beyond adding friction, since they surface context a resume cannot. That is a fair distinction, and it still stops at sorting the serious from the unserious. A sharper filter is not a better judgment of who actually performs once hired.

Tessa White, a former HR executive who left corporate roles in 2018, put it more bluntly to WIRED. Even AI-driven applicant tracking systems, she says, are “an antiquated way to look at people and skill sets.” White’s target is the technology doing the sorting. Ours is what happens after the technology steps aside, the point where an actual person judges an actual candidate, still without a consistent standard behind it.

Referrals Are a Retreat, Not a Repair

WIRED reports that the deluge of applications has pushed recruiters back toward referrals and internal hires as the only reliable route through the noise. Stockwell, job hunting himself after leaving Vendr in February, says a personal connection is close to essential just to land an interview. He also flags the cost: referral-heavy hiring favors candidates whose backgrounds mirror people already inside the company, which he says works against building a diverse workforce.

That retreat is the clearest evidence that the industry has not solved evaluation. It has abandoned it. If an organization cannot judge a stranger’s application reliably, the answer is not to stop looking at strangers. In our work with hiring teams, the interview stage a candidate reaches after clearing every filter is usually the least standardized part of the process, run on gut feel dressed up as a scorecard. Cutting the pool from a thousand to a hundred changes nothing about that stage. It only pushes back the point at which weak evaluation shows up.

Where the Fix Actually Has to Land

Add all the friction you like to the application. If the room behind it still cannot tell a good candidate from a good story, all that friction does is make people wait longer to find that out. Ask yourself what happens after the shortlist is set, because that is where the real test sits.

Every knock-out question, skills test, and AI interview in WIRED’s reporting does the same job: it decides who gets through. None of it decides how well those people get judged once they arrive. That is the harder problem, the one White gestures at when she says the system needs a “massive overhaul” while admitting she does not have the answer herself. Neither does more friction. The moment a resume becomes a person in front of an interviewer is still the part none of this touches.

Original reporting: WIRED.