AI Didn't Break Hiring. It Exposed It.
The AI résumé loop isn't a bug. It's a stress test. And the test is revealing that hiring was always performance theater--we just couldn't see it until the costumes got cheap.
The tweet landed like a punchline to a joke no one finds funny:
"YOUNG PEOPLE ARE USING CHATGPT TO WRITE THEIR APPLICATIONS; HR IS USING AI TO READ THEM; NO ONE IS GETTING HIRED"
-- @FirstSquawk, February 2026
Greenhouse CEO Daniel Chait calls it the "AI Doom Loop": job seekers use AI to auto-apply to hundreds of positions, employers drown in undifferentiated applications and deploy AI filters to cope, the filtering window tightens, candidates spray even more applications--and the cycle accelerates.
The framing is dystopian. The reality is clarifying.
The Economics of Cheap Signals
In 1973, economist Michael Spence won a Nobel Prize for explaining why hiring has always been a signaling game. Employers can't directly observe talent, so they rely on costly signals: degrees, pedigree, polished formatting, corporate tone. These signals "work" only when they're expensive enough that not everyone can produce them. Scarcity equals credibility.
AI just erased that cost structure.
Here's the chain reaction:
- AI commoditizes résumé polish. What once required professional editing, insider knowledge, and institutional fluency is now free.
- When a signal loses scarcity, it stops filtering. If everyone's application looks equally polished, polish differentiates nothing.
- Filtering failure raises evaluation costs. Employers can no longer skim--they must dig deeper or automate more aggressively.
- Companies climb toward higher-fidelity signals. Portfolios, shipped work, skills assessments, real outcomes.
Historical Precedent: Signals Die When They Get Cheap
Degree inflation: When bachelor's degrees became near-universal for white-collar work, the signal degraded. Employers started requiring master's degrees for roles that once needed none.
SAT prep industrialization: Once test prep became a $2 billion industry, SAT scores stopped measuring aptitude and started measuring prep spend. Colleges noticed. Many went test-optional.
SEO arms races: When everyone learned to keyword-stuff, Google pivoted to behavior-based ranking signals. The old game died; a new one emerged.
Whenever a signal becomes cheap to produce, it stops working. Markets adapt by moving up the fidelity curve--toward signals that are harder to fake.
That's exactly where hiring is heading now.
The Data: Policy vs. Practice
The shift is already underway--though messier than headlines suggest.
According to TestGorilla's 2025 survey, 85% of employers now use some form of skills-based hiring. 53% have removed degree requirements from at least some roles. LinkedIn's Skills-First Report finds that skills-based matching expands candidate pools 6.1× overall and 8.2× for AI roles.
But here's the gap: a joint study by the Burning Glass Institute and Harvard Business School found that at some large firms, fewer than 1 in 700 hires were non-college graduates--even after degree requirements were dropped.
Policy is ahead of practice. Old habits die hard. But the pressure is real, and it's building.
Why This Time Is Different
The AI résumé loop is different from previous signal collapses because it's happening on both sides simultaneously.
Candidates didn't gradually learn to game the system over decades. ChatGPT handed everyone institutional fluency overnight. And employers didn't slowly lose their filtering edge--AI screening tools scaled the same pattern-matching logic that was already failing, just faster.
The result: the system hit saturation in months, not years.
Chait's diagnosis is precise: "It's not that AI is bad--it's that the current AI tools are making the hiring problem worse for everyone." The firehose of optimized applications meets the firehose of automated filtering. Both sides optimize the surface layer until the surface layer collapses.
The Counter-Case
One objection deserves acknowledgment: companies might not move toward meritocracy. They could retreat to prestige--doubling down on elite degrees, insider referrals, and high-status filters precisely because they feel overwhelmed.
Some will. But that strategy has limits.
Prestige signals are scarce because they gate access, not because they identify talent. Over-reliance shrinks candidate pools, worsens diversity, raises costs, and eventually backfires--especially when competitors scoop up the overlooked high-skill, low-pedigree performers and outperform.
The LinkedIn data is instructive: skills-based matching increases the share of women in AI talent pools by up to 24% globally. Companies that cling to old filters aren't just missing talent--they're losing the demographic arbitrage.
Fidelity beats familiarity in the long run.
What This Means
If AI can generate a flawless résumé, then résumé quality was never a talent indicator--it was a performance of institutional fluency. High-status candidates used to win because they could afford polish. AI hands that polish to everyone.
Advantage deleted.
This doesn't kill competition. It shifts it.
The new edge isn't formatting skill. It's proof of work: GitHub repos, writing samples, shipped products, sales numbers, audience traction, reputation within networks. Signals that are hard to fake because they require doing something.
Greenhouse is already adapting. Their new "Real Talent" product combines intent matching, bad-actor detection, and identity verification to distinguish genuine candidates from spray-and-pray automation. IBM, Google, Accenture, and Walmart have shifted millions of roles toward skills-based criteria. The market is moving.
The Punchline
The AI résumé loop feels like chaos because we're in the saturation phase. Every signal collapse starts with saturation. Then comes recalibration.
What's dying is the prestige theater of hiring: the formatting arbitrage, the keyword games, the credentialist gatekeeping that confused polish for potential.
What's left is the signal that's hardest to counterfeit and easiest to verify: visible, shipped, undeniable competence.
And if we're lucky, this is the chapter where the labor market finally rewards the work--not the gloss.
Sources:
