Degrees say where you learned. AI-era hiring asks what you can ship.

The AI skills vs degree question has a blunt answer: a degree was never the capability. It was a proxy for it: proof you could learn hard things, finish them, and clear a filter. For generations that proxy was the best signal hiring had, because faking it was expensive and nothing cheaper measured the same thing.

AI broke it from two directions at once. First, the paper trail stopped being trustworthy. When every resume and cover letter can be polished by a model in seconds, a sharp application no longer proves the applicant is sharp. Recruiters know it, so the screens that follow keep getting more hands-on. Second, the skills employers most want certified are too new to certify. Nobody holds a 2019 diploma in directing AI agents. The transcript can't speak to the exact capability the role now turns on, so employers are pushed toward the only signal left: watch the person work. Live problem screens, portfolio reviews, take-home samples, and the increasingly common interview request, show me your actual AI workflow.

The prize for getting evaluation right is real. LinkedIn's Economic Graph found that a skills-first approach to hiring can enlarge talent pools roughly 6.1x globally. That's not a rounding error, it's most of the qualified people the degree filter was screening out.

The announcements skip one inconvenient finding. The Burning Glass Institute studied firms that publicly removed degree requirements and found actual hiring practice shifted for fewer than one in 700 hires. Companies changed the job posting; the people doing the hiring kept reaching for the old proxy. Policy moved, behavior mostly didn't.

So the practical read for professionals is not 'wait for the door to open.' It's to build proof that bypasses the proxy war entirely. A demonstrable AI workflow, real tasks you can run live with your own prompts, checks, and judgment, beats both a degree and a polished resume, because it's the one signal AI can't fake on your behalf. We covered why the credential pipeline can't keep up in higher ed's four-year pipeline problem, and why this is reshuffling who gets hired at the entry level. The pattern is the same everywhere: when paper is cheap, proof gets expensive, and the people holding proof get paid.

For employers the implication flips. If almost everyone else announced skills-based hiring and then kept keyword-matching diplomas, actually testing for capability is an arbitrage. Run work-sample screens for the AI-heavy parts of the role and you're fishing in the enlarged pool while competitors fight over the same credentialed shortlist.