You did the thing you were told to do. Three or four years, the best internships you could get, a transcript nobody is ashamed of. Now you are months into applications and the only reliable reply is an autoresponder.
Search why a degree is not enough anymore and the internet hands you a single story: degrees are dead, skills-based hiring won, go learn a skill instead. That story is half right, and the wrong half will cost you a year. Because in the same period everyone started declaring the degree obsolete, the share of job postings demanding one went back up.
Here is what the 2026 data actually says, including the part that contradicts the advice you have already read.
The short answer
The degree is still the ticket. It has simply stopped being the differentiator.
What a degree certifies is codified knowledge: the formal, documented, teachable kind. That is precisely the kind that got cheap in the last three years. What is scarce now is the knowledge you only acquire by being responsible for a real outcome, and no institution on earth issues a certificate for it.
Your problem is not that you learned the wrong things. It is that you have no evidence of the other kind.
What actually broke, and when
The class of 2026 numbers are bad in a specific way
The Federal Reserve Bank of New York tracks graduate outcomes quarterly, and in 2026 Q2 the unemployment rate for recent college graduates stayed elevated at about 5.6 percent while underemployment edged up to 42 percent.
Sit with the second number. Underemployment means working a job that never required the degree in the first place. Not unemployed. Employed, and holding a credential the role does not use.
This started before ChatGPT existed
Now the part that should change your strategy, because almost nothing on this topic mentions it. Recent graduates used to enjoy lower unemployment than the workforce as a whole. That advantage did not disappear when generative AI arrived. The gap crossed zero in February 2019, before generative AI and a year before COVID, and by early 2026 recent grads sat at 5.6 percent against 4.2 percent for all workers, the widest gap on record, with New York Fed research pinning roughly 64 percent of the rise on remote work rather than AI.
If you file this under "AI took the entry-level jobs", you will quietly wait for it to blow over. It will not. This is a structural drift that has been running for the better part of a decade, and AI accelerated it rather than caused it.
The degree did not stop mattering. It stopped counting as evidence.
This is where the popular story falls apart, and it falls apart on primary data.
Degree requirements are going back up, not away
Indeed's posting data shows 19.3 percent of US job postings required a bachelor's degree or higher in November 2025, up from 16.6 percent two years earlier, with the occupation-adjusted share trending upward since March 2024.
That adjustment matters more than the headline figure. Holding the mix of jobs constant, employers are now asking for a degree in roles that did not ask for one two years ago. In the middle of the skills-first era, degree inflation resumed.
Removing the requirement changed almost nothing
Employers say otherwise, of course. Lumina Foundation and Gallup surveyed 2,000 US hiring decision makers and found that 23 percent had removed degree requirements from roles that used to have them and another 20 percent were in the process, yet about three quarters still prefer or strongly prefer a candidate with a bachelor's even for roles that do not require one, and roughly seven in ten say recent graduates need a moderate or great deal of additional training to be effective.
Preference survives the policy. And the hiring data confirms it: across more than 1,000 large US employers, removing degree requirements was associated with only a two-percentage-point increase in the share of hires holding non-degree credentials, with little correlation between advertising for credentials and actually hiring people who hold them.
Two points. That is the entire revolution, measured at the offer letter.
So the honest reading is harsher than the popular one, and more useful. The degree is still a gate. It just no longer opens anything on its own. Both conditions at once.
The mechanism nobody names: your degree certifies the thing that got cheap
Here is the finding that explains the whole situation, and it comes from payroll records rather than a survey.
Stanford's Digital Economy Lab tracked millions of US workers through ADP data. In their August 2026 revision, employment among workers aged 22 to 25 in highly AI-exposed occupations sits about 19 percent below where it would be had it kept pace with less-exposed peers, experienced workers show no comparable gap, the adjustment runs through reduced hiring rather than layoffs, and the declines concentrate in occupations that rely on codified knowledge, meaning formal, standardized, documented knowledge, while employment rose among experienced workers in occupations relying on tacit knowledge acquired through practice, mentorship and repeated exposure to real situations.
Read those two words again. Codified. Tacit.
A degree is a receipt for codified knowledge. That is not an insult, it is the design. Standardised curriculum, documented material, an exam that verifies you absorbed it. For seventy years that receipt was scarce enough to be worth a career.
Then a machine learned to hold all of the documented knowledge at once, for a subscription fee.
Tacit knowledge is the other half, and it never had a certificate because it never needed one. Knowing which client email means trouble. Knowing when the number on the dashboard is wrong before you can prove it. Knowing how to tell someone their work has to be redone. You get it from practice, from being watched by someone better, and from consequences.
It used to come bundled with your first job. That is the rung that got pulled out, which is exactly why AI is thinning the entry level for anyone arriving with codified knowledge alone.
What employers screen for instead
They are not screening for the thing you think. A national survey of 3,147 US employers found 68 percent still say degrees are important while only 37 percent say higher education is preparing students with the skills needed to succeed, and 41 percent name on-the-job evaluations during probationary or trial periods as a top method of assessing whether someone has the skills.
Both halves of that sentence matter. The degree still counts, and it is not what the decision turns on. Meanwhile nearly 70 percent of employers now use skills-based hiring and projected hiring for the class of 2026 rose just 1.6 percent.
Put it together and the trial period has become the real interview. Which means the question a hiring manager is silently asking is not what you know. It is whether you have already survived a situation like this one. If you want the specific competencies that question resolves into, the AI skills employers actually test breaks them down one by one.
Stacking more credentials is not the fix
The instinct after a degree fails to convert is to buy another credential. Understandable, and mostly wrong.
A Brookings Institution analysis of more than 156 million US resumes found that a worker's first job-relevant credential carries a 3.8 percent wage premium against 1.8 percent for one that does not align to their role, workers without a bachelor's see 6.8 percent from their first job-relevant one, and certifications requiring proctored exams and periodic renewal show accumulating returns while badges and micro-credentials generate a one-time signal with little additional value.
Relevance beats volume by roughly double. One certification that maps exactly onto the job you want does more than five that look impressive on a profile. And none of them substitute for evidence, because a certificate is still a receipt for codified knowledge. You would be buying more of the cheap thing.
How to build the kind of experience that is actually scarce
Four moves. All of them are available to you this month, without anyone hiring you first.
Get inside a real deadline
Tacit knowledge does not form without consequences. Volunteer real work to a small business, a nonprofit, a student society with an actual budget, anyone who will be genuinely inconvenienced if you do not deliver. Announce the date out loud to a person who will notice. Motivation was never going to finish it. Mild public embarrassment will.
Ship something a stranger uses
One user outside your friend group changes the category of the work. Not a demo, not a design file, a thing somebody opened and used on purpose. Ugly is fine. Used is mandatory. This is the assembly job that build a portfolio with no experience walks through in detail.
Keep the failure stories
Write down what broke, what you missed, and what you changed. Interviewers probe here constantly because it is the one thing that cannot be faked or generated: the specific texture of having been wrong about something real. Almost nobody arrives with an answer. Arrive with three.
Make it openable in thirty seconds
A repository, a one-page site, a shared folder. If a hiring manager has to ask you to send something, the evidence has already failed. This is the whole practical core of how to get a job with no experience or degree, and it is the same habit behind future proofing the career after this one.
What this looks like from the other side of the desk
We run EX EPIC Academy out of Canggu, Bali, on a deliberately blunt premise: no grades, no exams, real ventures and real stakes, with participants from 26 nations working on 15 live projects. People are placed into operational roles inside real companies in the group, from waste-to-energy engineering at Zero X to automation at Gemino AI to wellness at LIV Urban Sanctuary, four to six months on site across 78 open positions.
What we see from the hiring side is consistent enough to state plainly. The people who become employable fastest are almost never the ones with the strongest transcripts. They are the ones who have already been the person a real outcome depended on, and who can describe exactly what that felt like when it went badly. That is the same honest case for a startup internship versus a corporate one: responsibility arrives earlier, and responsibility is the raw material.
Your degree is not wasted. It is unfinished. The missing half was never going to come from another course, because it was never the kind of thing a course can hand you.
FAQ
Is a degree still worth getting in 2026? Yes, and the data is calmer than the discourse. The New York Fed still puts the lifetime return on a degree near 12.5 percent, and young workers without a degree sit at about 7.2 percent unemployment against 5.6 percent for recent graduates. A degree still beats no degree. What it no longer does is beat the average worker, and that is the change people are feeling.
Which graduates are hit hardest by this? Early-career workers in AI-exposed occupations, with computer science the most visible case. Recent CS graduates now post some of the highest unemployment rates of any major, after the number of CS degrees more than doubled into a shrinking pool of openings. The major that was sold as the safest became the most exposed.
Should I skip university and go straight into work instead? Not on the strength of the skills-first headlines. About three quarters of employers say they would prefer a candidate with a bachelor's even where the role does not require one, and dropping the formal requirement shifted actual non-degree hiring by two percentage points. Skipping the degree does not level the field, it removes your ticket while leaving every competitor in place.
I already graduated with nothing to show. Where do I start? One project, one real user, one announced deadline. Not a course, not another certificate. The gap on your CV is not knowledge, it is a track record of having been accountable for something, and that can be built in weeks by giving real work away to an organisation that needs it and cannot pay for it.
Will employers actually check the projects I claim? Increasingly yes, which is good news if the work is real. Around 41 percent of employers name on-the-job evaluations during trial periods as a top way to assess skills, and roughly half assess AI competency through technical or skills-based assessment rather than taking a CV at face value. Claims get tested now. Do not put anything on the page you cannot open on a laptop.