On September 22, Andreessen Horowitz announced The Horowitz Andreessen Academy, a new school in San Francisco for people straight out of high school. What was most interesting to me was that students are chosen on proof of work: what they built, shipped, or earned. Transcripts and test scores are still collected, but the FAQ says admission is "heavily weighted toward proof of work."

To anyone who has ever hired an engineer, choosing on proof of work sounds obvious. Of course we look at the work, and most of us do it the moment a candidate sends through a GitHub link. A project someone built tells us more than lines on a resume ever can, because it shows us:
how they thought through the problem
what they chose to build, and what they left out
what they did when something broke
That matters even more now. AI makes a polished answer easy to produce - a cover letter, an essay, a neat written response to almost any question - but a working project, with decisions the builder can explain, is much harder to fake. The Academy's own FAQ says it plainly: "What we want to see is how you think and what you've actually built."
So is the Academy right to ask for the work? Yes, and I hope more schools and companies do this.
We built NextWork around the idea that proof of work beats credentials. Now one of the most influential firms in tech is choosing students on the same basis. That's good news, and it raises a harder question.
Choosing people on proof of work only helps the people who already have some proof to show. So who has proof of work to show? What did those applicants do before they applied? And where did they get the chance to make it?
The Academy is honest about the answer. It describes itself as being "for a specific kind of student: someone who already knows how to build," and its launch release goes further, admitting that "college is still the right path for most students." That leaves the bigger job to the rest of education, which is helping everyone who doesn't have proof yet go and make some.
Dropping the degree requirement was never enough
The idea that a credential shouldn't decide who gets in is not new. In 1971 the US Supreme Court decided Griggs v. Duke Power. The company required a high school diploma to move out of its labor department, the lowest-paid part of the business. Chief Justice Burger's opinion put it this way:
"Diplomas and tests are useful servants, but Congress has mandated the commonsense proposition that they are not to become masters of reality."
Put plainly, a credential shouldn't decide who gets in when it doesn't measure whether someone can do the work.
Fifty years later, plenty of employers took that idea seriously and dropped degree requirements from their job ads. Then researchers checked whether it changed who actually got hired. In February 2024, the Burning Glass Institute and Harvard Business School found that it barely did: across the US, dropping the requirement changed the outcome in "not even 1 in 700 hires last year."
To me, that says removing a requirement on its own doesn't do much. Whether the ask is a degree, a certification or proof of work, people still need real experience, and evidence of that experience, before anyone will let them in.
And the place most of us got that experience, the entry-level job, is getting harder to reach. Stanford's Digital Economy Lab tracks 22 to 25 year olds in the jobs most exposed to AI, and in August 2026 it found their employment had dropped by about a fifth compared with people the same age in less exposed jobs. Experienced workers in the same jobs saw no such drop.
Fewer entry-level jobs means fewer people shipping their first real thing, breaking it and fixing it, with a manager who can vouch for the work. So fewer people have anything to show when a school like the Academy asks for it.
What NextWork does
NextWork doesn't select people who already have proof of work; we give people a place to make it.
Every NextWork project is real work, done with the tools professionals use, so when you finish one you have something you built and can explain, and a record of what you're capable of. The point is to have that record before an employer, a school or any other gatekeeper asks you to show it. More than 220,000 people in 190 countries are building theirs on NextWork right now.
If proof of work is the new credential, then the job of education changes with it. It has less to do with sorting people by what they have already built, and more to do with giving someone with nothing to show yet a place to make something.
Proof-based selection is the right idea, but it is only fair if far more people get a real chance to make proof before someone judges them on it. Right now that chance mostly goes to the people who already know how to build - and widening it is the work that matters.
All my best,







