The $200,000 Permission Slip

Nine months ago I wrote The Proof of Work Protocol (Part 1). The shift I was predicting is no longer coming — it's here. Time for some actionables.

In March 2005, Paul Graham wrote an essay called "Undergraduation." His advice to smart undergrads was simple and radical: stop worrying about grades. Work on hard, interesting projects. Learn by building real things. Ignore the parts of college that feel fake.

Reading it in 2026 feels almost prophetic. The core truth he pointed to has only grown stronger. But the cost of ignoring that advice has become far more extreme.

A college degree today is a $200,000 permission slip.

The Fine Print

Let's break down what a degree actually gets you in 2026.

Knowledge? Mostly available for free. MIT OpenCourseWare, YouTube, Coursera, and even a $20 Kindle book often cover the same ground faster. The classroom is rarely where the best version of the material lives.

A network? Partially true — but weakening. That argument made more sense when access to smart people was scarce. Today, a motivated person can build a more relevant network on X, LinkedIn, or GitHub in a fraction of the time and cost. The network you build in public often outperforms the one you inherit from a university.

Credibility and signaling? This is the strongest remaining argument. Certain names still open doors. But that signal is weakening faster than most parents want to admit — especially as more companies quietly drop degree requirements and start hiring on demonstrated capability.

What it reliably gives you is the official document proving you showed up, followed instructions, sat still for four years, and completed tasks you probably didn't care about.

That document was once a smart investment. In the twentieth century, big companies needed predictable, compliant people who fit neatly into hierarchies. A degree was the perfect signal: this person won't break the machine.

The economy has changed. The degree hasn't.

The economy now rewards agency above almost everything else. The ability to think clearly, define problems precisely, use every available tool, and ship solutions without waiting for permission.

Most traditional colleges still optimize for the old world. They reward memorization, rule-following, and reproducing the professor's preferred answers. Using AI to solve problems is often labeled cheating. In the real world, refusing to use AI is the fastest way to become obsolete.

K-12 schools prime children for college entrance & SATs of the world. Colleges set the standard that our elementary, middle and high schools optimize for. The entire pipeline has been pointing at the same destination for a century. Nobody stopped to ask whether the destination still made sense.

You'll often hear that the college system is broken. That's not quite right either. It's doing exactly what it was designed to do — reliably and efficiently. The problem is that the world it was designed for no longer exists.

A system optimized for the wrong outcome isn't broken. It's obsolete.

And when a system becomes obsolete, the signal it produces loses value. That's exactly what's happening to the degree — slowly, then all at once.

Proof of Work Is the New Degree

This is what I referred to earlier as the Proof of Work Protocol in an essay last year. Something is shifting in how the best AI-era jobs are being filled. The candidates getting offers aren’t always the ones with the best resumes. They’re the ones with the most visible body of work. A GitHub repository that solves a real problem. A thread that went viral explaining something complex clearly. A product someone built and shipped alone. Recruiters and founders are finding these people. Not the other way around.

The people getting laid off aren’t being replaced by AI. They’re being replaced by people who use AI visibly and demonstrably. The threat was never the machine — it was always the person who mastered it faster than you did.

AI isn’t taking your job. Someone using AI better than you is.

This is why the degree matters less with every passing month. A credential tells someone what you studied four years ago. A body of visible work tells them what you can do today.

When I evaluate potential collaborators or hires, I almost never look at degrees or GPAs. They tell me very little. I look at what exists in the world because of them. A working product. A useful open-source project. A clear explanation of a complex idea that someone found valuable. A project that moved from zero to something real and used.

These things reveal whether someone can think and execute. A transcript cannot. This is the shift from trust to verification.

One is a record of compliance. The other is evidence of capability.

Your digital presence is your degree now. A live URL. A GitHub repository with real users. Revenue from a side project. Content that built an audience. These artifacts cannot be faked through shortcuts. They require taste, judgment, and the willingness to be judged by reality every single day.

Four years and $200,000 versus sixty seconds and a URL. That gap is the whole argument.

The Skills That Actually Matter Today

Nobody sits you down in college and teaches you how to start before you're ready, think clearly under pressure, build an audience, or read a room. These things are left to chance — picked up on the side, learned too late, or never learned at all. That's not an accident. Compliance is easier to measure than agency. So that's what gets taught.

Here are the six skills that actually determine who thrives in 2026.

1. Action Bias

The ability to move without permission, without complete information, without certainty. This is the rarest skill and the hardest to teach because most education systems do the opposite — wait for the assignment, follow the rubric, don't act until instructed. The people who build things that matter have an almost uncomfortable tolerance for starting before they're ready.

2. Clear Thinking

AI amplifies clear thinking and exposes muddy thinking. You cannot outsource your ability to define a problem precisely, identify what actually matters, and discard the rest. This is the foundation everything else builds on. Without it AI gives you faster wrong answers. With it AI becomes a force multiplier.

3. AI Fluency

Not knowing how AI works. Knowing how to work with AI. Directing agents, chaining outputs, building systems that multiply your output. This is the new literacy. A person with genuine AI fluency operating alone can outproduce a team that doesn't have it. This skill has the fastest monetization curve of anything on this list right now.

4. Public Writing

Not academic writing. Not internal memos. Writing that travels — that makes a stranger stop, read, and share. This is the highest leverage skill for building an audience, attracting opportunity, and establishing authority in any field. One essay that resonates can open more doors than a decade of networking. College teaches writing for professors. The market rewards writing for humans.

5. Judgement

Pattern recognition built from real experience. The ability to read what people actually want versus what they say they want. Judgment under ambiguity when there is no clear answer and no rubric to follow. Empathy that builds genuine trust. These are the capabilities AI cannot replicate — not because the technology isn't advancing but because they emerge from lived human experience that no model has had.

6. On-Camera Communication

The ability to communicate directly and humanly to people who have never met you. Too many people can present to a room full of colleagues but freeze the moment a camera turns on. That gap is not trivial. Distribution is the most underrated competitive advantage in 2026. The people growing fastest — in income, influence, and opportunity — are the ones who figured out how to show up on camera consistently. Podcasting, video essays, short form content, live streams — these are the new stages. Nobody teaches this. You learn it by turning the camera on and doing it anyway.

College optimizes for compliance. These six skills optimize for agency. The gap between those two things is now measured in income, opportunity, and whether your work compounds or becomes obsolete.

The Alternative

You don't need permission to start.

Pick a real problem. Use the best tools available right now. Build something. Show your work. Get feedback. Improve. Repeat.

No application. No acceptance letter. No waiting for the right semester to begin.

My son is a junior in high school and he already gets this. A degree is a checklist item to him. Handle it with minimum effort, minimum debt, minimum time. Community college first. Check the box. Move on.

He didn't read a book to arrive at this. He just looked at the math.

He thinks laziness is just a label society puts on people who refuse to waste energy on things that don't matter. Find the shortest path to the required outcome. Spend the rest on what gives you energy. That's not a character flaw. That's knowing what your time is worth.

The students who will thrive in 2030 are not the ones who collected the best credentials. They are the ones who started building the earliest. Who got feedback from reality instead of grades from a professor. Who learned that shipping something imperfect today beats planning something perfect indefinitely.

Too many students graduate fluent in the grammar of an extinct economy. They can write papers that please a professor but struggle to ship a basic product, talk to real users, or make something people actually want. They were trained to wait for instructions in an age that punishes waiting.

The alternative is to start now. With what you have. On a problem that's real to you.

Stop buying the map.

Start building the territory.

The permission slip is optional. The work isn’t.