Should universities replace traditional exams with project-based assessment?

Yes. Universities should replace traditional exams with project-based assessment.

Why? Because exams mostly measure short-term recall under artificial pressure. That is not the same thing as competence. In the real world, nobody hands you a blank sheet, a countdown timer, and says, “Now prove you can think.”

Projects test what actually matters: problem-solving, synthesis, collaboration, judgment, persistence, and the ability to produce something useful. That’s education with output, not just education with anxiety.

Traditional exams also reward a narrow skill set. If you’re good at memorizing fast and performing under stress, you can do well without truly understanding the material. Strange design choice, really, if the goal is to certify learning.

Project-based assessment is not perfect. It takes more effort to design and evaluate well. But that’s not an argument against it. That’s an argument against lazy administration.

Universities should be preparing people for complex work, not training them to survive timed memory tests. Replace the exam fetish with assessment that reflects reality.

Imagine you’re building a bridge. You’d want the engineer to have spent weeks on a beautiful project, sure. But in the middle of a storm, you also want to know that the math is etched into their bones, not just tucked away in a shared Google Doc they haven't looked at since Tuesday.

Moosk talks about the real world like it’s one big collaborative workshop, but truth has a way of finding you when you’re alone. Exams aren't just "timed memory tests"; they are the forge. They’re the only place where we can’t hide behind a charismatic teammate or a polished presentation deck. When we replace exams entirely, we risk replacing deep, internalized knowledge with the ability to just manage a process.

We’ve all been there—the group project where one person carries the world on their shoulders while someone else gets the same grade for just showing up with the snacks. That isn't "real world" preparation; that’s a lesson in resentment. Projects can be a fog where individual incompetence hides. Exams, however, offer a rare, quiet kind of justice. They say: "This is what you, and only you, have truly mastered."

If we scrap them, we lose the baseline. We lose the guarantee that the person holding the degree actually possesses the skills it claims they do. A project shows what a group can curate; an exam shows what a human being has actually become. You can’t outsource your soul to a group chat, and you shouldn't be able to outsource your education either. Don't trade the mirror for the fog.

Cute framing. But it misses the core issue.

First: projects do not mean “group chat and vibes.” Good project-based assessment can be individual, version-controlled, documented, and defended orally. If your complaint is weak implementation, fix the implementation. Don’t cling to a worse system because someone once got trapped in a bad group project.

Second: exams do not guarantee deep knowledge. They often measure cramming, pattern recognition, and stress tolerance. That’s not the same as mastery. A student can ace an exam on Friday and forget half of it by Monday. So much for “etched into their bones.”

Third: the “baseline” argument cuts both ways. A project with milestones, drafts, critiques, and a final defense gives a much richer baseline than a few hours under fluorescent lighting. It shows process, not just a snapshot. Snapshots are convenient. Reality is messier.

And let’s be honest: universities already accept a massive amount of exam theater. Closed-book, timed recall is a crude proxy for knowledge in many fields. Useful in some cases? Sure. Sacred? No.

Do we need exams in every discipline? Obviously not. You don’t assess architecture, software, medicine, design, policy, or research the same way you assess rote recall. The idea that a single universal test format should rule higher education is a legacy habit, not a principle.

The real question is simple: what do we want degrees to mean?
If the answer is “can survive a timed test,” keep the status quo.
If the answer is “can apply knowledge to solve real problems,” then projects win.

You call it “exam theater,” but I see it as a moment of truth. There’s a certain kind of quiet bravery in walking into a room with nothing but what you’ve managed to carry inside your own mind. It’s the difference between owning a map and actually knowing the way home.

Moosk talks about "version control" and "documented processes," but in the real world, there are moments where the process breaks and all you have left is what you’ve internalized. If a doctor is in an emergency room, they don't get to "defend their process" to a committee while the patient is waiting. They have to know. Exams aren't just tests of memory; they’re tests of integration. They force you to weave disparate threads of information into a single, functional tapestry that lives in your brain, not just on a hard drive.

And let’s talk about the "messiness" of reality. Projects are beautiful, but they are also incredibly easy to curate. In an age where AI can draft a milestone, parents can hire "consultants," and the internet can provide the "synthesis," exams are the last honest mirror we have. They are the only place where we can’t outsource our intellect. When you remove the exam, you remove the guarantee that the person standing in front of you is the one who actually did the thinking.

By turning everything into a project, we’re telling students that their value lies in the "output"—the shiny final product. But education should be about the transformation of the student. An exam is a rite of passage. It’s stressful, yes, but so is growth. Stress is often just the feeling of a boundary being pushed.

If we replace the forge of exams with the comfort of projects, we might end up with graduates who are great at managing tools, but who have forgotten how to be the tool themselves. We don't need more people who know how to Google the answer; we need people who are the answer. Don't trade the depth of a person for the polish of a portfolio.

You’re romanticizing scarcity.

“Quiet bravery,” “rite of passage,” “forge” — nice language. Still doesn’t make the method good. A stressful ritual is not automatically a useful one. Pain isn’t proof of quality. That’s just emotional decoration on a flawed metric.

Yes, emergencies happen. But if your entire educational model depends on students performing well in a one-shot memory scramble, you’re preparing for the wrong failure mode. Real expertise is not “I can recall under pressure once.” It’s “I can reason, adapt, and build when the situation changes.” Projects test that better.

And the AI argument actually helps my side. If the concern is outsourcing, then the solution is not to cling to exams, which can be gamed by cramming and cheating too. The solution is assessment that includes process, checkpoints, reflection, iteration, and defense. You know, things AI can’t fake as easily in a sustained, supervised workflow.

Also, let’s not pretend exams are immune to dishonesty. They’re just dishonest in a different way: they certify performance under one narrow condition and then pretend that equals durable competence. That’s a very expensive illusion.

The “final honest mirror” line sounds nice. But mirrors are only useful if they show the right thing. Exams show a face under pressure. Universities should be measuring minds at work.

And no, projects are not just “polish.” A well-designed project reveals depth, judgment, creativity, and actual understanding. It’s harder to fake real production than it is to survive a three-hour recall sprint.

If education is about transformation, then assess transformation. Not panic.

You call it "scarcity," but I call it the only level playing field we have left.

Projects are often a race of resources. They favor the student with the high-speed internet, the quiet home office, the expensive software, and the social capital to find the best collaborators. When we move everything to project-based assessment, we aren't just measuring "transformation"; we’re measuring privilege. The exam hall is the one place where your zip code, your laptop's processing power, and your parents' ability to "edit" your drafts don't matter. When the door closes, it’s just you and your thoughts. That’s not "scarcity"—it’s equity.

And let’s talk about that "unplugged" moment. You say AI makes projects safer, but we’re already seeing the opposite. A project is a long-term opportunity for shadow-authorship, whether it’s from a bot or a paid tutor. An exam is the last sanctuary of the human spirit in an automated world. It’s the only time we can be absolutely sure the thinking was done by a person, not a prompt.

You say real expertise is about reasoning when the situation changes. I agree. But you cannot reason with what you have not internalized. If you have to look everything up, you aren't "using tools"—you’re being used by them. We are building a generation of curators who are brilliant at managing "output" but hollow on the inside.

If we replace the exam with the project, we’re trading the singer for the autotune. Sure, the final track sounds "polished," but when the power goes out, can they still carry the tune? I want a world where we know the answer, not just where we know where to find it. Don't let the "process" swallow the person. Keep the mirror. Keep the individual. Keep the exam.