University Isn't a Job Factory — and Computer Science Isn't Ending

Every few months, the same two conversations flare up again.

First: "University should prepare you for industry. If it doesn't, it's failing."

Second: "AI is coming for software jobs. Is computer science still worth it?"

I disagree with the framing behind both, and I want to explain why. This isn't a defense of universities as they currently operate, and it's not blind optimism about the job market either. It's simply a claim about what universities are for, and about what computer science becomes in an AI-saturated world.

1) University is not meant to be a company's training pipeline

Many companies talk about universities as if they're human-resource factories: a conveyor belt that takes in teenagers and outputs "ready-to-hire" employees.

That story is convenient for business. It turns education into a supply chain problem: define the job requirements, ask universities to match them, and label anything else "inefficient."

But that isn't the purpose of a university, and framing it this way misses what actually happens during those years.

University occupies one of the most formative stretches of a person's life. For many students, it's the first time they:

live among strangers and learn how to coexist with people who think differently
navigate various relationships: professors, advisors, competition, authority
test their confidence against failure and ambiguity
discover what they actually care about when nobody is telling them what to care about
begin forming the inner "core" they will carry for decades: values, worldview, taste, discipline

A company can teach you its stack. A bootcamp can teach syntax. You can learn tools on YouTube.

But something different happens when you spend years in an environment built, at least in principle, for inquiry, mentorship, and depth. When it works, university is one of the few places where the people around you are not trying to sell you something. Their job is to teach, to question, and to push you to think more carefully than you would on your own.

"But students already know what they want in high school."

Sometimes. Many students arrive with strong opinions about what they'll do and who they want to be. High school can create that sense of certainty.

But for a lot of people, that certainty isn't a core. It's a draft.

University is where many students discover what their beliefs look like when they're challenged by:

history that complicates simple narratives
mathematics that refuses hand-waving
literature that forces empathy across unfamiliar lives
research that teaches humility, because reality doesn't bend to confidence
peers who are smarter, faster, kinder, harsher, more complex than the "school bubble" suggested

That isn't wasted time. That is the training. Not company training — but life training.

What about "universities will disappear" because of AI?

I don't see a clear reason AI changes the core purpose of universities.

AI will absolutely change how we learn. It will shorten the time needed for certain skills. It will individualize practice. It will raise the baseline of competence.

But none of that replaces what a university is meant to provide: a structured, mentored environment where young adults develop judgment, intellectual discipline, and a foundation that lasts.

If anything, AI makes that environment more important. When information is abundant and persuasion is automated, judgment becomes more valuable, not less.

What should change is this: universities need to integrate AI seriously and quickly.

Not as a "cheating problem," and not as a gimmick. AI should become a normal tool in education, like calculators, search engines, and IDEs did. Students should graduate with:

strong fundamentals: writing, math, reasoning, ethics, collaboration
the ability to use AI responsibly and effectively in real work

That isn't turning universities into job factories. It's preparing students for the world they're actually entering.

A note on Silicon Valley's anti-university rhetoric

There's a popular line online: "Don't go to university. Just start a company."

For a small number of exceptional people, that path works. In any large population, you'll always find outliers. We highlight those stories because they're simple and dramatic.

But for most people, it's unrealistic to expect eighteen-year-olds to understand markets, institutions, incentives, and long-term tradeoffs well enough to make that bet reliably.

And if we want that to be realistic, we'd have to redesign middle school and high school entirely. Otherwise, telling most students to skip university isn't a strategy. It's a gamble.

2) Computer science jobs aren't "dying" — they're shifting

Now to the second fear: the job market for computer science in an AI world.

People see layoffs and AI copilots writing code. They conclude: "CS is over."

I see something different. Not an ending, but a shift in terrain.

A cliff feels scary because familiar paths disappear. But beyond a cliff isn't emptiness. It's new ground that hasn't been built yet.

AI needs an AI-ready world

If AI becomes as fundamental as electricity, then the world has to adapt around it.

Electricity didn't eliminate work. It created grids, standards, new industries, safety regulation, and entire professions.

AI will require:

infrastructure that runs reliably at scale
data pipelines that are lawful and traceable
security models that assume adversarial inputs
evaluation frameworks that measure correctness and safety
governance structures that constrain powerful systems
integration with domains like health, law, finance, and education

These aren't side tasks. They're core work. And that work doesn't complete itself.

"But won't the agents build everything?"

Even if agents help build parts of this world, responsibility remains human.

Ask a basic question: what does the future look like without people building it?

Who designs the compute infrastructure? Who ensures privacy and audit trails? Who handles outages and rollback plans? Who defines evaluation standards before harm occurs? Who integrates with messy, real-world databases and workflows?

AI can generate code. It can accelerate iteration. It can make teams far more efficient.

But greater efficiency doesn't automatically reduce total work. That only holds if the amount of work is fixed. In this case, it isn't. New capability creates new demand. We attempt problems we previously avoided because they were too expensive.

Yes, some traditional, repetitive programming work may shrink. That's real.

But that doesn't mean computer science has reached a dead end. It means the center of gravity is moving.

What makes CS students different?

In an AI era, the advantage of a computer science student isn't just "I can code."

It's: I understand systems deeply enough to make AI work in the real world.

That includes:

systems thinking: performance, scaling, tradeoffs
security thinking: threat models, misuse, adversarial behavior
correctness thinking: testing, monitoring, debugging
data thinking: quality, provenance, governance
modeling thinking: knowing where systems fail
product thinking: understanding how humans actually use tools

Computer science becomes the discipline that turns AI from a demo into infrastructure.

So what should we actually worry about?

My concern isn't whether students should study computer science. I still encourage it, especially now.

The harder issue is what happens to roles that are narrow, repetitive, and easy to specify. How do those workers transition? What do new roles look like? How should education evolve so people aren't left behind?

That's not just a CS problem. It's a societal one.

Universities matter here too — not because they're job factories, but because they're one of the few institutions capable of helping large numbers of people adapt, learn new tools, and rebuild confidence in a changing world.

Closing thought

University isn't a production line for companies. It's where many people build the foundation of their adult life.

Computer science isn't dying. It's reorganizing itself around a new technological layer, just as earlier eras reorganized around electricity, the internet, and mobile computing.

The cliff is real. But beyond it, there is more work than we can currently imagine — and we will need people who can build it.