Don't Study Computer Science

A friend told me last week his son decided to study computer science at a university in Buenos Aires. The kid is excited. The parents are proud — and quietly calculating the cost of flights, rent, and living away for years. They asked me, the “computer guy” they know, what I thought.

A year ago, I would have hesitated. I would have mumbled something about “the industry is changing” and “maybe consider data science instead.” I would have tried to soften the blow while planting doubt.

Today? Today I’m honest.

“Don’t do it. Not if you want to be a programmer. That career is dying.”

The Awkward Truth

They looked at me like I’d insulted their religion. And I get it — I’ve been coding for over a decade. I built a career, a company, a life around this skill. Telling someone not to follow that path feels like betrayal.

But betrayal of what? The past? An idea?

Here’s what I didn’t say, but should have: coding is a solved problem. Not perfectly, not completely, but solved enough. The AI writes functional code in seconds. It debugs better than most juniors. It never gets tired, never asks for a raise, never needs mentorship.

The entry-level programming job — the one that used to train generations of developers — is vanishing. And with it, the traditional path into this profession.

The Paradox

But here’s the part that keeps me up at night: if nobody becomes a junior, what happens to the seniors?

Every profession needs new blood. Doctors need residents. Lawyers need clerks. Engineers need apprentices. The knowledge doesn’t transfer through osmosis; it requires people learning at the bottom rung, making mistakes, growing into the role.

If we dissuade an entire generation from studying computer science — if we succeed in warning them away — who maintains the systems in ten years? Who understands the legacy code? Who becomes the architect that the AI assists?

I’m telling kids not to enter a field that desperately needs them. And I don’t know how to resolve that contradiction.

What I Should Have Said

Maybe the advice isn’t “don’t study systems.” Maybe it’s “don’t study systems to become a code writer.”

The field isn’t dying. It’s morphing. What we need now — what we’ll desperately need in five years — are people who understand systems, not just syntax. People who can think architecturally, who understand business context, who can translate human problems into technical requirements.

We don’t need more people who can write a for-loop. We need people who can decide whether to write a for-loop, where it should live, and what happens when it breaks.

So study systems, yes. But study them like you’d study philosophy — to learn how to think, not to acquire a specific trade skill that might be obsolete by graduation.

The Generational Gap

There’s another layer to this, one I’m acutely aware of at 49 with nearly three decades in tech (my first Linux install was 1996). The tools I learned on — the ones that shaped my understanding — are archaeological artifacts now.

But the way I learned them matters. I learned by breaking things. By spending nights trying to understand why a server wouldn’t boot. By writing code that failed spectacularly and having to figure out why.

Will the next generation have that experience? Can you develop deep intuition for systems if you’ve never struggled with them? If an AI always provides the answer, do you ever develop the muscle of debugging, of hypothesis-testing, of that sixth sense that tells you “something smells wrong here”?

I don’t know. And that’s terrifying.

The Uncomfortable Middle

So I’m stuck in this uncomfortable middle place. When parents ask, I tell the truth: the traditional programming career is disappearing. Don’t expect the path I walked to still be there — but know that I don’t regret a single step of it.

But I also know — I know — that if everyone takes my advice, we’re building a cliff. In ten years, when the current generation of senior engineers retires or moves on, there will be no one to replace them. The AIs will be powerful, yes, but they’ll need direction. Context. Wisdom. And that comes from experience that no one will have.

It’s like watching a slow-motion extinction and not knowing whether to warn the species or let nature take its course.

What I’m Actually Afraid Of

Let me be more vulnerable than I’m comfortable with: I’m not just worried about the industry. I’m worried about what we’re losing.

Every time I tell someone not to study programming, I’m acknowledging that the world that formed me — the world of tcpdump and Stevens’ TCP/IP Illustrated, of Perl one-liners and ipfilter queues, of building HA systems from first principles — that world is fading. Not because it wasn’t valuable, but because the terrain has shifted.

But staying silent feels worse. Watching bright-eyed kids invest years and money into learning a trade that’s being automated feels like watching someone train to be a blacksmith in 1920. The automobiles are already on the roads. The change isn’t coming. It’s here.

The Real Answer (That I Don’t Have)

I wish I could end this with a clear prescription. “Study this instead.” “Focus on that.” But I don’t have answers. The ground is shifting under all of us, and anyone who claims certainty is selling something.

Maybe the right advice is: study what fascinates you, but don’t expect the credential to guarantee a job. Learn to think, to communicate, to solve problems creatively. The tools will change — they always do — but those core skills might survive the transition.

Or maybe that’s just cope. Maybe I’m trying to salvage meaning from a profession that’s being hollowed out.

The Conversation I Keep Having

It keeps happening. Different parents, different kids, same dilemma. Sometimes I soften the message. Sometimes I’m brutally direct. I haven’t found the right balance.

One father pushed back hard: “But AI can’t replace everything. Someone has to build the AI.”

“Sure,” I said. “Maybe 1% of the people who study systems will build AI. The other 99% will compete for the shrinking jobs that remain. Are you comfortable betting your kid’s future on being in that 1%?”

He wasn’t. Most parents aren’t.

What I Hope

I hope I’m wrong. I hope this is like the automation scares of the past — the Luddites, the assembly line fears, the outsourcing panic — where new jobs emerged to replace the old ones. I hope that by warning people away, I’m creating a shortage that makes the career viable again. Irony upon irony.

But I don’t think I’m wrong. Not this time. The technology is different. The pace is different. The entire paradigm of “learn a skill, practice it for 40 years, retire” is crumbling, and programming is just the canary in the coal mine.

The Hardest Part

The hardest part isn’t the technical change. It’s the human one. These kids are excited. They grew up with technology, they love it, they want to build things. And I’m standing there like the ghost of Christmas future, telling them their passion has an expiration date.

Who am I to crush that? But who am I to lie?

So I tell the truth, and I watch their faces fall, and I wonder if I’m helping or just spreading anxiety. If I’m saving them from a dead end or closing a door that might have led somewhere unexpected.

Afterword

The kid from last week? He’s still going. His parents decided to let him follow his passion, consequences be damned. I admire that. I hope he proves me wrong. I hope he finds a path I can’t see.

But I also know — with the certainty of someone who’s watched this industry for three decades — that the path I took is closing behind me. And pretending otherwise helps no one.

So I’ll keep having these conversations. Awkward, honest, necessary. Telling the truth as I see it, even when the truth is uncomfortable. Even when I don’t know if I’m right.

Especially then.

If you’re reading this and you’re the kid whose parents showed them this post: I’m sorry. I really am. I wish I had better news. But the world changes, and we change with it, or we get left behind.

Study systems if you must. But study them with eyes open. Study them to understand, not just to execute. Study them because the thinking matters, not because the credential guarantees anything.

And maybe — just maybe — you’ll find something in that study that the AI can’t replicate. Something human. Something essential.

I hope so. God, I hope so.