The Numbers First

Undergraduate CS enrollment dropped more than 8 percent last year — the largest absolute decline of any major, according to The Atlantic. Graduate-level enrollment fell even harder: 14 percent.

Students are voting with their feet. The internet is flooded with videos from new CS grads who can't find work, and the top comment on one popular YouTube rant says it all: "Your first mistake is not being born earlier."

The AI Factor Is Real — But Misread

Anthropic co-founder Jack Clark recently said about 90 percent of the company's new code is now AI-generated and warned that "the value of more junior people is a bit more dubious." That's a direct gut-punch to anyone studying CS hoping to land an entry-level job.

John Coogan, co-host of the tech podcast TBPN, publicly floated whether studying CS is now "contrarian" given that coding jobs are disappearing. The Economist told students last week to "forget Python, study Plato."

But the coverage misses something important.

What the Panic Merchants Are Getting Wrong

CS grads still significantly outearn graduates from nearly every other major, according to The Atlantic. Yes, their unemployment rate is spiking. But their underemployment rate — people stuck in jobs that don't require a degree — remains comparatively low. Nearly half of philosophy majors end up underemployed. The figure for CS grads is a fraction of that.

One plausible explanation: CS grads aren't settling for barista jobs. They're holding out for real positions. That inflates the unemployment numbers without telling you anything about the long-term value of the degree.

The employment data currently circulating tracks class of 2024 graduates. AI capabilities have improved substantially since those students graduated. The threat is real and accelerating. But the data being used to declare CS a dead end is already outdated.

Schools Are Finally Waking Up — Slowly

While students flee, at least some institutions are trying to catch up. UC San Diego launched the GenAI in CS Education Consortium in October 2025, backed by $1.8 million from Google.org — Google's philanthropic arm — as part of a broader $1 billion commitment to U.S. education Google recently announced, according to UC San Diego's own reporting.

The consortium pulls in researchers from UCSD's Jacobs School of Engineering and School of Social Sciences, plus faculty from the University of Toronto. The stated goal: help educators worldwide teach CS in a way that makes sense in the AI era. The effort claims to reach thousands of educators across dozens of countries, who in turn will reach tens of thousands of students.

Whether $1.8 million genuinely moves the needle across a global education system remains unclear.

Separately, according to codeeasy.io, schools are increasingly integrating machine learning, natural language processing, and robotics into core curricula — shifting away from teaching students to write boilerplate code toward teaching them to build and direct AI systems.

The Real Story on Campus

College students booed multiple commencement speakers this spring for mentioning AI, including former Google CEO Eric Schmidt at the University of Arizona, real-estate executive Gloria Caulfield at the University of Central Florida, and Nashville record exec Scott Borchetta at Middle Tennessee State University.

NBC News called AI "unpopular" with graduates. The Wall Street Journal declared an "American Rebellion Against AI." Fox News framed it as grads telling Schmidt exactly what they thought of him.

All of that is spin, according to The Atlantic's Jeffrey Selingo, a university professor and administrator. His read: students use AI constantly. They cheat with it. They depend on it. The booing isn't a coherent rebellion against technology — it's what he calls "a cosmic howl" against uncertain futures, debt, and a higher education system that often failed them before AI ever showed up.

The media turned boos at graduation ceremonies into a political narrative. The reality is messier and more human than that.

What This Actually Means

The real problem isn't that CS is dying. It's that the version of CS education built around teaching people to write code from scratch is becoming obsolete — and universities are years behind in building what replaces it.

The students who will win in this market aren't the ones who can type Python faster than a chatbot. They're the ones who understand systems, architecture, data, and problem-solving well enough to direct AI tools and catch their mistakes.

That requires a better CS education, not less of it.

If you're a 17-year-old deciding whether to major in CS right now, look at the wage data. Look at the underemployment data. Then look hard at what specific schools are actually teaching — because the gap between a program that prepares you for 2030 and one stuck in 2015 is enormous.

The degree still has value. The question is whether the school selling it to you has figured out what that value actually looks like now.