Are Computer Science Majors "Cooked" because of AI?

Introduction

I wanted to write about this topic, mainly for my own understanding, but I figure this might help shed some light on the current state of the software engineering job market as well. I'll answer the question: "Are CS majors cooked because of AI?" and you'll find that, while it's definitely a difficult market, AI really isn't the main culprit or even the top factor.

I think it really boils down to four key issues: historical unemployment trends that predate AI by years, massive market oversaturation from the "learn to code" bubble, structural hiring problems that eliminate true entry-level positions, and the convenient scapegoating of AI to avoid addressing deeper market failures. Now, let's dig into the actual data.

TL;DR

• CS majors do have high unemployment (6.1%) - ranking 7th highest among all college majors (A Popular College Major Has One of The Highest Unemployment, College majors with the best, worst job prospects)
• AI is NOT the cause, CS unemployment was already highest in 2018 (5.6%) long before ChatGPT took off in popularity (Computer Science had a high unemployment rate even in 2018)
• Real cause is market oversaturation - CS degrees doubled from 51,696 to 112,720 (2013-2023) (Computer Science Graduates Face Worst Job Market in Decades)

The Historical Context: CS Unemployment Was Bad Before AI

The timing here proves the AI story is wrong. Computer science majors already had the highest unemployment rate among all college majors in 2018 at 5.6% (Computer Science had a high unemployment rate even in 2018), a full four years before ChatGPT was released to the public. This wasn't a one-time anomaly either, it was a consistent pattern that researchers were already documenting.

The data from the Employment America analysis is where the AI argument falls apart: Recent graduates had worse job outcomes way before AI came around (Don't Blame AI For The Rise in Recent Graduate Unemployment). Their research shows that recent college graduates actually held worse unemployment rates than the general working-age population in the lead-up to the pandemic and again in early 2021, well before AI tools became mainstream.

Even more telling, the employment edge for recent college graduates had been eroding since the 2010s. The gap between recent graduate employment rates and all workers peaked at 12.1% in summer 2021, then fell 0.6 percentage points by the time ChatGPT was released in November 2022. This suggests market forces completely independent of AI were already reshaping entry-level employment.

The pattern becomes clear when you look at broader tech unemployment data. During the height of the pandemic in April 2020, when AI was barely a consideration for most companies, the unemployment rate for computer occupations actually declined to 2.8% while all other occupations jumped to 15% (Analysis of Employment Data for Computer Occupations). Yet by 2018, during a supposedly strong economy, CS majors were already struggling with higher unemployment rates.

The Real Culprit: The "Learn to Code" Bubble

The true driver behind CS unemployment becomes obvious when you examine enrollment and graduation numbers. Computer science degrees more than doubled from 51,696 in 2013-2014 to 112,720 in 2022-2023 (Computer Science Graduates Face Worst Job Market in Decades). That's a 118% increase in just a decade.

That's not natural growth based on what companies actually needed. It's the result of what I call the "learn to code" bubble, which is a everyone telling people that sold computer science as a guaranteed path to prosperity. Universities, politicians, and yes, tech executives themselves spent the 2010s telling everyone that coding was the future and CS degrees were golden tickets. What did you think was going to happen?

The numbers tell the story of a market flood. Computer science showed the highest increase in bachelor's degree earners among all majors - up 4.3% year-over-year even as overall bachelor's degrees declined 3.0% (Computer Science Has Highest Increase in Bachelor's Earners). While other fields were contracting, CS kept pumping out graduates at an unsustainable rate.

Coding bootcamps amplified everything. The bootcamp market exploded from virtually nothing to graduating 24,975 students by 2020, representing 1047% growth since 2013 (2020 Coding Bootcamp Market Size Study). University-based bootcamps alone grew from 5 in 2015 to 120 in 2020, with the total bootcamp industry generating over $700 million in revenue by recent estimates (The Rise and Fall(?) of Coding Bootcamps).

What makes this especially brutal is the timing. As one analysis I read noted, "universities doubled CS enrollment just as demand collapsed". The market was simultaneously flooded with new graduates while experiencing the post-pandemic correction and economic pressures that reduced hiring. It's basic supply and demand. When you double the supply of something while demand stays flat or declines, prices (and employment rates) crash.

The "learn to code" narrative was so strong that it created a vicious cycle. More students enrolled because they heard CS was safe and lucrative, which led to more graduates competing for the same number of positions, which drove down entry-level opportunities and wages, which made the field less attractive to students, but by then millions were already in the pipeline.

Structural Market Problems: Experience Inflation and Geographic Concentration

Beyond the oversupply issue, the CS job market suffers from structural problems that make entry-level hiring more difficult. The most glaring is what I call "experience inflation", the paradox where supposedly "entry-level" positions require 2-3 years of experience (The entry level tech job crisis).

An analysis of LinkedIn job postings found that for most entry-level jobs, an average of two years of experience was required. The only postings that truly require no experience are internships, which make up only a small fraction of job postings and usually require current enrollment in an undergraduate program. This creates a paradox: you need experience to get experience.

Why does this happen? Companies use years of experience as a convenient filtering mechanism, even though "experience inflation" means entry-level requirements keep creeping upward as the job market becomes more competitive. What used to require zero years of experience now requires 2-3 years, similar to how housing and other costs inflate over time.

The geographic concentration problem makes this worse. The San Francisco Bay Area employs 407,810 tech workers, representing 11.6% of total Bay Area employment - more than double the 50-market average of 5.6% (San Francisco Bay Area Ranks #1 in CBRE's Annual 'Scoring Tech Talent Report'). This concentration means that 76.6% of the Bay Area's software engineers work in the tech industry, the highest concentration in the U.S.

This creates fake scarcity. Most tech jobs are concentrated in a handful of expensive metros: San Francisco, Seattle, New York, Austin, where the cost of living makes entry-level salaries inadequate. Meanwhile, companies in these areas can be extremely picky because they have massive applicant pools.

The result is a divided market: experienced developers in tech hubs get high pay and have multiple opportunities, while new graduates compete for a small number of true entry-level positions against hundreds of other applicants. Companies can demand unrealistic experience levels because they know someone will meet those requirements.

The data shows this clearly: new grads now account for just 7% of hires at Big Tech firms - down 25% from 2023 (The State of the Software Engineering Job Market for 2025). Companies have essentially abandoned entry-level hiring in favor of poaching experienced talent from each other, leaving new graduates to fight over scraps.

AI as Scapegoat: Why the Timeline Doesn't Add Up

When you examine the timeline closely, the AI explanation falls apart. ChatGPT was released in November 2022, but CS unemployment problems were well-documented years earlier. Tech layoffs began in earnest in 2022, primarily due to pandemic over-hiring corrections and rising interest rates - economic factors that had nothing to do with AI capabilities.

The scapegoating becomes obvious when you look at what companies actually say. "Tech CEOs might invoke AI in earnings calls to justify 'cutting' their teams, but analysts note this often masks deeper issues like prior over-hiring or inefficient projects" (Layoffs & AI: The Impact on Software Engineering (2023–2025)). As one economist put it: "Tech companies are laying off because of economic uncertainty + high interest rates. AI has become a convenient explanation."

The numbers support this. Between 2022 and 2024, more than 500,000 tech workers were laid off (Q&A: UW researcher discusses the "cruel optimism" of tech industry...), but only 27,000 job cuts have been directly tied to the advent of AI since 2023 (AI is leading to thousands of job losses, report finds) according to Challenger, Gray & Christmas data. That means AI accounts for roughly 5% of tech layoffs - hardly the dominant factor the narrative suggests.

Even more telling, many of the 2023-2024 layoffs hit non-technical departments - HR, recruiting, marketing, and middle-management layers - rather than core engineering teams. Google's 2023 cuts largely affected People Operations, Amazon cut communications and support roles, and Meta eliminated project management functions. These companies often retained their critical developers while slashing other staff.

The AI adoption timeline also doesn't support mass displacement. According to the Census Business Trends and Outlook Survey, just 3.7% of firms reported using AI in September 2023. By late 2024, 23% of employed workers used generative AI for work at least once per week (Is AI Contributing to Rising Unemployment?) - significant adoption, but hardly the kind of widespread use that would justify massive engineering layoffs.

The reality is that "AI sometimes became a convenient scapegoat to justify tough decisions, but the reality is that many layoffs would have happened with or without the latest AI tools." It's much easier for a CEO to say "AI is transforming our business" than to admit "we made terrible hiring and strategic decisions during the pandemic boom."

Conclusion: Market Correction, Not AI Apocalypse

It's pretty clear: computer science majors aren't "cooked" because of AI - they're facing the crash after years of hype. The unemployment rate of 6.1% for CS majors is certainly concerning, but it's what happens when you flood the market with graduates while actual demand stagnated.

The real factors driving CS unemployment are straightforward: way too many graduates from doubled enrollment and tons of bootcamps, experience inflation that eliminates true entry-level positions, geographic concentration that creates fake scarcity, and economic pressures that reduced overall tech hiring. AI became a convenient narrative to explain these problems, but the timeline and data don't support it as a primary cause.

This doesn't mean AI will never impact software engineering jobs... it probably will, eventually. But the current crisis isn't about AI replacing developers. It's about a market that encouraged too many people to pursue the same career path at the same time, creating a supply-demand imbalance that was always going to end badly.

The "learn to code" bubble has burst, and we're seeing the hangover. The good news is that this is likely a temporary correction. Software is still eating the world, demand for quality developers remains strong, and the inflated graduation rates are already starting to moderate. The bad news is that it may take several years for the market to rebalance, and current graduates will unfortunately bear the brunt of that adjustment.

For CS majors facing this market, the solution isn't to panic about AI, it's to differentiate yourself in an oversaturated field through specialized skills, real experience (internships, projects, contributions), and persistence in a temporarily difficult market. Software engineering is still a good field; it's just no longer the easy path to prosperity that everyone was promised.

So for now, don't panic about AI if you're already in the CS market. Just focus on 1) building projects, 2) networking, and 3) Leetcode (sorry). It's all worth it, there's still not any other career I can think of where you can make $180,000+/year right out of college. Here are some tips (plug for my own article): New Grad Job Search for Software Engineers.