They Booed Eric Schmidt. The Labor Data Had Been Building to That Moment for Two Years.

A generation sold on credentialism graduated into a market where the entry-level jobs that credential was supposed to unlock are contracting. The AI executives telling them to adapt helped build the tools doing the contracting.

On the night of May 18, former Google CEO Eric Schmidt took the stage at the University of Arizona's commencement ceremony and got booed every time he said "artificial intelligence." He told the graduating class AI was unavoidable. He compared the moment to the PC revolution of the 1980s. The arena did not receive the comparison well. The video circulated widely enough to reach European press before the ceremony ended. It was the second commencement in a week to produce the same reaction.

The boos are easy to dismiss as anxiety or ingratitude. The labor market data makes them harder to dismiss.

The Promise That Got Priced In

For roughly two decades, the social contract for American higher education was specific. A degree was the floor, the credential that converted into employment stability, higher earnings and a defensible position through recessions and disruptions alike. The data supported it. College graduates consistently out-earned non-graduates, got hired faster and rode out downturns better. Parents, counselors and policy makers priced that promise into the $1.7 trillion in outstanding student loan debt currently sitting on the books.

Gen Z absorbed the contract in full. They enrolled in college at higher rates than any previous cohort. They graduated into the precise moment the entry-level roles that contract was supposed to unlock began contracting.

Where the labor market sits for new graduates as of Q1 2026:

• Average student debt at graduation: $37,000
• Entry-level corporate job listings: down 15% YoY (Handshake)
• Applications per entry-level listing: up 30% YoY (Handshake)
• Unemployment rate, college graduates ages 22-27: 5.8%
• National unemployment rate: 4.0%
• Spread between the two: widest in BLS data for the cohort in decades

Michael Macaluso, 22, graduated from the University of Connecticut with a mechanical engineering degree in May 2025. He submitted roughly 200 applications. He is now working as an assistant pool director at a country club in Eastchester, New York. His trajectory is common enough across the cohort that it has stopped reading as anomalous.

Where the Jobs Are Going

The mechanism is direct and well-documented. The roles disappearing fastest are the ones that require the least institutional knowledge and the most repeatable cognition. They are also the roles that have historically served as the first employer-funded training program for new graduates, the rung where a 22-year-old learns to do the work above them.

The categories contracting first:

• Research drafts and literature review
• Data summarization and basic financial analysis
• Entry-level code review and QA
• Client intake and documentation
• Junior paralegal and discovery support
• First-pass marketing copy and content production

LinkedIn's chief economic opportunity officer, Aneesh Raman, said in February 2026 that AI is "breaking" the entry-level layer. Employers using AI in their job descriptions have increased 400 percent over the past two years, according to Handshake's January 2026 labor market retrospective. The jobs being created are real, but they skew toward the power buildout, AI operations and model fine-tuning. They require different preparation than the degrees the current graduating class spent four years completing.

The displacement figures from major institutional research:

• Goldman Sachs, March 2026: AI reducing U.S. monthly payroll growth by 16,000 jobs
• Goldman Sachs, 2023 baseline: 300 million jobs globally exposed to automation
• McKinsey, November 2025: 40-60% of knowledge-work tasks reachable by current AI agents
• BLS, 2026: first-ever published displacement risk estimates by occupation

The Velocity Problem Schmidt's Analogy Skips Over

The comparison to the 1980s PC revolution was the part of Schmidt's argument that drew the loudest skepticism, and it deserves more scrutiny than it received in the room.

The PC transition played out over roughly twenty years. Workforce adaptation lagged the technology by a cycle, but it lagged by years, not quarters. Workers had time to retrain. Institutions had time to adjust curricula. Policy had time to respond.

The current transition is not running on that timeline.

The release cadence of the current AI cycle:

• Q1 2026 AI model releases: 255+ across more than 30 labs
• Average gap between major capability releases: roughly three days
• Parallel development tracks at frontier scale: OpenAI, Anthropic, Google DeepMind, Meta, xAI
• Models considered state-of-the-art in January 2026: superseded by April

Gen Z's assessment of the credential they hold:

• Believe a four-year degree will not protect them from AI displacement: 65% (BestColleges, April 2025)
• Say AI has already made their college degree irrelevant: 45% (Harris Poll/Indeed)
• Describe their degree as a waste of money: 51%
• Same figure among Boomers: 20%

Schmidt's timeline assumes adaptation can happen on the PC schedule. The graduates in the arena had already concluded it cannot.

The Environmental Arithmetic

Gen Z is the first cohort for whom climate change has been a background fact of their entire conscious education. Carbon footprint calculators were built into their school curricula. They grew up watching wildfire perimeters expand annually and listening to institutional pledges about net-zero targets.

They also know what the technology Schmidt was championing costs to run.

The resource profile of the current AI buildout:

• 2025 U.S. data center electricity consumption: equivalent to a mid-sized country
• IEA 2030 projection: figure doubles
• AI share of U.S. electricity by 2030: equivalent to adding another California to the grid (Food and Water Watch, February 2026)
• AI share of global emissions by 2030 under high-adoption scenario: 1.5-3% (PLOS One, March 2026)
• CO2 emissions from training xAI's Grok 4: 72,800 tons
• GPT-4o annual inference water consumption: exceeds drinking supply of 12 million people

Hyperscaler water consumption growth, most recent fiscal year:

• Microsoft: +34%
• Google: +17%
• Amazon: +9%

For a generation taught to take shorter showers and avoid single-use plastics, the resource profile of the technology being presented as the century's great opportunity lands as a specific kind of dissonance. The carbon commitments made in their names are in direct tension with the deployment being accelerated in their names.

Who Is Telling Them to Adapt

The executives championing AI's restructuring of the labor market are not, in the main, the workers absorbing the restructuring.

Schmidt is 70. His net worth is approximately $25 billion. His career was built before the capabilities he celebrated on May 18 existed. The Google he ran was a search company. The AI he was describing at a commencement ceremony is an autonomous agent economy. He is not the person who will spend the next decade navigating a job market shaped by tools that did not exist when he formed his career.

What the Gen Z cohort is reporting back about that asymmetry:

• View AI as a threat to career stability: 60% (Deloitte 15th Annual Gen Z & Millennial Survey, May 13, 2026)
• Describe their degree as a waste of money: 51% (Harris Poll, up 11pp in two years)
• Believe AI will eliminate their intended career path within ten years: 38%
• Trust tech industry leaders to manage the transition responsibly: 22%

The people being told to adapt are the ones carrying the debt. The people doing the telling built their careers on the credential those debts were supposed to fund.

Schmidt left the University of Arizona stage to applause from the faculty section.