CS graduates are hitting a job market that has structurally changed around them. A Tapestry News analysis published March 15, 2026 puts the numbers together: entry-level tech job postings have fallen 67% in the US and 46% in the UK since 2022. Revelio Labs tracked a 35% overall decline in entry-level postings since January 2023. A Harvard study by researchers Hosseini and Lichtinger, analyzing 62 million workers across 285,000 firms, found that companies adopting generative AI hired 3.7 fewer junior workers per quarter than non-adopters, with junior employment dropping 7.7% within six quarters of AI adoption. CS graduates now face a 6.1% unemployment rate — higher than philosophy majors. SignalFire data shows a 50% decline in new role starts by people with less than one year of post-graduate experience at major tech firms between 2019 and 2024.

Tools like GitHub Copilot and Cursor automated the work that used to justify hiring a junior developer: boilerplate code, unit tests, debugging, and feature-building from specs. That was the on-ramp. Anthropic CEO Dario Amodei has forecast AI may eliminate 50% of entry-level jobs, a prediction reinforced by <a href="/news/2026-03-14-andrej-karpathy-scores-ai-exposure-of-342-us-occupations-using-gemini-flash-llm">analysis showing software developers rank highest among 342 US occupations in AI exposure</a>. CNBC has framed the deeper damage as the destruction of the career ladder — the junior-to-senior pathway that has historically powered upward mobility in tech. The Computing Research Association's survey of 134 institutions found 62% reported undergraduate computing enrollment declines in Fall 2025, with traditional CS, software engineering, and information systems hit hardest. AI and cybersecurity specializations bucked that trend, growing at 56% and 58% of institutions respectively.

Where people land on this depends heavily on what they thought college was for in the first place. Degree skeptics in the Thiel-Musk mold see the labor data as confirmation of something they already believed — that credentials were always overrated, and AI has simply made the case undeniable. They point to Thiel Fellowship alumni like Figma's Dylan Field and Scale AI's Lucy Guo, and to expanding apprenticeship pathways, as evidence that alternatives to the four-year track have always worked. A harder-nosed reading of the same data, though, doesn't dismiss the degree so much as identify CS specifically as the casualty: the job market collapse is concentrated enough, and the automation of junior work specific enough, that the degree's vocational logic has broken down even if the broader credential hasn't.

The counter-argument isn't nostalgic. Harvard's Michael Sandel and University of Chicago's Martha Nussbaum have separately argued that peer networks, mentorship, and the habits of civic reasoning universities cultivate aren't things a coding agent can replace. The labor data backs a version of this: Stanford produced 122 venture-backed founders in 2024, MIT 87, Harvard 73. Elite network effects haven't collapsed alongside junior hiring. MIT Sloan's Andrew McAfee has staked out the synthesis: the degree survives, but not alone. Without demonstrable AI fluency, a CS credential is no longer sufficient to get hired at the companies that are actually growing.

The agents displacing junior developers are not hypothetical futures — they are production tools running at scale right now, and their footprint is showing up consistently enough across Revelio Labs, SignalFire, the CRA survey, and the Harvard study that this looks structural, not a post-pandemic correction. The harder question for the agent ecosystem is whether anything rebuilds the pipeline. Senior engineers come from somewhere. If the junior role has been compressed out of existence, the industry is borrowing against a talent base it is no longer replenishing.