The credential reckoning is the structural event, visible in enrollment data and hiring patterns by the mid-2020s, in which the university degree lost its monopoly as the default sorting mechanism for professional labor markets. The reckoning has two components: AI makes the skills the credential certifies acquirable outside the institution, and AI makes direct capability assessment — through portfolio evaluation, project-based interviews, and AI-augmented testing — cheaper and more reliable than credential-based sorting. The combined effect is not the death of the degree but the end of its monopoly. Elite credentials retain value through network effects; vocational credentials retain value through direct training; the broad middle of American higher education, which was always selling a bundle of moderate training and moderate signaling, faces the most acute crisis.
The credentialing function grew with the twentieth-century expansion of the professional labor market. As the economy shifted from manufacturing to services, employers needed a mechanism for sorting an undifferentiated pool of applicants. The degree served efficiently — not by telling the employer what the candidate knew, but by signaling that the candidate had been selected by an institution with standards, endured four years of demands, and emerged with certification that served as proxy for intelligence, persistence, and institutional reliability. Economist Michael Spence formalized this as signaling theory: the degree's value lay not in knowledge certification but in selection and endurance.
The signaling function was always more important than the training function. Dale and Krueger's landmark study demonstrated that students admitted to highly selective universities but choosing less selective ones earned roughly the same as those who attended elite institutions — suggesting selection, not education, was the primary driver of earnings premiums. The credential was always somewhat circular: it certified qualities the student possessed before enrolling, and its value depended on the institution's prestige, which depended on selectivity, which depended on student quality, which had nothing to do with what the institution taught them after they arrived.
AI dissolves the monopoly in two directions. First, AI makes the certified skills acquirable outside the institution — Segal's Trivandrum engineers acquiring cross-domain capabilities in days, non-technical founders prototyping products over weekends. Second, AI makes direct capability assessment cheaper: employers who can evaluate portfolios, see what candidates built with AI augmentation and how they exercised judgment, possess a more informative signal than any transcript. Portfolio-based assessment was always possible in principle; AI makes it practical.
Segal's parallel to the software death cross illuminates the structure: in both cases, value migrated from what was being commoditized (code, credentials) to what could not be commoditized (ecosystem lock-in for software, network effects and judgment cultivation for universities). The institutions that survive are those whose value was never primarily in the commoditized layer. The institutions that do not survive are those — disproportionately in the broad middle of American higher education — whose value proposition was primarily the eroding layer.
The reckoning was underway before AI — total undergraduate enrollment in the United States had declined approximately fifteen percent from its 2010 peak by 2023 — but AI dramatically accelerated it. By 2025, major technology companies had shifted substantially toward skills-based hiring; portfolio assessment had become standard in creative and technical fields; and the SaaSpocalypse had demonstrated what institutional repricing looks like when a value layer commoditizes.
End of monopoly, not death of degree. Elite credentials retain network-effect value; the broad middle faces the acute crisis.
Signaling before training. The degree's historical value was primarily as a proxy for selection and persistence, not certification of specific knowledge.
Direct assessment cheaper. AI-augmented portfolio evaluation provides more informative signals than transcripts, at lower employer cost.
Hollowing of the middle. Regional universities and non-selective four-year colleges — selling bundled moderate training and moderate signaling — face the sharpest enrollment declines.
Institutional parallel to SaaS. The pattern matches the software death cross: value migrates from commoditized layer to ecosystem and judgment.
Defenders of the credential argue that elite employers still require degrees and that portfolio-based hiring favors candidates with existing social capital, potentially widening rather than narrowing access gaps. Critics respond that the credential was always a device for maintaining social reproduction under the guise of meritocratic sorting.