Average Is Over (Revisited)

In the AI Story

Cowen built the thesis on the freestyle chess model: human-computer teams beat both the best humans and the best computers playing alone. The economy would evolve similarly—winners would be the best collaborators, not the most credentialed or traditionally skilled. The AI moment vindicates this with precision The Orange Pill could not have anticipated: the backend engineer who never wrote frontend code but built a complete feature in two days using Claude is Cowen's freestyle champion made flesh. She is not the best traditional programmer; she is the best collaborator, and collaboration—not execution mastery—is what the new economy rewards.

The bifurcation mechanism operates through the marginal utility shift. When AI provides competent execution at near-zero cost, the marginal value of median human execution approaches zero. Median knowledge work—the reliable, professional-standard output that junior associates, staff engineers, and analysts produced—was valuable when execution was scarce. When AI delivers equivalent quality for a hundred dollars monthly, the human producing that quality faces a market that has repriced her labor. Not worthless—there remain verification, judgment, and context-integration tasks AI handles poorly—but worth dramatically less than twelve months prior. Meanwhile, the workers whose contribution was always above the execution layer—product vision, strategic judgment, taste—face the opposite: their marginal value soars because they provide the scarce complement to abundant execution.

Cowen acknowledges the thesis has proven more accurate than even he expected, compressed onto a faster timeline. The 'average is over' he predicted over decades arrived in months. The legal profession discovered it when AI tools began producing first drafts junior associates spent days writing. Accounting discovered it when tax analyses generated in minutes. Software development discovered it most violently with the December 2025 threshold and the subsequent SaaSpocalypse. The speed of arrival has practical consequences: workers who might have had a decade to transition have months, producing the acute identity crises, the desperate retraining attempts, and the flight-to-the-woods exits that The Orange Pill documents as the fight-or-flight dichotomy of the AI moment.

Origin

Cowen developed the framework in the early 2010s observing three trends: chess computers surpassing humans, machine learning handling increasingly complex pattern recognition, and labor market polarization into high-skill/high-pay and low-skill/low-pay with a hollowing middle. Average Is Over: Powering America Beyond the Age of the Great Stagnation appeared in 2013, predicting that the comfortable median—stable employment, middle-class security—would erode as machines competently handled routine cognitive work. The title came from a Lake Wobegon inversion: in Garrison Keillor's fictional town, everyone is above average; in Cowen's economy, being average is no longer enough. The 2025-2026 AI moment has elevated the book from speculative economics to operational manual.

Key Ideas

Freestyle collaboration beats pure human or pure machine. The winners are not the most skilled humans or the most powerful AI but the best human-AI teams—collaboration as the decisive competitive advantage.

Median performance now competes with algorithmic adequacy. The comfortable middle—professional-standard output, credential-protected positions—faces compression as AI matches median quality at near-zero marginal cost.

Returns bifurcate rather than disappear. Total knowledge-work demand may rise as judgment requirements expand, but distribution splits violently: premiums for directors, compression for executors.

The transition timeline compressed catastrophically. What Cowen projected over decades arrived in months, eliminating the adjustment period that might have softened the identity and economic dislocations.