In The Complacent Class (2017), Tyler Cowen documented a multi-decade trend: Americans had stopped moving, stopped switching jobs, stopped starting companies, stopped taking the risks that had once defined the national character. Residential mobility fell to post-WWII lows. Interstate migration declined. Entrepreneurship rates dropped among younger cohorts. This was not laziness but rational response to an environment where the costs of change were high and expected benefits uncertain. Moving meant leaving accumulated social capital. Switching careers meant abandoning hard-won expertise. The complacent choice—staying put, optimizing existing arrangements—was individually defensible. Collectively, it produced fragility: a society of people who had arranged their lives around stability discovered that stability was not guaranteed, and when disruption arrived, they lacked the adaptive capacity the previous generation possessed.
Cowen's analysis identified three mechanisms producing complacency. First, matching technologies—from dating apps to real estate algorithms—allowed people to pre-optimize their environments, reducing the tolerance for friction and surprise. Second, regulatory accumulation and credentialing proliferation raised the switching costs for careers, making professional reinvention prohibitively expensive. Third, the psychological shift toward risk aversion that occurred after the 2008 financial crisis taught a generation that ambition was dangerous and preservation was wisdom. These mechanisms operated simultaneously, each reinforcing the others, producing a population that genuinely preferred the known to the unknown and the stable to the dynamic.
The complacent class confronts the AI moment as its most direct challenge. The senior engineer who spent twenty-five years accumulating expertise in a particular stack, language, and framework represents complacent optimization in its purest form: deepening within a domain rather than broadening across domains, maximizing the value of sunk costs rather than exploring new capability. This strategy was rational under the assumption that his specialized expertise would retain market value. When AI commoditizes that expertise, the sunk cost becomes a liability—not because the knowledge is useless, but because the identity built around it resists the transition to judgment-based contribution. The complacent equilibrium punished breadth and rewarded depth. The AI transition inverts the returns.
The response pattern Cowen predicts is visible in the December 2025 developer discourse. Denial came first—AI cannot really code, the output is superficial, experts will always be needed. Anger followed—this is theft, this is cheating, this devalues genuine skill. Bargaining appeared—perhaps AI handles routine work while experts do the complex work, preserving a division of labor. Depression is settling in now among mid-career professionals who recognize the trajectory and have not yet found a path forward. Acceptance, when it arrives, will look like identity reconstruction: the engineer stops being 'the person who writes the best code' and becomes 'the person who knows what code should exist.' The transition is psychologically equivalent to a midlife crisis compressed into twelve months.
Cowen's framework explains why the complacent populations are least prepared for AI despite being most exposed. The upper-middle class professionals who optimized their lives around credential-based stability—lawyers, doctors, engineers, financial analysts—have the most to lose from a repricing of credentialed skills and the least experience with the kind of rapid reinvention the AI moment demands. The poor have less sunk cost; the very wealthy have capital buffers. The complacent middle has neither the flexibility of the poor nor the security of the wealthy, and its entire identity infrastructure is built on the expectation that expertise, once earned, retains value. The AI transition is testing that expectation with an empiricism that no amount of credentialing can deflect.
Cowen developed the complacent class thesis through the 2010s as he observed declining American dynamism across every measurable dimension. The Complacent Class: The Self-Defeating Quest for the American Dream (2017) synthesized data on mobility, entrepreneurship, cultural homogenization, and political polarization into a unified diagnosis. The AI application is Cowen's own extension, first articulated in his 2024-2025 lectures and interviews where he noted that the populations most complacent—professionals who spent decades optimizing for stability—face the most acute disruption from AI. The framework dovetails with his earlier Average Is Over thesis: the middle that is complacent is the same middle that is hollowing.
Complacency is rational response to high switching costs. When career transitions are expensive and benefits uncertain, staying put is individually optimal—but collectively produces a society that cannot adapt when the environment shifts.
Sunk costs become psychological anchors. The complacent professional has invested decades in expertise that AI is repricing; accepting the repricing means accepting that the investment was, in some sense, stranded.
The complacent are most exposed to AI. Upper-middle professionals whose security depends on credential premiums face the sharpest compression, because their skills are exactly the cognitive execution AI performs cheaply.
Adaptation requires identity reconstruction. The transition from execution-identity to judgment-identity is not a skills upgrade but a psychological crisis—letting go of who you were to become what the market needs.