Alex Imas didn’t chase optimism. He wrestled with dread first. The University of Chicago economist, who pores over AI’s labor punch, spent years mapping doomsday paths—job wipeouts, wage crashes, demand implosions. Now, in his Substack Ghosts of Electricity, he charts a counter-scenario. Scarcity doesn’t vanish. It relocates.
“The economics of scarcity won’t disappear, it’ll just relocate,” Imas writes in his essay What Will Be Scarce?. AI floods the world with cheap goods. Production commoditizes. Humans? They become the premium. Nurses drawing bed-side chats. Baristas flashing genuine smiles. Therapists unpacking real pain. These relational roles—already employing 50 million Americans—swell as incomes climb and desires mimic exclusivity.
History backs him. Farms claimed 40% of U.S. workers in 1900. Today? Under 2%. Output boomed, yet labor fled to services with elastic demand. A 2021 Econometrica paper by Diego Comin, Danial Lashkari, and Martí Mestieri pins over 75% of that shift on income effects. Richer folks crave more human touch—dining out, live shows, education. Top earners spend 4.3 times more overall, but ratios explode in relational categories, per BLS data.
Imas’s experiments seal it. With Kristof Madarasz, he showed willingness-to-pay doubles when goods exclude others—mimetic desire, Girard-style. Follow-up with Graelin Mandel: AI-tainted art loses premium allure; humans hold value because they can’t copy infinitely. Starbucks gets it. CEO Brian Niccol pushed handwritten notes and ceramic cups, ditching automation for that human spark. The $112 billion chain reversed course, hiring more baristas for the details AI can’t fake.
But. Speed kills the dream. Past shifts unfolded over decades. AI? It’s jagged, probabilistic. Imas calls it that on a recent Bloomberg Odd Lots podcast. “These systems have been showing incredible improvements,” he told the New York Times. Valleys rise fast. Truckers and warehouse crews face walls of robots already. White-collar entry jobs? Dario Amodei eyes 50% evaporation soon, echoing cuts at Meta, Microsoft, Salesforce.
Doomsday’s Shadow Looms Large
Imas takes collapse risks dead seriously—unlike complacent peers. Citrini Research’s “ghost GDP” parable haunts: AI automates fishing on an island. Owners gorge on 200 fish. Machines churn 10,000. Workers starve, demand craters. In models, GDP halves, output spirals down at zero rates. His own OLG simulation? Wages drop 25%, capital stock 65%, national income 4.2% below baseline. Yet he deems it improbable. Transitions lag. Policies adapt. New wants emerge—AI pals, virtual worlds. Satiation? Rare in humans.
Still, economists wake up. When ChatGPT dropped, Imas shifted: “It was just a paradigm shift for me,” he said, per the New York Times. A fresh survey flags faster growth but inequality spikes, millions displaced. Daniel Rock of Penn: “I don’t think A.I. has hit the labor market yet… but I think it’s coming.” Imas trains execs one-on-one now. “My primary role right now is to sit people down… and get them trained on top-flight technology.”
Wall Street nods. Morgan Stanley cites Imas. His Substack surprises him—rigor for economists, accessible to all. A medical data scientist gripes: AI prototypes shine, but scaling? Bureaucracy blocks. Humans navigate the rapport no bot matches. Jobs evolve. Widget-makers fade. But truckers? Pure tasks, no chat. Others tilt relational as AI strips drudgery.
“The speed of change really matters,” Imas warns in Fortune, “whether we get to this hopeful version versus the more worrisome one.” Billionaires crave podcasts, X threads, live gigs—not endless movies on solo islands. Provenance. Presence. That’s the draw. AI abundance amplifies it.
Policy and the Human Edge
And here’s the rub. Disruption breeds winners, losers. Imas, replying to Garry Kasparov on X, agrees: History nets more jobs, but unevenly. “The hope is that we have a better understanding of economics than in the 19th century, so that disruption can be managed.” Social nets. Retraining bursts. Tax shifts from labor to land, capital—fund the relational boom.
Real-world tests mount. A data whiz at a med firm: Docs grant 15 minutes max. Precision rules. No AI apes that trust register. Starbucks bets on hospitality driving stays. If gradual, labor flows. Relational demand surges—no ceiling, as incomes balloon and exclusion stokes want.
Imas arrived scared. “My first reaction was to be very scared. I needed to work things out carefully.” He did. History. Preferences. Models. The verdict? AI doesn’t end work. It revalues humans. But rush it, and ghost GDP bites. Watch the pace. Economists, technologists, policymakers—take note. The relational economy awaits. Or the drag.
WebProNews is an iEntry Publication