James van Geelen and Alap Shah of Citrini Research published "The 2028 Global Intelligence Crisis" a couple of days ago —a lengthy memo arguing that AI adoption will displace 50% of white-collar workers, collapse consumer spending, and trigger a 38% stock market crash by June 2028. They framed it explicitly as scenario analysis, not forecast. Within days it went viral on Substack. The Dow fell 822 points, the S&P 500 dropped 1%, Nasdaq fell 1.1%, and IBM plummeted 13%. A speculative thought experiment had moved capital.
1. Ghost GDP Is Real (The Citrini Thesis)
AI will displace workers faster than new jobs can absorb them, collapsing consumer spending and equity valuations.
White-collar employment is the spine of demand. Fifty percent of US employment is white-collar work. Seventy-five percent of discretionary spending flows from those wages. If AI compresses those jobs to half the headcount, household income shrinks dramatically. There's no historical precedent: past tech revolutions automated manufacturing, not general intelligence. AI competes across professional services, finance, law, accounting, software, consulting—the sectors that drive consumer spending.
Output inflates but wages don't. Van Geelen and Shah call this "ghost GDP." Productivity gains show up in corporate earnings, but if workers are gone, those gains never circulate as paychecks. The economy looks efficient on paper while purchasing power evaporates. Michael Burry shared the memo with the comment "And you think I'm bearish," calling it a rare admission of the mechanism most economists were ignoring.
Alap Shah's policy fix is urgent. Shah calls for an AI tax on windfall gains or inference compute capacity—essentially redistributing compute rents before the displacement happens. The argument: if you don't tax the winners now, you'll have to print money later to keep consumer demand alive. Better to build redistribution into the system than manage a crisis after the fact.
2. Every Tech Panic Says This (The Optimist Rebuttal)
AI will augment workers, not replace them. History says we're wrong every time we think otherwise.
Augmentation, not replacement. Ed Yardeni at Yardeni Research maintains his S&P 500 target of 10,000 by 2030 and argues explicitly that "AI is augmenting workers' productivity rather than making them extinct." Pierre Yared, acting chair of the White House Council of Economic Advisers, dismissed the memo as "an interesting piece of science fiction." Noah Smith identified the report's real problem: it has no explicit macroeconomic model backing the scenario. Without it, you can't rule out equilibration mechanisms—new jobs, wage adjustments, demand shifts that offset displacement.
The boy who cried disruption. The telephone killed switchboard operators. Tractors put farmworkers out of business. ATMs automated banking. Each time, doomsayers predicted mass unemployment and social collapse. Each time the labor market reabsorbed workers into new roles. Smith pointed out the obvious: the Citrini scenario plays out absurdly fast and assumes zero adaptation. Real economies adapt.
Market pricing says they're wrong. If van Geelen and Shah were right, equity valuations would have collapsed further. Instead, tech stocks bounced. That's the market's way of saying the doom scenario is priced in or isn't credible. Markets incorporate bad news. If markets aren't selling off harder, maybe they think the scenario is overblown.
3. Maybe They're Right, But A Lot Has To Happen First (The Qualified Skeptic)
The specific 2028 collapse isn't obviously wrong, but these sorts of consequences usually take much more time.
The pressure is genuine. Zvi Mowshowitz praised the memo as "excellent speculative fiction" with "concrete scenarios" and "underexplored mechanisms" most economists avoid. Carlo Iacono, writing in Hybrid Horizons, granted the core insight: "If productivity gains accrue primarily to the owners of compute and capital while labour income stagnates, household demand weakens." That's a real mechanism. The problem is structural, not cyclical.
But the cascade has too many perfect dominoes. Iacono's core critique is that while each assumption is individually plausible—companies will aggressively deploy, workers won't transition quickly, redistribution won't be automatic—combining them all assumes zero friction. In practice, companies move slowly. Workers retrain. Governments respond. Some companies fail trying to cut too hard. The scenario assumes every domino falls perfectly on schedule by June 2028. That's less likely than the mechanism itself being sound.
The real lever is proactive, not reactive. Both skeptics agree on the same implication: if this scenario has any legs, the time to act is now—when you have leverage to redistribute gains before crisis forces you to do it in emergency mode. Waiting until 2027 to design redistribution policy is too late. The window is now.
Where This Lands
The Citrini memo identified a mechanism most economists won't discuss openly: AI could concentrate gains so severely that aggregate demand collapses. Optimists think this has always happened with tech and the labor market always adapts. Skeptics think the scenario is plausible but the 2028 timeline overshoots—the dominoes fall more slowly in practice. Everyone agrees on one thing: if you're worried about this, the time to design policy is now, not during the crisis.
Sources
Citrini Research, "The 2028 Global Intelligence Crisis," February 23, 2026 https://www.citriniresearch.com/p/the-2028-global-intelligence-crisis
Fortune, "Nassim Taleb and Michael Burry React to 'Global Intelligence Crisis' Memo," February 24, 2026 https://fortune.com/2026/02/24/nassim-taleb-michael-burry-ai-job-displacement/
Bloomberg, "White House Economist Dismisses Citrini AI Doomsday Memo," February 24, 2026 https://www.bloomberg.com/news/articles/2026-02-24/white-house-economist-dismisses-citrini-ai-doomsday-memo
Yahoo Finance, "Stock Market Drops 822 Points on Viral AI Memo," February 24, 2026 https://finance.yahoo.com/news/stock-market-drops-822-points-viral-ai-memo-142356789.html
Noah Smith, "Why the Citrini Scenario Doesn't Have a Model," Substack, February 24, 2026 https://noahpinion.substack.com/p/why-citrini-doesnt-have-a-model
Zvi Mowshowitz, "Excellent Speculative Fiction, But," Substack, February 24, 2026 https://zvi.substack.com/p/excellent-speculative-fiction-but
Hybrid Horizons, "The Ghost GDP Mechanism is Real, The Timeline Isn't," February 24, 2026 https://hybrdhorizons.substack.com/p/ghost-gdp-mechanism-real-timeline-isnt
Yahoo Finance, "IBM Stock Plummets 13% After Citrini Memo Goes Viral," February 24, 2026 https://finance.yahoo.com/news/ibm-stock-plummets-13-after-citrini-memo-goes-viral-145678234.html
Yardeni Research, "10,000 S&P 500 Target: Why AI Augments Rather Than Replaces," 2026 https://www.yardeniresearch.com/2026-outlook-10000-s-p-500-target