Before last week the name Alap Shah didn’t ring a bell for many people. The 45-year-old financial analyst and tech entrepreneur had spent the past two decades working in relative obscurity. Then last weekend he coauthored a blog with the research firm Citrini titled “The 2028 Global Intelligence Crisis.” It was a “thought exercise” about the impacts of artificial intelligence, and it predicted that in June of that year, AI would jack up unemployment past 10 percent and force the Dow down, down, down. Writing in a confident, Nostradamic tone—as if auditioning for starring roles in the next Michael Lewis book—the authors painted a picture of a flywheel in reverse: AI agents take jobs from workers, people spend less, and struggling corporations conduct layoffs on top of layoffs.
There wasn’t much in it that hadn’t been previously heard, or speculated about. Tech leaders like Anthropic CEO Dario Amodei have already estimated that half the entry level white collar jobs will soon be gone, and earlier this year, Anthropic’s release of new agentic tools spurred a Wall Street selloff. Nonetheless the report hit with the force of the blizzard blowing through lower Manhattan. When the closing chimes sounded on the New York Stock Exchange, the Dow was down 800 points. The name Alap Shah was now ringing bells.
The achievement is less impressive than it seems. Wall Street, like the rest of us, is in a persistent state of anxiety about AI, and it doesn’t take much to trigger a mini-panic. Financial markets don’t necessarily map to reality, but the jitters reflect a wider disquiet. The AI future is in a William Gibson zone—it’s here, but unevenly distributed—and the news from those already living in the agent-packed, AI code-writing universe is both exciting and unsettling. Emphasis on unsettling.
No one—no one!—knows exactly how AI will impact the economy, but clearly it will be significant. Right now stocks are soaring, so it seems to make sense to keep the party going. But then along comes the latest doom manifesto, or a paper indicating that a traditional business sector might be threatened by AI, and suddenly money managers are reminded that the biggest issue of our time is totally unresolved. Case in point: earlier this month, a tiny company (valuation under $6 million) that had previously sold karaoke machines pivoted to AI-powered shipping logistics and put out a report saying that it had discovered some efficiencies in loading semi-trucks. That was enough to erase billions of dollars from the share prices of several major logistics companies, none of which had karaoke experience.
After it did its job on Wall Street, the Citrini report came under considerable fire. Critics climbed over each other to proclaim its flimsiness. For one thing, they pointed out, AI has had very little discernable impact on the economy so far. Others cited the long history of resilience after technological upheavals. A mocking response by the respected trading firm Citadel Securities read, “For AI to produce a sustained negative demand shock, the economy must see a material acceleration in adoption, experience near-total labor substitution, no fiscal response, negligible investment absorption, and unconstrained scaling of compute.”
The most withering critiques disputed the report’s contention that much of the economy involves non-productive “rent-seeking” by middlemen and market makers, taking advantage of the laziness of the general population. When everyone has a few dozen AI agents working on their behalf, writes Shah, consumers will be able to effortlessly find the best goods for the best prices. Apps will be rendered unnecessary—just type what you want into the LLM and an army of agents will do everything for you. The “poster child” for this phenomenon, Shah says, is DoorDash. Instead of being limited to the restaurants on the app, consumers will send out AI agents to find their ideal meal options, contracting directly with restaurants and delivery people—no apps needed. Zero friction! The DoorDashes of the world are avocado toast!
