An Evercore ISI economist criticizes Citrini's AI report, calling its assumptions “extreme and improbable”, but says it's a thought-provoking exercise
Most of the sell-side has remained hilariously silent at a mere Substacker seemingly shaking markets, even though the anguish and frustration is almost palpable.
Financial TimesRobin Wigglesworth
Context & Ripple Effects
Citrini Research’s scenario framed an AI-led shock through white-collar layoffs and weakening consumer demand. Evercore ISI’s response puts that scenario under institutional scrutiny without dismissing its usefulness as a stress test.
Citrini’s report becomes a contested input to AI-market debate rather than an unchallenged narrative, with Evercore ISI explicitly questioning the plausibility of its assumptions.
Investors and analysts assessing AI exposure must separate the report’s provocative end-state from the assumptions required to reach it.
Second-order effects
The exchange directs attention toward whether AI can be deployed economically at scale; Citadel’s critique reinforces that compute requirements and labor-cost comparisons are central constraints on the scenario.
Sell-side research faces pressure to engage more directly with fast-moving independent AI theses when they influence market discussion.
Third-order effects
If such debates persist, AI valuation narratives may be judged less on headline productivity claims and more on explicit adoption, compute-cost, and demand-transmission assumptions.
The broader risk framework for AI shifts toward scenario analysis: severe labor-market outcomes can inform stress tests even when forecasters dispute their likelihood.
The trend: AI-market narratives are moving from generalized enthusiasm or alarm toward scrutiny of the economic constraints that govern real-world deployment.
An instantly disqualifying line in the Citrini paper: “In every way AI was exceeding expectations, and the market was AI. The only problem...the economy was not.” Sorry, this in implausible.
The 2028 Global Intelligence Crisis That Wasn't A Macro Memo from the Actual June 2028, Not the Fanfic Version The unemployment rate printed 3.8% this morning, roughly where it's been all year. The market yawned. The S&P 500 is at 7,400, which is somehow both a record high and a
Whenever this cycle ends, the Citrini “report” will be an extremely funny relic of the mania like a Bored Ape or Peloton's stock chart. The DoorDash analysis is a third grader's understanding of marketplaces.
Gratifying rebuttal to the Citrini article. The bit about real estate agents is particularly good. The internet should have killed that business. Not clear why AI agents will get it done if Zillow couldn't. https://stratechery.com/...
For the past few months, I have been writing a long-form podcast articulating an AI Bull Thesis for the digital economy. I call it The Prosperous Society. In light of the Citrini report, I decided to chop it up and release it episodically. Episode 1 is live. [image]
The Citrini essay includes a paragraph about an industry I know something about. Redfin tried to do this exact business model (bare-bones buy-side agent who rebates 2/3 of the 3% fee) and consumers hated it. I see no reason to think an AI version will work better. [image]
Reading the Citrini report felt like being forced to listen to 2 mid-IQ bros, stoned out of their respective minds, debate something without respect for complex systems while the captain of the football team, who just suffered a severe concussion, tells them how smart they are.
I really want people to see the story above the story here, which is that whether you're reading Citrini, or listening to Jamie Dimon at a cocktial party, the conversation about AI is a marketplace of competing science fiction narratives. That's not to say I think the
>longs assmad Citrini put out a piece that didn't have any particular relevance to their holdings but caused them to lose a shitload of money anyways [image]
Here is what I wrote to our team re Citrini: I just read the Citrini article. It's actually really good (so much better than the Matt Shumer post). However, the market has no business reacting the way it did. It's a total hypothetical disguised as an absolute because of how it's
If you're a bull, the worst thing you can say is “Citrini's scenario happens too fast” If we don't have enough compute and scaling laws happen over a longer time scale, this means that AXP will get its day of reckoning in 2032, not 2028! A decade of uncertainty! Ouch!
The meta-thinking proposed by the Citrini piece is correct: Our economy has been making huge leveraged bets on assumptions that will fail, and bad things will happen as a result. See also: Commercial real estate vs COVID.
“Any software engineer can now produce much more value than before” “No one will want to hire software engineers” Both of these cannot be true at the same time.
Yep. I have a realtor friend, and his huge advantage is that he knows literally everyone. Contractors, home inspectors, government officials, lenders, other agents. There are no surprises for him—anything that pops up, he's seen before. No AI agent is going to have the personal
The Kobeissi Letter just published a sharp rebuttal to Citrini's “2028 Global Intelligence Crisis” thesis. Save this breakdown - you'll want to revisit it when the next wave of panic hits. Kobeissi's core argument = the bearish AI case is “too obvious” and probably wrong. [image]