When Fan Fiction Hits the S&P
Why AI narratives are moving markets more than AI products
Two essays rattled financial markets this month. One, from Citrini Research, was a fictional macro memo from June 2028 describing an AI-driven economic crisis with cascading white-collar layoffs, a private credit implosion, and cracks in the $13 trillion mortgage market. The other, from AI startup founder Matt Shumer, was a personal account of how the latest AI models had rendered him unnecessary for the technical work of his own job and urged readers to prepare now for the death of white-collar work.
Both spread like wildfire across social media and generated a pile of breathless commentary. But they’re very different pieces of writing. The Citrini essay is genuinely captivating: a multi-dimensional speculative narrative that weaves together labor economics, private credit mechanics, mortgage market dynamics, and geopolitical consequences into a scenario that’s not totally implausible. The Schumer piece reads, ironically, as if it were substantially produced by the very tools it’s evangelizing.
Why did two posts of wildly different quality from people most investors had never heard of cause such a panic?
Narrative Consumers vs. Tool Users
The investors, analysts, portfolio managers, and finance influencers who consume pieces like these are sophisticated evaluators of financial logic. They can spot a flawed model from across a conference room. What they can’t do is evaluate whether the tech assumptions underneath those arguments are realistic on the stated timelines. That requires something financial analysis can’t provide: daily, hands-on experience with the actual products.
When Citrini posits that coding agents will let companies replicate mid-market SaaS products in weeks, or that consumer AI agents will route around credit card interchange fees, the finance audience evaluates the logic downstream of those assumptions. And the downstream logic is good! If those things happen, the cascading effects Citrini describes are plausible.
But someone who uses Claude Code every day to ship real software knows something the scenario takes for granted: the gap between “impressive demo” and “reliable production deployment at scale” is enormous, unpredictable, and not closing on any neat timeline.
Shumer’s piece reveals the same asymmetry but more crudely. His account of AI coding capability is credible: the latest tools really are remarkable for building software. But then he projects that experience uniformly across law, medicine, finance, accounting, and every other domain of knowledge. Anyone actually working in those fields would push back on that generalization, but the audience has no basis to.
AI coding tools are further along than AI tools in virtually every other profession, for structural reasons Shumer himself half-acknowledges: the labs built coding capability first because it helps them build more AI. His experience is real. His blanket application of it to everyone else’s job is not. (At least not yet.)
That this extrapolation was wrapped in AI-generated prose and still landed with the force of prophecy tells you everything about the state of the audience.
The Tomorrowland Effect
Walt Disney’s original Tomorrowland was compelling because it was a narrative about the future first and an engineering project second.1 Walt worked with scientists and engineers, but what made it resonate was the storytelling, making you feel something about what was coming, not just showing you a spec sheet.
The Citrini piece is a high-quality markets version of Tomorrowland: speculative fiction with financial fluency, built by people familiar with constructing market narratives. The Shumer piece is the gift shop version: mass-produced, shiny enough to catch the eye, but not something you’d put on a shelf.
Yet both move the same audience to the same action, and that’s actually the more concerning case. If only the Citrini piece had moved markets, you could chalk it up to the quality of the analysis. But when a clearly-AI-generated blog post produces a comparable reaction, you’re seeing something structural: the audience is grading the narrative, not the premise. They lack the messy, unglamorous, deeply unsexy middle layer of product experience that separates “technically possible” from “happening at scale next quarter.” Without it, anything well packaged enough lands as insight, and even things that aren’t particularly well packaged land too.
We saw this dynamic play out in miniature a few weeks ago. The market reaction to Anthropic’s Claude Cowork legal skills was notable. Legal services stocks dropped double digits when traders noticed the feature. But the plugin had been out since the previous Friday. Anyone who had actually installed it over the weekend would have had a different read. The legal skills are decent starting points for rolling your own workflows, not finished replacements for professional expertise. Enabling the legal skills does not conjure a Supreme Court justice into your laptop.
But if you haven’t opened Cowork, all you have is the press coverage, the feature announcement, and someone’s breathless Twitter thread. Those read very differently from the product experience.
Belief as Economic Force
Friend of the newsletter Kyla Scanlon coined the term “vibecession” to describe what happens when economic sentiment detaches from economic data.2 What we’re watching in AI is a sector-specific version of the same phenomenon: the market’s feelings about AI capability have detached from the product reality, and the feelings are moving real money.
A commenter on the Citrini piece, identifying themselves as working in a Mag7 AI division, made this concrete. It doesn’t actually matter whether a company can replicate a SaaS product with AI, they argued. What matters is whether they can credibly threaten to do so in a procurement negotiation.
The belief in AI’s capabilities, even one untethered from current reality, is already having real economic effects. If a procurement manager walks into a renewal conversation and says, “We’re exploring having AI replace your product entirely,” the SaaS vendor has to respond to that threat, whether it’s technically feasible or not. The bluff works because the vendor can’t be sure it’s a bluff. Margins compress, deals get restructured, real money moves on the strength of a narrative about capability, not capability itself.
This is where the calibration gap becomes self-reinforcing. Pieces like Citrini’s and Shumer’s feed the narrative. The narrative gives procurement managers ammunition. The ammunition produces real business outcomes. The outcomes appear to validate the narrative. It’s a strange loop where market fiction generates market fact.
Where This Puts Communicators
The calibration gap hits differently depending on where you sit as a comms person.
If you’re in comms at an AI company, this gap is awesome right now. The market is imbuing your products with god-like capabilities they don’t quite have yet, which drives investment, talent pipelines, enterprise deals, and the kind of procurement leverage that makes your sales team’s job easier. The rational short-term move is to say just enough not be technically lying and let the narrative do the heavy lifting.
But we’ve all seen this movie before. Every hype cycle has a correction, and the violence of that correction is proportional to the size of the gap between narrative and experience. The companies that survived those resets were the ones that had built enough real product credibility that the narrative coming back to earth didn’t take them out, too. The calibration gap is a tailwind until it’s a headwind. The comms teams at AI companies who are thinking about this now, while building genuine product credibility alongside the hype, getting real user stories into the market, and being specific where the zeitgeist is vague, are the ones who’ll have something to stand on when the music changes tempo.
If you’re in comms at literally anywhere else, be it a SaaS company, enterprise vendor, or startup that isn’t an AI company but lives in the blast radius of AI hype, the gap is a genuine strategic problem. Your customers are reading Citrini and Shumer and perhaps walking into renewal conversations with leverage built on narratives, not product evaluations. Your stock might be moving on someone else’s speculative fiction.
The opportunity here is real, though. There’s a premium on grounded perspective that’s genuinely scarce right now. Communicators who can translate between product reality and market narrative—who use these tools daily and can speak with specificity about what they do and don’t do—have a form of credibility most people in the conversation lack. Not “AI is going to change everything” hand-waving or “AI is overhyped” contrarianism. The specific, boring, valuable knowledge of what the tools actually do today, how fast that’s changing, and where the gaps remain between demo and deployment.
That gap won’t last forever. Eventually, enough decision-makers will use these products daily that the audience will develop its own immunity to narratives untethered from reality.
But right now, a work of well-crafted speculative fiction and an AI-generated blog post land with comparable force, and that tells you everything you need to know about how much room there is for someone who’s actually used the thing everyone else is just reading about.
Go watch this video from the wonderful YouTube channel Defunctland for a fascinating deep dive.
If I could dictate national high school curricula, every kid would read Kyla’s book “In This Economy?” It’s the most accessible and digestible publication on markets and economics I’ve ever read.