The friction economy is over

On Monday, the Dow Jones Industrial Average fell 822 points. The Wall Street Journal tied the selloff in part to a single piece of writing that had gone viral over the weekend: a fictional macro memo from CitriniResearch, dated June 30, 2028, reconstructing the collapse of what they called the Global Intelligence Crisis.

It is a scenario, not a prediction. The authors say so plainly. But the market did not treat it that way, and neither should you.

The piece deserves to be read in full. It is one of the most grounded, financially specific, and unsettling analyses of what AI actually does to an economy. Not in theory. In plumbing. In the real mechanics of how money moves through a system built entirely on the assumption that human intelligence is scarce.

That assumption is now in question. And if you are a CEO, the only question worth asking this week is the one buried near the end of that memo: how much of your business model is built on friction, intermediation, or human cognitive tasks that are now exposed?

The moat that wasn't

For decades, the most durable competitive advantages in business were not the best products. They were the best friction.

Insurance companies earned 15 to 20 percent of their premiums from policyholder inertia. People who simply did not get around to re-shopping their coverage every year. Travel booking platforms built billion-dollar businesses on the idea that assembling a complete itinerary was so tedious that consumers would pay someone else to approximate it. DoorDash's real moat was not logistics. It was because you were hungry and tired, and it was already on your home screen.

The CitriniResearch piece calls this "habitual intermediation," and it describes its collapse with precision. AI agents do not have a home screen. They do not get tired. They do not default to the familiar option. They check every platform, every price, every alternative, every time, in milliseconds. The entire consumer subscription economy, built on the behavioral science of inertia, is now negotiating with a counterparty that has no inertia whatsoever.

This is not a future risk. It is happening now. The piece notes that, in their scenario, by early 2027, the average American was consuming 400,000 tokens per day from AI systems. Those systems were renegotiating subscriptions, routing around interchange fees, re-shopping insurance, and dismantling every business model whose value proposition was ultimately "navigating complexity you find tedious." The agents found nothing tedious.

The ServiceNow lesson every CEO should internalize

The most instructive story in the memo is about ServiceNow.

ServiceNow sold workflow automation. It was disrupted by better workflow automation. Its response was to cut headcount and use the savings to invest in the very technology disrupting it. The company that sold seats lost revenue when its customers cut 15 percent of their workforce and cancelled 15 percent of their licenses. The AI-driven headcount reductions that boosted margins for Fortune 500 clients were mechanically destroying ServiceNow's own revenue base.

What makes this story important is not the outcome. It is the logic. Every company's individual response was rational. Cut costs, deploy AI, maintain output. The collective result was catastrophic. Every dollar saved on headcount flowed into AI capability that made the next round of cuts possible.

The piece describes what happened to incumbents across all sectors: they did not resist because they could not afford to. Boards demanded answers. Stocks were down 40 to 60 percent. The historical disruption model said incumbents die slowly by resisting change. What happened instead was that incumbents accelerated their own disruption by adopting the technology aggressively, thereby validating and funding the very loop that was consuming them.

Any CEO running a company today needs to ask whether the revenue model survives the aggressive AI adoption of their own customers. Not whether AI will disrupt the product. Whether the efficiency gains clients extract from AI will mechanically cancel the seats, licenses, or contracts they currently hold.

The cognitive task inventory

The piece draws a clear line. Blue-collar employment remained relatively stable through the scenario's initial disruption. White-collar employment collapsed. The jobs that disappeared were the ones that wrote memos, approved budgets, routed decisions, managed workflows, and lubricated the middle layers of the economy.

This is not a statement about the relative worth of those workers. It is a statement about what AI can do right now, and what it cannot. Physical presence, dexterity, and in-person judgment remain genuinely hard problems. Cognitive intermediation, the process of taking information, applying learned pattern recognition, and producing a recommendation or deliverable, is increasingly being solved.

Every CEO should conduct a cognitive task inventory of their organization. Ask two questions. First, what percentage of our labor costs is paying for human cognitive tasks that an AI agent could perform at a fraction of the price? Second, what percentage of our revenue is being paid to us by customers who are asking the same question about us?

The first question reveals where the cost structure is exposed. The second reveals where the revenue is exposed. If the honest answer to either is "a large percentage," then the CitriniResearch scenario is not fiction. It is a roadmap.

The intelligence premium is unwinding

The most philosophically significant passage in the memo is near the end. The authors write that throughout modern economic history, human intelligence has been the scarce input. Capital was abundant or at least replicable. Natural resources were finite but substitutable. Technology improved slowly enough that humans could adapt.

Intelligence was different. The ability to analyze, decide, create, persuade, and coordinate could not be replicated at scale. Every institution built over the last century, from labor markets to mortgage underwriting to the tax code, was designed for a world where that assumption held.

That assumption is now unwinding.

This is not doom. The authors are careful about that. The economy will find a new equilibrium. It always does. But the path to that equilibrium will be disorderly, and the businesses that fail to see it coming will not fail slowly, as Kodak or Blockbuster did. They will fail the way ServiceNow did in this scenario: rationally, quickly, and while trying their hardest to survive.

The canary is still alive. The S&P 500 is still near all-time highs relative to any historical baseline. The loop has not started. But a piece of financial scenario writing going viral enough to move the Dow 822 points is itself a data point. The market is signaling that these questions are no longer academic.

What personal leadership demands right now

At CEO.com, the core belief around personal leadership is straightforward: before you can lead an organization, you have to be able to lead yourself. That requires being honest with yourself. To be able to see clearly.

Seeing clearly right now means resisting two temptations.

The first temptation is dismissal. Calling this speculation, noting that scenarios are not predictions, pointing out that the doom-and-gloom crowd has been wrong before. Some of that is true. But the underlying mechanism is not speculative. AI agents that eliminate friction are not a 2028 hypothesis. They are a 2025 product. The question is not whether they exist. It is how fast they scale.

The second temptation is paralysis. The piece is written with enough detail and conviction that it can feel deterministic. It is not. The authors close by noting that there is still time to be proactive. As investors, as executives, as a society.

What personal leadership demands is a clear-eyed audit. Not of your AI strategy in the abstract, but of the specific assumptions underlying your business model. Where are you charging a premium for something that is no longer hard for humans but is still hard for machines? Where are your customers paying you to navigate complexity that agents will soon handle for free? Where is your moat made of friction?

Answer those questions honestly, and the next step becomes clear.

The piece ends with a line worth sitting with: "Getting there is one of the few tasks left that only humans can do. We need to do it correctly."

That is still true. For now.