Citrini’s Argument Is Most Vulnerable to Criticism on the Grounds That It Takes for Granted This Is America

Nvidia reported record earnings on Wednesday afternoon, beating expectations on all three metrics. Despite a quick pop early in the after-hours trading session to a high of roughly $203, about a 3.5% increase from close, the stock cooled off shortly thereafter and ended up settling nearly at the exact same market price it closed the trading day at, $195.92 a share.

Some in the financial media world are using this as a signal that the fears of an AI bubble are overblown. Those fears, while not linear, have steadily become increasingly common as the current bull market continues to produce increased returns. Alarm bells rang once again this weekend. Markets sold off steeply on Monday of this week, with many attributing much of the selling pressure to the narrative surrounding the Substack post “THE 2028 GLOBAL INTELLIGENCE CRISIS” published by Citrini Research on Sunday.

Like many people, I found the conclusions of the article both plausible and concerning for the fate of human society. Over the last few years, I have been a self-proclaimed AI skeptic for two reasons.

First, its use case. In periods of rapid technological advancement, there is almost always overpromising. Sales teams need to close deals, start-ups need to raise funds, and executives want to convince shareholders that the multiple expansion of their shares will be worth it in the long term. Go listen to a Tesla earnings call from 2024 and you’ll probably hear their supreme leader predicting that 50% of Americans will have their own personal humanoid robots by Q1 2025. (Kind of hyperbole from me.)

My second and main concern was the human element. If AI is as disruptive as some predict, is it something that will even benefit human life? Sam Altman made some alarming comments last week, giving rare insight into the way he views humans. More laborers than persons. If the wealthiest business owners can replace the labor force with a fully submissive, more efficient workforce that need not eat, drink, or sleep, is investing in AI a path we should even consider?

Yet, as I am quite contrarian in nature, the Citrini article made some claims that I feel need pushing back on. Oftentimes, when someone begins to predict a doomsday scenario such as this, it’s easy to connect the dots that are laid out in front of you without pushing back.

Many questions arose in my head as I made my way through the piece, but there are two I would like to explicitly push back on.

The report makes the argument that AI disruption in the white-collar economy will create a circular “AI Driven Economic Feedback Loop” with no natural brake. In theory, as AI capability continues to improve, companies will need fewer workers. As these companies lay workers off, consumer demand will decrease. When demand begins to decrease, the companies will, in return, invest more in AI to protect their margins. Thus, AI companies will have increased revenue to improve the capability of their models, and a never-ending cycle ensues.

This AI doom loop only works if the assumption is made that the wage-consumption paradox does not exist, subsequently making the claim that a firm can reduce labor costs to zero without undermining its own customer base.

The entire argument relies on the assumption that aggregate demand will not contract in an economy with a skyrocketing unemployment rate.

While this argument is rather abstract in nature, as the relationship between consumer demand and wages has been debated for centuries, Citrini’s thesis that the widespread introduction of AI agents to consumers will shrink corporate margins ignores real-world evidence.

The article argues that early on in the cycle, the consumer will experience a small win, as AI agents will be able to take over the tedious process of bargain hunting online.

“Over the past fifty years, the U.S. economy built a giant rent-extraction layer on top of human limitations: things take time, patience runs out, brand familiarity substitutes for diligence, and most people are willing to accept a bad price to avoid more clicks. Trillions of dollars of enterprise value depended on those constraints persisting.
It started out simple enough. Agents removed friction.
Subscriptions and memberships that passively renewed despite months of disuse. Introductory pricing that sneakily doubled after the trial period. Each one was rebranded as a hostage situation that agents could negotiate. The average customer lifetime value, the metric the entire subscription economy was built on, distinctly declined.
Consumer agents began to change how nearly all consumer transactions worked.
Humans don’t really have the time to price-match across five competing platforms before buying a box of protein bars. Machines do.”

This assumes that these AI agents will only be introduced on the side of the buyer, claiming that AI will be able to efficiently search online marketplaces in seconds, finding the cheapest prices, while those selling the goods hopelessly sit on the sidelines, forced to continuously undercut each other to remain competitive.

Almost in real time, as I was reading this, my mind shot to algorithmic pricing. American companies have already shown their willingness to quickly change the prices of the goods they are selling based on a variety of algorithmic factors such as weather, hour of the day, or surges in foot traffic. It is almost a given that sellers will implement these same agents to maximize their margins. This is without even considering the possibility of collusion throughout the market from said AI agents to artificially prop prices up.

As I began to connect these dots in my head, I began to realize that the potential doomsday scenario in front of us could be a completely different type of dystopian economy.

Over $12 million were wagered on the State of the Union address over the last several weeks. As Polymarket and Kalshi continue to expand in popularity in the United States, so does the volume of money wagered on their platforms. I’ll refrain from soapboxing about my qualms with these two platforms and lay out how I think an underlooked risk of AI disruption is turning the entire economy into both a casino and a two-way financial market.

If both the American buyer and seller have near-infinite access to real-time pricing tools, with autonomous agents constantly adjusting and reacting, it is not difficult to imagine the entire online shopping economy morphing into a live financial market. Recurring Amazon purchases begin to fluctuate with real volatility. The line between shopping and trading starts to blur.

Here enter the degenerates.

Once prices move continuously, it’s only a matter of time before the WallStreetBets bros try to trade them. Automated, real-time arbitrage of everyday goods. Buying a Netflix subscription at $12 a month only to flip it at $12.33. Instacart orders packaged and exchanged like mortgage-backed securities. Eventually, markets will develop on the volatility of the markets themselves.

That is my personal doomsday porn.

The article lays out a scenario I find very plausible. I do fear it is not being taken seriously enough by business leaders, civilians, and most importantly, politicians. Economic predictions are rarely accurate. Economic theory, in general, is almost never reflective of how humans actually respond. But we are now introducing an autonomous element that is not human. How will that move the market? If AI turns consumption into a continuously repriced market, we should not be surprised when society begins to treat it like one.