Markets, AI & The Great Dumbing

As AI takes hold, markets lose their clarity, leaving profound questions about risk and the true path ahead

With a NASDAQ and S&P 500 (narrowly driven by seven, mega-cap monopoly powers) enjoying a seemingly immortal ride North, all feels eerily normalized in the land of Wall Street Oz.

Surfing the S&P Wave

An entire generation of wide-eyed investors and a string of clueless policy makers have conditioned themselves (and others) to assume there is no dip that the Fed can’t save, valuation be damned.

In such a seductively sunny backdrop, retail investors are increasingly jettisoning risk management (and risk managers) to passively ride this market wave on ETF-indexed surfboards with very little fear of drowning.

Rocks Beneath the Water?

Meanwhile, a minority of market veterans (Grantham, Buffett, Dalio, etc.) bravely (but in vain?) continue to warn of historical market risk greater in scale than the Nikkei of ’89 or the DOW of ’29 as insiders (Bezos, Zuckerberg, etc) quietly dump billions in private shares in a topping market…

So, whose right or wrong in this bear-bull circus of hidden risks and open optimism?

Are the bears just crying perpetual wolf in a world where the centralization of “once-free markets” by central banks of endless liquidity has outlawed the very laws of stock market gravity?

Madness of Crowds?

For years, the risk-focused bears have been shaking their experienced fingers before what we (and Charles Mackay) have called “the madness of crowds.”

And for the most part, the “madness” has prevailed in one profitable dip-buy after the next—from a global shutdown and historical bank failures to a broken carry trade or an embarrassing liberation day.

In short, nothing seems to shake this S&P, which long ago divorced itself from the mean-reversion warnings of Bob Farrell, the valuation signals of Ben Graham or even the stubborn math of red financial statements and broken balance sheets.

The net result is the most optimistic yet paradoxical (and doomed) market I have ever witnessed.

Margin debt, now greater than anything seen in the pre-GFC or dot.com implosions, has surpassed $1T, as AI-consuming stocks like NVDA, with a market cap of $4T, equal more than 13% of US GDP.

In such current euphoria, boring things like PE ratios, history or even common sense have left the room, and words like “bubble” are literally replaced with the dumbest words of all time, namely: “This time is different.”

Hmmm…

Blindness from Above

Standing upon frothy tops, perhaps it becomes harder to see the lessons of history or the valleys of busted cycles.

Perhaps investors losing oxygen at nose-bleed market levels forget that following the 1929 crisis, markets never sold at 21X earnings again until December of 1997.

Now, such PEs are considered almost “tame.”

Perhaps they also forget that the crashed markets of 1929 did not recover until 1958, just as the crash in 72 only recovered in the mid-90’s or that the great tech bubble of 2000 did not recover its losses until 2011.

And then of course there is the Nikkei of 89, which took over 30 years to regain its lost levels.

But why worry? In the moral-hazard-rich “new normal” of central bank “accommodation” (i.e., currency destruction), every dip is an easy buy, right?

In other words, nothing can stop this new direction in centralized and immortal markets, right?

Wrong.

Will the Fed Save Us?

That classic oxymoron, the Federal Reserve (which is neither “Federal” nor a “reserve”), is no miracle cure.

Regardless of who runs it today or tomorrow, and regardless of its balance sheet and rate tricks, eventually the mechanizations of man can and will bend before the natural laws of supply and demand, debt and currency debasement, and alas, human hubris before karmic implosion.

As for the long list of needles pointing at the US debt and market bubble, any number of foreseeable white swans (spiking rates, a USD out of Fed control, a deep recession) or unforeseeable black swans (wars, assassinations, social unrest, etc.) could bring the madness of these markets and crowds to a sobering and historical uh-oh moment.

The Ignored Needle: AI

But there’s another needle pointing at these mad crowds and madly inflated (and historically unprecedented) markets, which few are willing to see, and it’s currently winning hearts, minds and income statements at equally maddening (and misleading) levels, namely: AI.

The Dumbness of Crowds

Like so many current and past memes of market salvation (from electricity and railroads to dotcoms and mortgage-backed securities), the now omnipresent “AI” wave will eventually (and empirically) drown a large swath of trend-trusters and bubble victims who refuse to see the forest for the trees.

In line with the other oxymorons of late (from the not-so-patriotic Patriot Act to the not-so-genius GENIUS Act), Artificial Intelligence, by its very title, is a comical signal of ignored irony.

To remind: “Artificial” is the antithesis of genuine, and any intelligence that is artificial is inherently the very opposite of intelligence.

Such simplicity, of course, will not dissuade those mega-computing “moderns” (from Sam Altman to Eric Schmidt) who are currently making fortunes on microchips and data-synthesizing “miracle tech” under the banner of “Our Human Future.”

Again with the ironies…

In fact, a deeper dive into the numerous ripple effects of AI suggests there is nothing very “human” behind AI nor anything that “humane” ahead of it.

In other words: When it comes to AI, be careful what you ask for…

First: The Illusion

For anyone glued to their ChatGPT, there’s no denying that AI processes data (both good and bad) incredibly fast. But faster does not mean wiser.

As Antoine de Saint-Exupéry wrote in the 1940s, “Today we can manufacture 10,000 pianos a day, but not any pianists worthy enough to play them.”

In other words, technology often outpaces wisdom, which cannot be augmented by robots or implanted chips.

Wisdom requires time, effort and exposure, not an app.

And once wisdom leaves the room (as we see in nearly every political headline), well… we are all in major trouble.

In a society where actual reading, research and critical thinking (like piano-playing) has already been critically eroded by Google searches and neck-to-iPhone addictions, just imagine the “great dumbing” to come for a newer, faster generation who gets their thinking from ChatGPT rather than, well, actual thinking, research and debate?

Very soon, AI-driven protocols, policies and decisions won’t have to explain themselves at all. Pivotal decisions and “debates” will simply be resolved with: “ChatGPT said so.”

Oh, the horror…the horror…

And no matter how seductive, fast and seemingly dispositive AI-generated conclusions are processed, their outputs are only as good as their inputs.

Diagnostic Robotics, for example, used AI for COVID prognostics, which were impressively fast, but sadly, embarrassingly wrong.

Even DARPA’s Director of AI for detecting IEDs in combat zones confessed that the results were no better than a “coin toss.”

Next: The New, Exciting, Sexy and Life-Altering Bubble

But there’s no denying that AI is new, exciting, sexy and life-altering.

As such, demand for anything touching the AI space (from chips and software to electronic baristas, robotic romance partners and high-school essay cheaters) is skyrocketing.

The $4T market cap at the center of this entirely familiar pattern is, of course, NVDA, whose micro-chip sales are critical to the evolution of this space. Jeremy Grantham describes NVDA as “sellers of shovels in an AI gold rush.”

That is why the other mega-cap tech companies, which keep the S&P precariously afloat alongside NVDA, are buying its microchips to the tune of $200B per year. Once the AI star falls in price, so too will the companies holding the S&P together.

In this race to stay current and ahead of the AI madness, valuation can get cloudy, as names like DeepSeek can arguably do for $6M that which “social visionaries” like Sam Altman will do for $500B.

But hey, why worry about valuation when tomorrow’s prices always seem higher than yesterday’s?

But therein lies the risk.

As in every other prior example of new, exciting, sexy and life-altering bubbles, such as the railroads of the late 19^th century or the evolving electricity, automobiles and even internet and dot.com fortunes of the 20^th century, there is no denying their extreme impact on our lives.

The AI of the 21^st century is undeniably no less of a game-changer than these prior game-changers.

But from a market veteran’s lens, there’s equally no denying that each of these prior game-changing “technologies,” though massively important, were eventually massively over-bought and then, you guessed it: Massively over-sold.

Finally: The Implosion

Over-sold, of course, is just a nicer way of saying, “the market tanked.”

But the real risk behind the current AI hype, bubble and madness (of which the seven stocks leading the S&P are dangerously “all-in”), is not just the pattern recognition of a technology bubble in which fortunes are made before markets are gutted.

In fact, the AI revolution is far more threatening at a broader economic level than just another boom-bust tech cycle.

In the brave new dystopia in which otherwise “speedy and efficient” technology is taking code-writers to Silicon Valley wealth, this unwise yet smart segment of the Orwellian world in which we now live brags about a Starbucks with 100 robots and only 2 employees as “progress.”

The word I’d use is a bit less sexy but a lot more honest: “Deflationary.”

I’ve sat at VC seminars in the US and Europe in which unemployment levels driven by AI alone (from white-collar to blue-collar, Starbucks to UBS) are projected to reach at least 10% within two years.

Here, it’s worth reminding that 5%-6% unemployment rates are traditional tipping points whereby passive investors get clobbered as their index-friendly ETFs turn from friendly to deadly.

10% unemployment would be a true needle to pop the current (mad) market and send 401Ks, as well as the AI-deep “Mag-7” fatally southbound.

And yes, we saw 10% unemployment during the GFC of 2008, and even 15% unemployment during the COVID crisis, which was, of course, supported by trillions in fiscal stimulus, which the US can’t afford today without effectively knee-capping the USD (and sending the gold price further moon-bound).

In other words, AI is more than just another tech bubble cycle; it’s a job-killing technology for which an already debt-driven (rather than GDP-driven) US economy (and S&P) cannot and will not be able to “land softly.”

In summary, if you still think AI is what will save this market and economy, you may wish to think again – or at least ask ChatGPT what it thinks for you…

---