Is AI in a bubble? The Dot Com boom, accelerated

It's still the usual suspects.

If there’s something a little bad in your life today that you remember being quite a lot better before, the reason is almost always the same: private equity.

Private equity is why rent is high, why dental care is bad, and why your company was just acquired three times.

It's also why AI seems to be accelerating at such a break-neck pace.

Shiny object syndrome, at scale

Private equity investors provide capital in exchange for stake in a company. But by becoming equity partners, private equity investors also gain some level of control over the organization. And they pursue profit at all costs—unlimited growth and unlimited expansion. Frequently, this leads to the Matroshka doll effect: a company acquiring a company that then acquires a company.

When private equity firms started buying up stakes in veterinary care clinics, for instance, veterinary costs skyrocketed. You were still getting the same service. The veterinarians were still getting the same pay. But private equity started taking a cut.

So, private equity firms don’t even look at a technology like AI and ask “What can it do?” They look at it and ask “How much money can I make?” And that has led to a stunning proliferation of AI companies across the board, regardless of how silly or how niche.

For a technology that actually does have some incredible use cases, it's actually tragic. Investment dollars can fuel innovation, but they can also set it back. Private equity firms make money fast and dirty. They don’t care whether an organization has longevity; they intend to get in and get out.

So, if the way we are building AI feels shaky and uncoordinated right now, that’s by design.

Fast to market, first to market

But that doesn’t mean that all AI companies are sloppy or uncoordinated. There are quite a few incredible companies out there doing wonderful things with AI. But not only are these companies outnumbered… they’re also a bit slower.

Being first to market isn’t always best long-term—it’s not always sustainable. Just as Google wasn’t the first search engine (and Apple wasn’t the first smartphone), being second, third, or even fourth to market can yield better results. The market is now prepared to adopt your technology, the technology has become more feasible, and you've spent the time you needed to develop a solid product.

And coming to market with a truly novel product takes time. If you’re in a space such as healthcare, education, or finance, there are additional artifacts of compliance you need to meet.

These factors coalesce into a very specific environment: an environment in which the fastest movers may not necessarily be the best examples of technology or application.

Do you remember the Dot Com boom?

If you do, your back probably hurts in the mornings.

From 1995 to 2000, the NASDAQ grew by 800%. By October 2002, it had fallen 78%, largely erasing its prior gains. Of course, raw percentages don't tell the full story; there were winners and there were losers.

But there were mostly losers.

The World Wide Web was the “new thing,” and investors were flush with cash. Startups materialized overnight with nothing more than an idea—a promise. These startups then burned through venture capital, largely producing nothing. Eventually, it all collapsed.

That’s not altogether different from what is happening to AI. It’s inarguable that the World Wide Web strongly changed the foundations of our society. It’s inarguable that tremendous value has been produced through the internet. But in its infancy, it strongly over-extended. Individual investors did everything they could to simply replicate the world in the web, even if the world wasn't ready for it.

Now, we have investors chasing the new big thing. They are willing to invest in practically anything that says “AI.” And more to the point, some investors don’t really understand AI; at least, they don’t understand it to the depth that they can really know the difference between a great technology and a not-so-great one. So, everyone is racing to incorporate AI into their products—even if it might not be a fantastic idea.

A house of cards, creaking

An AI boom actually has the potential to be even more disastrous than the Dot Com boom, if only because of the infrastructure. When the Dot Com boom occurred, the internet wasn't such a widespread part of our lives. The sphere of influence was limited, even if it felt at the time omnipresent.

Now, AI can take advantage of the internet to move across our entire globe with a swiftness. And we're already feeling the consequences of that. 70% of workers involved with AI say that AI is actually increasing their job responsibilities, not decreasing it; a consequence of demanding even more labor with every technological invention.

And we are radically adjusting our society quickly to centralize ourselves around the idea of AI. New laptops and smartphones are being pumped out already "AI-enabled." The tech industry has made the decision for us; we are only passengers.

A boom doesn't necessarily have to bust. There could always be, to steal a term, a "soft landing." But if there is a bust, it has the potential to take out a widespread number of industries if only because it is so deeply embedded.

And what remains

AI's many defenders will say that AI is a strong technology.

So was the web.

The web didn't have to be a bad technology to create a boom-and-bust cycle. It didn't have to fail to live up to its promises for there to be a bubble. We are here, decades later, and the web is now everything that the Dot Com boom really promised that it would be.

And that's an important thing to take away: an "AI bubble" does not inherently indicate that AI technology is not worth investment. It's not an attack on AI; there are other, better attacks that can be used against AI.

Rather, a boom-and-bust can occur simply due to investors who cannot adequately separate grifters from innovators.