Who Says Another Google is Coming?

Why has AI hype been so overwhelming? I’ve argued that, at the bottom of this strange moment, there’s a broad desire to be freed from the mundane. AI sounds like something that might relieve us from the burden of living in boring reality, and that’s an attractive notion even for people living generally contended lives. The AI utopia and AI doomer scenarios are more alike than different in that they seem to offer the possibility of that big break, the fissure with the past that lets you (say) finally tell your boss to fuck off. The promise of The Jetsons future, just a little bit delayed, the palpable need for something different, powers AI dreams. Another big driver of AI hype, though, is the recent investment environment.

Fed funds interest rates, via

You will have heard this basic story before. Following the financial crisis of 2008 and 2009, inadequate government response led to a decade of anemic labor markets, acting as a serious drag on the income of ordinary Americans. (Thanks, Obama.) To try and address this slack market and make up for insufficient demand, the Federal Reserve system kept interest rates remarkably low, resulting in cheap borrowing costs meant to spur economic activity. That is to say, the inadequacy of the fiscal response led to an unusually aggressive monetary response. The low-interest rate days ended with post-Covid inflation and the Fed scrambling to tamp it down; as a man paying a 6.25% fixed-rate mortgage, I am very well aware of the effects of this new interest rate environment. (Over a full 30-year term, without refinancing, I would pay more than twice the principal over the life of the loan.) But for about a dozen years, we had really crazy low interest rates in historical terms. This had consequences.

For one, money was cheap, which is a phrase that always makes me grin a little for some reason. With low rates and quantitative easing (the Fed buying government securities to increase the supply of money), borrowing became incredibly inexpensive for firms, which gave them broad latitude to spend more than they earned and incentivized them to invest and grow knowing that their debt servicing obligations would be easy to manage. For another, a lot of traditional investment vehicles became pointless. You could go and buy a CD, but even with the consistently-low inflation of the 2010s, your return in real dollars would be pathetic. Most savings accounts were barely doing more for your money that putting it under your mattress. Both of these things contributed to a ton of speculative investment activity, including the crypto craze. It became common for people to say that a lot of investors were just looking for a place to put their money. Lots of loose cash, low interest rates, an eager investor class - all of it contributed to the conditions you can see in this graph. Everyone was trying to throw their money into startups.

The trouble has been that the investors in this wave had extremely high standards. A phrase you hear sometimes is “hockey-stick growth,” meaning a graph of stock price that looks like a hockey stick lying on its edge. They want exponential gains, not slow and steady increases. You could read a lot about venture capital investors who were taking a buckshot approach, throwing money around knowing that almost everything they invested in would fail. They didn’t care if nineteen of the startups they funded lost them money, though, so long as one of them returned at such an absurd rate that it covered their losses. They were hoping, in other words, that they could get in early on the next Google or Facebook. (Well, Alphabet and Meta, now.) Those companies are the paradigmatic Silicon Valley success stories; they’re the storybook startups that became billion-dollar propositions in a handful of years. Other giant tech firms like Microsoft and Apple took decades to achieve the valuations they enjoy now, and in 2024 America, nobody wants to wait. And with debt so cheap for so long, both investors and founders were incentivized to let the good times roll.

The trouble is that there hasn’t been a new Google or Facebook since Google and Facebook. There’s been plenty of profitable new companies, mind you. Uber has gotten rolled up into many people’s conceptions of huge successful Silicon Valley firms, and setting aside any questions about their business model, it’s true that the company has grown quite large. But large is relative; Uber’s market cap is less than 10% of Google’s. And besides, it’s too late to get in early on Uber. The company will be old enough to drive next year. (See what I did there?) There simply hasn’t been any success stories on the scale of Google and Facebook since. A lot of profitable new firms, meanwhile, aren’t sexy in the way tech firms are perceived to be sexy. Dirt-cheap Chinese clothing company Shein was a darling for awhile, but it’s a clothing company. They sell people physical goods. Not sexy! Stripe, Zoom, Slack - potentially sexy, to a certain kind of person, but nobody’s 100x’d their money. Now that the party’s over and the Fed’s free beer era is over, the macro conditions are putting more and more of a squeeze on potential startups - it’s much more expensive to borrow and there’s less incentive for investors to chase moonshots. So your idea has to look really impressive.

It’s worth saying that the market has been doing very well. The market isn’t the problem. The problem is that tons of people are chasing returns based on a couple of companies that emerged in truly sui generis situations. I mean, the internet isn’t new to most people the way it was in the late 1990s and early 2000s. What if we’ve simply picked all the low-hanging fruit that’s out there? What if all of the space for innovation in tech is already being filled by these preexisting giant firms, which have tons of capital, can hire the most qualified people on earth, and have a lot of other incumbent advantages? Who says there’s a trillion-dollar idea left to fund? There’ll be plenty of profits and lots of opportunity moving forward, but the dream of a new Google or a Facebook (and, of course, of recognizing one when you see it) has no particular reason to be fulfilled. I mean, look, Aramco and Exxon and Chevron etc. remain huge firms despite being the opposite of sexy, right? The oil and gas business is a really good one to be in. Big profits, huge market caps. But nobody is basing their whole investment strategy on the idea that there absolutely has to be some hot new oil company startup that definitely will emerge. Why would anyone think that? And the same question applies to all of the people hunting for those infinity returns. Tech is a very mature field now, and despite startup mythology, what tech startups are trying to do now is often capital-intensive. The free money spigot has been turned off. There’s no law of nature that says another Google or Facebook is going to appear out of the mist and offer you an opportunity for pre-IPO investment.

And yet a quick perusal of the “hockey stick growth” search term on Twitter reveals how many people are chasing this dream.

Which is where AI enters the picture. What current “AI” systems (LLMs/neural nets) can do is underwhelming relative to the promises, in my opinion, certainly on the consumer side. Even the most promising applications like protein folding are still firmly in the “I’m gonna” world, not the “I am” world. Some of the stuff that might be accomplished is cool, but a trillion dollars worth of investment cool? No. But the tail has to wag the dog; investors want to invest, companies want to grow, and our culture has a burning desire to believe that Silicon Valley is an endless firehose of progress. And so you’ve got this irrational attachment to the idea that the big idea is floating around out there, waiting to make the lucky few fabulously wealthy. It’s just like I was saying about professional gambling and day trading - there’s this growing mob of people who explicitly reject any talk of slow and steady gains for a comfortable retirement, and they’re becoming more and more aggressive in their demand that they have to get rich and get rich quickly. There is such a palpable anger in these spaces. The whole meme stock world is defined by an explicit breezy jokiness, but underneath that culture is filled with anxiety and need.

A company I haven’t mentioned yet is Nvidia. They’re geriatric by tech company standards, but have recently become exponentially more valuable. For most of their history they were making money designing graphics cards for PC gamers, a niche concern. But it turns out that the kind of computation that LLMs need (and they need an immense amount of compute) is very well suited to the type of chips that Nvidia builds. This has made Nvidia an immensely valuable company in the past several years; their growth looks pretty hockey sticky. If you bought a bunch of Nvidia stock a decade ago, well, congrats. But conceptually, that valuation only makes sense if you have a maximalist perspective on the eventual profits of all those companies using AI chips. The money’s gotta come from somewhere, eventually. We don’t know that AI can live up to the overheated predictions that are being made about it, and in many cases, even if the technological predictions come true it’s entirely unclear how they’ll be monetized and what kind of profits can actually be wrung out of them. Right now these firms are just throwing bales of money onto a giant bonfire, all based on hope.

Of course, the same irrationality that’s powering all of this also lives in the markets, and there’s no reason to think that the bubble will pop any time soon. The need to find value where it might not actually exist is an essential part of capitalism. But it certainly could happen that there’s some event that causes the markets to suddenly recognize that what these AI products actually do could never justify the truly immense costs of building and running them, and if that happens, watch out. The amount of cash that’s been poured into all of this, the media’s casual association of the future of AI as the future of our economy, the degree to which investor confidence rides on hopes for artificial intelligence…. It could be sufficient to put the country into a depression, if things broke the wrong way. But that’s how badly people want to believe, how much Americans need to feel that hope springs eternal. If there is no prophecy telling us that we’re about to become fabulously rich, we’ll just write one for ourselves.