HOW THE ECONOMY REALLY WORKS (OR, TECH, AI, THE PAST, AND THE FUTURE
I. THE MAGNIFICENT FIVE AND A HALF
There’s a myth in finance, economics, and business. That things change. They don’t.
Right now, the US economy is powered almost single-handedly by “tech.” The American stock market certainly is.
But “tech” is a misnomer, which is why I put it in quotes. At least if you want to understand the economy. Here’s how “tech” really works, and what it is.
The “Magnificent 7” is made of Google, Meta (aka Facebook), Apple, Nvidia, Amazon, Microsoft, and Tesla. Let’s drop Tesla, because it’s a has-been of a company. Microsoft is sort of hanging in there, trying to constantly play catch up. That leaves us with a Magnificent five and a half.
III. WHAT BUSINESS IS “TECH” REALLY IN?
What business are the Magnificent five and a half really in? The answer is going to surprise you, and maybe even make you chuckle.
Google and Meta are in the advertising business.
Amazon is a mail order company.
Apple is something interesting, which is a luxury company masquerading as a “tech” company. That’s why Apple commands a premium, by the way. Feel your iPhone or MacBook. Doesn’t feel much like junky “tech,” does it? It feels (and behaves) more like a nice watch or car or suit or what have you. I’ll write more about this. For now, let’s leave Apple on one side.
That leaves us with Nvidia, who makes the chips that power AI.
IV. WELCOME BACK TO THE 19TH CENTURY
So. The US economy, and its stock markets, are powered almost single-handedly by “tech.” But tech is…surprisingly…old school.
The majority of the Magnificent Five and a Half are in…businesses that’ve existed forever. Amazon’s the new Sears, more or less. Google and Meta, meanwhile, are just in the advertising business.
There’s nothing new under the sun when it comes to finance, economics, and business.
I’m emphasizing this to demystify what’s going on. All of this is surprisingly…almost...19th century. The economy is powered by a) two big advertising companies and b) a gigantic mail order behemoth (and a handful of companies who make the infrastructure for this to work.)
Now. If you doubt me, consider for a second how much of a mail-order Sears Amazon really is. Less so, in fact, than Sears. Sears claim to fame was that it’d vet what you’d get. But on Amazon, you can scroll for hours, we’ve all done it, sort of biting your nails, wondering if you’re going to get garbage, crap, a fake, or something that, out of a bad date, looks “nothing like the pictures.”
The business models are as old as time (the Mad Men sold advertising, too.)
Google’s business is about 80% advertising, and Meta’s is about 99% advertising. Amazon, meanwhile, makes about 75% of what it does from it’s gigantic “marketplace,” in various ways.
It’s just mail order and advertising.
IV. WHERE SUPERNORMAL PROFITS COME FROM
Why am I teaching you all this? So that you understand the reality behind the spin and the hype, number one, and number two, so that you understand how to think about what’s supposed to happen next.
That is that AI is supposed to catapult these incredibly profitable companies profits to the moon. Will it?
What’s different about mail order and advertising today, what makes these companies more profitable than Sears and the Mad Men, is that they’ve automated parts of it.
And that automation hasn’t always been good. Amazon is like the world’s shittiest bazaar. Google and Meta ads, meanwhile, have all the charm of a tech bro trying to propose marriage to you. Still, despite that, automation has done what it does, which is create supernormal profit margins for them.
So: it’s just mail order and advertising, with automation behind it, to jack up the profits. (And, beneath that, a healthy dose of what are called network effects, meaning, these giants have created monopolistic marketplaces for advertising and mail order stuff.)
V. THE DREAM OF A NEW BUSINESS MODEL, OR NORMAL PEOPLE DREAM OF LOVE, BUT TECH BROS DREAM OF $$MONEY
Enter AI. How is AI going to change all this?
The dream is that…there’s going to be a new business model.
Nobody’s “figured it out” yet, which is Silicon Valley code for “it doesn’t work very well so far.” It’s sort of important to note how money-obsessed this culture, this thing, really is. Normal people don’t spend their lives dreaming of business models. They dream about love, life, happiness. And for that reason good businesses aren’t just about money.
Let me try to explain.
Google makes about $3 per month per user. From advertising. That adds up to the billions it earns over the course of the year.
Did that number just make you shudder? Maybe you know what I’m about to say next.
The dream is that AI will become a “business line” itself. And it’ll do that because people will pay for it. They’ll pay, according to Silicon Valley, something like $30 a month for it. That’s sort of what it has to make to be profitable, given the tremendous investments in it. We can’t say for certain, because the companies are cagey about the numbers, but rest assured, those numbers are not in the single digits.
So. Google currently makes $3 per month per user from advertising. The AI dream is to make $30 per month or so per person.
That’s ten times what Google makes from its existing business model.
Meta’s situation is even crazier. Its revenue per user, $5-6, is even higher than Google’s, even though, LOL, it’s a terrible, soulless sort of experience.
And yet that’s still not even remotely close to the notion of charging people double digits for some sort of AI subscription.
Are you beginning to see the problem here?
VI. HOW BUSINESSES (REALLY) WORK
“Tech” has business models that work very, very well. They do that for a reason. They’re simple, and they’re old. They’re time honored and battle tested, stretching back through the centuries, because they solve problems in relatively efficient ways, while providing profitability.
But to make the AI dream work, “tech” is going to have turn it into a new business model, and charge people for it, forever. And it’s true that many “will pay for it.” No doubt about that. That’s not the point. The point is that that is a very, very different business. In its entirety.
You don’t pay Google or Meta for much of anything right now. You pay Amazon begrudgingly a very, very small amount, for the absolutely awful, just dismal, horrible service it barely provides. To say that Amazon offers “quality control” is like saying Donald Trump talks a lot.
Getting people to accept a new business model is hard. How hard? Nike failed at it. So did most of the car companies, not once, but over and over again, for decades. So did most of healthcare. Etcetera.
A new business model depends crucially on having a brand people genuinely trust, respect, and admire. This is the Number One Thing in business, in fact. A brand is the most valuable any successful business has. Always.
Does “tech” have people’s trust and respect?
VII. THE PROBLEM WITH A VERY SUCCESSFUL BUSINESS MODEL IS…THAT IT IS ONE
But even that’s not the real problem
The real problem is this. AI is going to eat these companies existing business models alive.
This is an old problem in b-school lingo: cannibalism.
Let’s think about it.
There you are, and me too. We type something into Google, and now we get “AI results.” But guess what? The AI results aren’t as profitable as the regular ones, because Google hasn’t figured out how to put ads into them. It will, but even then, the economic problem remains very simple: if AI works, then the ads become more efficient, meaning that you have to hit people with less of them to sell something. See what I mean by “cannibalism”?
The same is true for Amazon. Let’s say that AI can tell you, hey, don’t buy this whatever, it sucks, it’s terrible, it looks nothing like the pictures, the seller will never give you your money back, the whole thing will just be a hassle, buy this one instead, or maybe try not buying any of them at all, and try this thing, that’ll do all three of the things you want at once. See what I mean again? And again, this is economics at work, which is that the uncertainty of buying some piece of crap at Amazon goes down, and so far, that uncertainty works very, very much to Amazon’s favor.
So even if AI is “embedded” into the existing models, the outlook is…dubious.
Can you make advertising and mail order more efficient using AI? Sure you can. But right now, these supernormal profits are coming precisely frominefficiency. From the fact that you have to buy a thousand crappy text ads, or pay a dozen grinning influencers, or order three of the same thing just to get one that doesn’t suck.
So: efficiency is the enemy, in many ways, of business models. It’s good for the economy, that’s true, but it drives down profits.
VIII. THE FUTURE OF THE ECONOMY
Where does that leave us? I just wanted to share some thoughts with you. About the economy, finance, business, and what really happens underneath the hood of it all.
I think that the case of AI is dubious. Not because it’s socially creepy or culturally toxic or just plain theft, which it is. But even the hard-nosed business case for it.
Eventually, I think that Wall Street and so on are going to have sit down, think this all through clearly, do some simple math, and see through the hype. Let me try and do that for you now.
The US economy is powered by advertising and mail order. LOL, it’s funny when you think about it that way. It’s not stuff that’s really…rocket science…quantum physics…to begin with…which is why…applying AI to it…isn’t likely to yield as much as people think.
How much more efficient can you make a pretty dumb, already very automated thing?
And that’s also why, by the way, if AI takes the massive number of jobs it’s already beginning to, the economy is hypersensitive to the catastrophic consequences: advertising and mail order depend on people spending money.
Think about all that. Let me know what conclusions you reach.
Love,
Umair (and Snowy!)
❤️ Don't forget...
📣 Share The Issue on your Twitter, Facebook, or LinkedIn.
💵 If you like our newsletter, drop some love in our tip jar.
📫 Forward this to a friend and tell them all all about it.
👂 Anything else? Send us feedback or say hello!
Member discussion