Dreams make good stories, but everything important happens when we’re awake.

-Duncan Idaho, Dune

Technology companies have a sort of mystique around them — especially software companies. Meta (formerly Facebook) did roughly $200 billion in revenue in 2025. That’s a decent chunk of change! But software companies make money much differently from most organizations. They are built around the idea of scale.

Scale is Everything

See, Facebook only has to program “Facebook” one time, and then it can serve 100 people, and then a thousand, and then a billion, with (relatively) minimal changes. They can scale with very little effort — maybe lighting up a new server every 100,000 new users.

Walmart made almost $700 billion in revenue in 2025, but scaling looks much differently for them. If Walmart wants new customers they have to go to a new city, find a plot of land, figure out if there are sufficient customers around, build a store, staff it, fill it, and only THEN do they see an increase of revenue.

Because of this difference we’ve come to idolize software companies — especially in the world of venture capital. Anywhere between a third and a half of venture capital funds go to software companies, while the next highest category (“hardware” — still tech based) got roughly 15%. In 2025 ALL of the top ten investments were software companies, the majority of which were AI. Only one company had anything to do with the “physical” world at all (Project Prometheus — a company dedicated to “applying AI technology to physical tasks” soooooo … still a software company at heart).

Venture capital funds love these companies because they can create a product and then scale quickly — in other words, they can create a fast turnaround on the investment. This type of ROI creates headlines, but it also creates some pretty unfortunate attitudes.

The $1.5 Trillion Company that Doesn’t Matter

Meta (formerly Facebook) has a market cap of $1.5 trillion dollars. That’s a lot of money!

If Facebook disappeared tomorrow, what would happen?

Well, if we averaged out all the positives and negatives my guess is that it the disappearance of Facebook would be a net positive for the world. Our political discourse would improve. Teenagers all over the world would get a self esteem boost. Our focus would sharpen. And I would never have to look at a captioned picture of a minion ever again. Net positive!

I’m picking on Facebook because they’re a perfect microcosm of the problem. Facebook, allegedly, exists to connect people. It doesn’t really exist to do that — it exists to sell the attention of the masses to advertisers — but let’s take them at their word. They exist to connect people.

There are a million ways to connect people! Forums. Chat rooms. Signal. Email. Text. Phone. Letters. Smoke signals. Facebook is “easy” but the ease doesn’t help anything — it actually makes a lot of things worse. Facebook is just a middle-man, inserting itself between you and your friends and saying “I think I could profit off your friendship.”

In the end Facebook doesn’t matter — except to one group. It matters to investors. Because it scales easily. Is it valuable? Is it … good? Is it even just … not evil? Probably no to all of those. But it scales like crazy.

All hail the middle-men

In the end most software companies follow the same pattern — there are two things that are connected in some way, and they get between them and charge one or both sides.

Sometimes this is perfectly acceptable. Think of Microsoft Word. There are writers, and there are readers, editors, publishers, etc. In the old days the technology that connected these people was the typewriter.

In comes word processors, offering to make writing easier for the writers, editing easier for the editors, and the moving of the written word easier for publishers and readers. And you know what they priced it like?

A typewriter! You paid one price, you got access to the software for as long as you wanted. George R.R. Martin (author of the Game of Thrones series) famously still uses Wordstar 4.0, which came out in 1987.

Obviously modern companies don’t use that model — they use subscription pricing, so that a word processor that once cost $100 every 3-5 years can now cost $10 a month (or $600 over that same 5 years).

Spotify is a middle-man between artists and listeners. Youtube sits between creators and watchers. Grub Hub connects restaurants and delivery drivers, AND connects delivery drivers and hungry customers.

The pattern is the same — squeeze into the gap between two people and start charging. With software you can do that and scale up and make tons of money for yourself and for investors.

The problem is, ultimately, truthfully, none of it matters.

What matters to you?

I want you to think for a second. Maybe even grab a paper and pencil. Now write down the ten most important things to you. What are the things you couldn’t live without? What are the things you’d take with you to an island? What are the things that give you joy on a daily basis?

How many of those things are purely digital? They have no analog or physical component whatsoever — they can’t be purchased physically, they take no physical form, they are 100% virtual?

I am going to guess very few of the most important things in your life are digital.

And that’s because digital is the domain of the middle-men. Maybe you said “I couldn’t get by without my phone.” Well, for one, that’s a physical object. But secondly — why? What is it on the phone that you need?

Movies? Physical things. Games? Even games can be bought physically, and require something physical (the phone or game console). Music? Come on now, aren’t you into vinyl yet? All the cool kids are doing it. Your journal? Your to-do list? Your books? All things that are physical. A service has just gotten between you and the actual physical item and said “I can probably do that better/easier/faster (and make a profit in the mean time).”

And sometimes that’s fine! There are some amazing services out there that take something difficult to do in the physical world (like take your vinyl collection everywhere) and make it much easier. I’m not saying that’s bad.

What I’m saying is that, because these companies can be immensely profitable, and make lots of money for venture capitalists, we fetishize them. We assume that they are the most important things on the planet because the numbers next to them are the biggest. But they’re not.

Why hasn’t AI taken over the world?

Ultimately, this is why AI hasn’t, and won’t, take over the world. The people who are creating AI swim in the digital world — everything they touch is mediated by computers. And that’s how they like it!

But the most important things in life involve, you know, meat-space. They involve the physical movement of goods and the physical performance of services. They involve direct human interaction. And although tech companies can squeeze in the spaces and charge rent, they can’t replicate the thing they’re charging rent on.

Facebook put itself in the middle of human connections, but the valuable thing wasn’t facebook, it was the connection. The fact that they have convinced everyone that THEY provide the value is a credit to their marketing department, but they are wrong. They don’t provide the value. At best they facilitate value creation by others.

Grubhub can put themselves between a hungry customer and a restaurant (and a driver), but they can’t make a meal, and they can’t move it anywhere.

Airbnb can put themselves between someone with space to rent and someone who needs space, but they can’t create rentable space themselves — if they could, they would! But that’s a whole different type of company and it doesn’t scale like software! It scales like builders, and like everything in the physical world, that scales SLLLLOOOOOWWWWW.

So let’s stop freaking out and remember that what creates anything of value in this world is people, not algorithms — regardless of what the people who create algorithms would like us to think.