Sixty Billion Reasons They've Ruined Everything

First of all, can we talk about the Yahoo Finance website for a moment? Because I'm trying to read one sentence about computer chips and I'm being assaulted by seventeen different notifications asking if I want to enable something or subscribe to something or confirm that I understand I'm being watched. There's a banner at the top, a banner at the bottom, a banner that slides in from the side, and I'm fairly certain one of the pop-ups is eating another pop-up. In 1982, if you wanted financial news, you picked up a newspaper. It was already printed. Nothing could interrupt it. Nothing could autoplay. Nothing could ask you to confirm your location via seventeen different methods. You just read the thing and moved on with your life.

But anyway: Supermicro. Sixty billion dollars in backlog. That's the headline they're excited about. Sixty billion dollars in orders for computer chips. And the stock jumped because their profit margins got wider. This is supposed to be good news and I'm supposed to care and frankly, I don't understand why anyone cares about anything anymore.

In 1982, you knew what a computer was. A computer was a thing. You could see it. It sat on a desk or it was the size of a refrigerator in a basement somewhere, and it did one job or maybe two jobs if you were lucky. You could *understand* what it did. My Commodore — and I'm not going to pretend I had a Commodore because I didn't, that's the thing — but somebody's Commodore could play a game and it played the same game every time you turned it on. Same pixels, same beeps, same everything. Predictable. Honest.

Now Supermicro is sitting on sixty billion dollars of backlog that I'm fairly certain nobody actually understands. What are these chips for? Artificial intelligence, they say. Machine learning. Large language models. Which is apparently what we're calling the thing where a computer pretends to be smart by regurgitating everything it read on the internet, including all the lies. And every company in the world suddenly needs trillions of these things and they're all willing to pay for it which means Supermicro's margins go up which makes shareholders happy which makes me want to move to a cave.

And another thing — we're not even calling them computers anymore. They're "systems." They're "infrastructure." They're "the backbone of the cloud." The cloud! You know what the cloud was in 1982? Weather. Just weather! Now it's this invisible thing that everything relies on and if you want your email or your documents or your entire life, you have to trust that some corporation's data center in three states you've never been to is keeping it safe. And probably charging you for it. I give it two years before you have to subscribe to access your own memories.

And don't get me started on the backlog. Sixty billion dollars means nothing until it ships and then it ships and then what? The people who ordered it have to maintain it and update it and patch it and replace it. They're going to need MORE chips. It's a treadmill that only goes faster. In 1982, if your computer broke, you fixed it or you got a new one and that was the end of the transaction. Now there's always another thing you need to buy to keep the other thing working, and you're renting the whole stack from someone you've never met, buried under terms of service that nobody's read and probably can't read because they're twelve thousand pages of legal text designed to confuse you.

Also, these chips are probably going to be obsolete in six months, which is how they keep you buying new chips forever. Planned obsolescence used to be crass. Now it's just called "innovation."

Anyway, Supermicro stock is up. Congratulations to whoever owns that stock, I guess. The rest of us just get to watch everything get more complicated and more expensive while pretending we understand why.