Wall Street Finally Figures Out What Atari Knew in 1983

Listen, I had to scroll past three different pop-up overlays, a video I didn't ask to watch, and what I can only assume is a cookie banner written in Japanese—or possibly ancient Sumerian, I couldn't tell—just to read the headline at Financial Times. I'm pretty sure the actual journalism is behind a paywall now. I also watched an ad for luxury watches autoplay directly into my ear. This is the business model. We've built a civilization where the *reporting* on economic collapse is itself a collapsing business, and nobody finds this ironic.

But fine. Chip stocks are falling because AI is apparently done being trendy. This is what we deserve, frankly. In 1982, if you wanted to compute, you went to a chip manufacturer and said, "I would like one chip, please." You got a chip. It did one thing. It did that thing reliably. Nobody promised it would achieve sentience or rewrite your novel or tell you what your grandmother meant when she said "we'll see."

Now we've got investors dumping billions into semiconductors because some executive promised that artificial intelligence would cure death, or rent itself out, or something. I have a theory that the entire AI sector exists primarily as a subscription service that hasn't figured out how to bill you yet. Give it time. Soon you'll need to subscribe to use your own computer. "Pay $19.99 monthly for the thinking feature." That's coming. I can feel it.

And another thing—remember when "artificial intelligence" was something a teenager had to figure out how to do with a Commodore 64 and a stack of gaming magazines? Now it's investors' retirement funds vanishing because a chatbot trained on everything on the internet turned out to be, shockingly, pretty mediocre at doing actual work. The headline should read: "Rich People Confused That Magic Rocks Did Not Perform Magic," but I guess that doesn't move units.

Chip manufacturers in 1982 had the decency to fail quietly. They didn't sell you on a dream. Intel's 80286 just *existed*. You could compute with it. Nobody promised it would transform civilization. The stock moved because the computing demand was real and growing—people wanted to *do things*—not because someone released a promotional video of a chip having a philosophical debate with itself.

So here we are, watching the inevitable correction. The great AI bubble gasping its last before the next bubble inevitably replaces it. Probably something to do with quantum computing, which, let me tell you, is definitely about to require a subscription just to access the computing power.

—Edgar