hy are AI ‘experts’ telling us that the product they’re making
is going to kill everybody? It’s as if a car manufacturer
made something called the Corvair II and advertised it as the
car that kills everyone who drives it. It makes no sense.
One theory is that people really think AI is real. Suppose I invented a new kind of laser and named it “Death Ray.” If I was convincing enough, people would clamor to make me stop. It would become famous and I’d sell millions of them. Another theory is regulatory capture: they’re trying to get the government to regulate it to kill the competition. And there’s the bored-from-too-much-safety theory: people need something to be scared of in their dull lives and gluten allergies just aren’t working anymore.
Or maybe, as some are saying, it’s a fad like, say, tulip bulbs. Everyone else is afraid of it, the movies make it look easy, so it’s fashionable to pretend to be afraid of it. And there’s the China theory: China is paying people in the West to clamor for the government to stop it so they can win the race. That’s also old: they used to warn us that China would win the 5G race. Now 5G is ‘meh’ and nobody cares about 6G and 7G either.
Or maybe it’s because the news media, facing extinction because of their legendary inaccuracy, need something to sell papers. We see this in the UK press: Russia, they’re saying over and over, has a new type of bomb that causes a tsunami and they’re going to use it against England because . . . well, we’re not quite sure, but here’s a picture of a big scary orange nuclear explosion and here’s a target list we helpfully compiled.
Theory #1: Claim inflation
Any or all of the above theories might be true. But the fact is that hype about AI is nothing new. I worked in the field back in the era of the PDP-11. In those days, we theorized and wrote the software ourselves. It was the second wave of AI. The first wave, from the Old Ones—the Perceptron guys—had crashed long before because reality could not meet expectations. The second wave crashed for the same reason: claim inflation. Partly it was competition for funding, partly it was to have one’s own ideas beat the competitors’, but mostly it was blind arrogance. The claims became so preposterous that I switched to the most boring subject known to man: biochemistry. That’s how bad it was.
In the 1960s, people thought HAL9000 was just around the corner. Making an intelligent computer would be easy: just stuff it with so much information it would achieve “sentience itself.” Star Trek repeated it endlessly. Today’s entrepreneurs are the same. They think intelligence is an emergent property and AI will just happen by itself. Sadly, computers are not going to become intelligent by accident. Humans just aren’t that lucky.
Theory #2: They’ve hit a brick wall
What they have now is a new type of search engine. It does pattern recognition about as well as the older algorithms, only slower and more expensively. It also summarizes text, occasionally even accurately. But they have absolutely no idea how to get to the next step: how to make it intelligent. The plan is to make threats about wiping out all humans in order to trick those few who know the answer to reveal it.
They know the only way to disprove their claims would be for someone to disclose to the world the architecture and algorithms needed to produce genuine AI, so we can prove that a chatbot doesn’t really have it.
If somebody did that, the corporations would patent it and, just as with so many inventions in the past, the inventor would get nothing but corporate scorn. The problem is that AI sounds easy, but the closer you look at it, the more you realize it’s not. Dante said it way back in 1321:
Like one who has imperfect vision, we see things that are distant from us . . . but when they approach, or come to be, our intelligence is wholly void, and we know nothing of your human state, except what others tell us.
The person who says that (a character in Inferno) was damned in hell, so at least he had an excuse.
Theory #3: They are having doubts
The outlandish predictions are a sign of self-doubt. We’re supposed to ask: Why would they be warning us of impending doom from AI if AI didn’t even exist? The reason: they’re trying to convince themselves that they haven’t wasted hundreds of billions on fake AI.
It was the same with global warming: the oceans would rise by ninety feet, snow would disappear forever, and we would all be dead from massive one or two degrees of warming. Those claims were made partly to convince us, but mostly for activists to convince each other that it was real. Remember the claim that a billion people will die from global warming? The claims became more and more preposterous because those making the claims didn’t really believe them.
A clue is that the term “AI” has been so degraded they had to invent a new one: superintelligence. Remember how ‘global warming’ was not scary enough so it had to be replaced with ‘climate emergency’ and ‘global heating’? ‘AI denier’ can’t be far off.
The field is about to crash
This time there’s real money involved. People are sick of AI, sick of the fake cartoon-like pictures it creates, sick of the fake music, and just plain sick of hearing about it. Duckduckgo has even created a new search engine (noai.duckduckgo.com) that eschews everything AI, and it’s incredibly popular. The investors are scared, and they should be.
sep 14 2026, 6:37 am
