Ten brutally honest predictions about the future of AI.

Observing that enthusiasm for so-called AI is diminishing by the day--a recent survey suggests 86% of the public distrusts AI--that tech "experts" over-promise again and again, and that most predictions about the technology have been proven dead wrong ("you could get more reliable information by consulting tarot cards or a Ouija board"), cultural commentator Ted Gioia jumps in to give in his own "ten brutally honest predictions about the future of AI."

The first is about where the tech might find a home. Because in the creative world, where it's been desperately oversold, despite all the hype it's seen increasingly as "the microwave meal version of culture."

Food is the useful analogy here. When frozen dinners and microwave meals were introduced, many consumers embraced them as the tech-driven foods of the future. But over time, the public figured out that more tech isn’t always better when it comes to food.

The same will happen with AI music, AI writing, AI art, and all the rest. Like the microwave meal, it will be cheap and easy. But anyone with discernment and taste will prefer handcrafted alternatives.

The slop is just too obviously slop.

Which brings him to money, which is where, here reckons, "the AI crisis will intensify the most" -- and mostly because there will still still be too little coming in to justify all squillions of capital spending ...

The cash getting invested in bots is beyond anything ever seen before in the history of capitalism—already $3.7 trillion has been spent, and the outlay is growing by another trillion per year.

This investment will never generate an adequate return. It can’t. The numbers just don’t add up.

Let me make a comparison. For less than that amount, Silicon Valley could have acquired every movie studio, every video game company, every major record label, and all the big publishers in the world—and still have plenty of cash left over.

So do you think the average person will pay more for AI each month than for all that entertainment? Is AI really worth that much? Not at all—at best, the true believers will pay roughly the same amount as a single Netflix subscription. But the market cap of Netflix is a tiny percent of total AI expenditures. So the mismatch between the level invested and the resulting demand is unprecedented. ...

This is an easy prediction to make. Get ready for write-offs and financial ugliness of the most extreme kind. It’s hard-baked into the business plans, and keeping the mess in the oven for longer won’t make it taste any better.

Why so little revenue? Because it's so increasingly unpopular.

Silicon Valley has consistently underestimated how much people hate AI. ... The public is shifting from passive grumbling to active resistance.

This is a problem even billionaires in their bunkers can’t ignore. They thought that the right response was to push harder. But they thought wrong.

They are still living in denial, but that won’t last for much longer. The public is getting more fierce, and the hyperscalers’ cash resources—which once seemed almost infinite—are dwindling. Silicon Valley has the weaker hand here, and nobody believes its bluffs anymore.

So what happens next? Read all ten brutally honest predictions on the future of these large learning models.