Op-Ed: AI investment losses are getting neurotic and expensive. Is sanity finally creeping in?

The general reassessment of AI investment by the markets is long overdue and it can do the job that market babble refuses to do. The overall sales pitch of “It’s weird, chatty, untrustworthy, and we don’t really need it for most daily and business things” somehow isn’t quite as appealing as it was. At long last, some real evaluation is finally happening where it matters. The big money is getting cautious and demanding.

Please excuse the number of links. This subject is more like quantum physics than most. Every aspect entangles with everything else. The links are intended to add some scope in context and current pictures of what’s happening and not happening.

This generation of AI is at the two-year-old stage and will be for a while if not regulated. It’s a toddler, difficult to manage, and it’s messy. Worse, it’s astonishingly expensive. You’d think that the markets, aka The Last Word in Moneygrubbing, would have started asking questions a lot sooner.

It was unproven tech at first, now extremely complex in practice, with real risks. The questions still aren’t being answered. Self-regulation as proposed by the US is a bad joke in far less demanding environments, and that euphemism’s not getting a lot of traction, either.

In AI environments, self-regulation simply cannot work. AI economics are by necessity integrated with the entire financial and wider economic systems. If your property is solely protected by the whims of vested interests, how protected can it be? How can they “protect” external entities from themselves?

Just as well nobody stops to think about what else you could do with those few trillion bucks, isn’t it? For Big Tech, it’s looking like they brought their own cliff with them to have something to jump off. Such huge capital commitment, and no real sense of direction isn’t selling, either.  

Financing AI is also looking suddenly tentative and more wary by the headline. AI debt is now a real major issue, with refinancing another noose around real development. The lack of hard market products and dubious, very mixed reports on AI ROI aren’t helping. So far, “neurotic” is the right answer. Tolerance is a commodity, too, and it’s running out.

The beginning has not yet begun

There’s more neurosis to come. The shockwave hasn’t really hit yet. This is the early event horizon. AI isn’t yet mainstream in the same sense as most common digital systems. This is the very early beginning of true macro-scale automation, and it’s hardly what you could call systematic.

Nor is the process of AI integration in any sense stable. This tech can easily make itself obsolete overnight. You could find yourself owning a museum at that sort of cost. The never-ending Next Big Things can and mostly will wind up on the scrap heap almost instantly. Nothing about AI is being done systematically as usual. Even the energy demands of AI have been approached like a lottery. Somebody might win, but what about paying for and managing the huge infrastructure loads?

The wholesale destruction of the employment sector and repositioning of every goalpost in life doesn’t seem to work for some reason. Destroying the foundations of modern economics with no ideas about what happens next doesn’t yet include any economic planning at all.

Daily life could become unrecognizable in a decade without so much as a snappy little brochure about how to live in it. What are 8.3 billion people supposed to do with this constantly evolving mess? How do people fit into this existential demolition derby on any level in real life?

Sanity

You may remember sanity. At one point, it even included knowing what you were talking about.

Enough with the Endless AI Fashion Week Bling-A-Thon. You’d have to be almost totally techno-illiterate to believe a word of it. It’s wasted space. Enough of blundering into every predictable practical obstacle like nobody saw it coming. To hell with panic. This situation needs managing, not rhetoric.  

Let’s look at a few actual situations for a change:

Consumer-level AI is at roughly high school-level comprehension. Ironically, that’s good. Some general basic training would help. Gen Z are getting high doses of AI as it glues itself onto daily processes. Familiarity breeds fluency, eventually.

Business-level AI isn’t really all that flashy or particularly impressive. It’s either useful or it isn’t. It shouldn’t take a crystal ball to figure out what AI is worth having. That’s the only real issue.

What’s needed is high functionality and efficiency. Nothing else. AI can do wonders in the sciences. It can save billions in processing tests and projections, and above all, time. This is the sort of work at which AI is truly priceless. Never mind “chatbot culture”, or whatever this idiocy is supposed to be. The priorities must go to high-value delivery, not meaningless gimmicks.

Automation can free up a lot of human skill sets. The one and only valid argument for AI in general is that it eliminates drudgery. That argument worked for the Industrial Revolution, and it took 50 years to adjust society to manage the displacement of people and old economics. Those freed-up skills can rewrite humanity on humanity’s terms.

So far, we have fluency, hard business values, and efficiency. You could call that sanity. Does it sound worth it to you?