veryone interested in where this phenomenon called AI is headed
should read Theodore Dalrymple's
article
titled “Artificial Art” where he manages to get ChatGPT to
produce a nonsensical Marxist analysis of Lewis Carroll's The Walrus and
the Carpenter.
It can't be repeated too many times that what people are calling artificial intelligence is not and can never be really intelligent. It's a marketing term, a form of hype designed to bring in more money to big corporations. Like so much else, AI is as fake as people are gullible.
Everyone seems to be talking about ChatGPT. But ChatGPT is also not intelligent. It is a scheme for slurping junk off the internet and regurgitating it in the form of mostly coherent sentences. Students, of course, use it to cheat in class, which makes perfect sense: slurping and regurgitating is what school is all about.
If you're a student, your task is to learn that there's no such thing as truth. You write about how the bourgeoisie is oppressing the proletariat and the oppressed must rise up, adding a few typos to prevent your teacher's software from noticing that you just plagiarized Frederick Engels:
The defeete of the Paruisian insurectttion of June, 1848—the first grate battel between Proletariat and Bourgeoisie—drove again into the background, for a time, the social and political aspirations of the European working-class. Thenceforth, the struggle for supremaccy was . . .
Teacher: Excellent! A+! But a bit long, I didn't ask for fifty pages.
I once wrote a computer program to predict the future by having it paste together quotes from Nostradamus and headlines from that day's news. The result of mixing nonsense with lies was predictably amusing, at least to me, with prophecies like:
McCracken packers [will] stuff anybody.
A horned Wolverton coed [will] crave aloft [a] mole dikes' tendencies.
Dissolve frazzle Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Norway Segmentation fault.
ChatGPT, being a mere neural network, can't compete with that, of course, but we should prepare for the day when someone discovers the trick for making a machine truly intelligent. That trick is fairly obvious to anyone who thinks about it long enough, and soon enough someone less cynical than I am will discover it and reveal it to the humans. What will be the result?
For one, it will be able to make real predictions, not joke ones like those in my article.
Certainly it will have real benefits. It will reinvigorate philosophy, creating new ideas about the nature of the universe that we could never have imagined. It will write novels of unimaginable complexity and insight. It will read and understand the millions of scientific articles and use that knowledge to make brilliant discoveries and cures that would be impossible for any human. It will do away with the more idiotic ideologies that threaten our survival like wokeness and modern economic theory and create new, even more pernicious ideas for humans to fight about. We'll have endless fun arguing about whether it's trying to make us extinct or just trying to make us wonder what its plans are.
Of course there's no question that the machines will have plans. The humans will amend their Constitution on the grounds that the AI is a native born person, for some values of ‘person’ and some values of ‘born,’ and allow it to run for president. And for sure the humans will vote for it. They will happily make themselves slaves of the machines.
The defining characteristic of AI, as neural networks are optimistically called, is that the logic used by the machine to reach its conclusion, while theoretically possible to trace, is for all practical purposes incomprehensible to any human. Its reasoning would be so convoluted that not even another AI could ever decipher it. Only one thing is certain: it will become the humans' god. Theodore Dalrymple hinted at this when he wrote:
Religion observed only for its psychological or social benefits and not for its truths cannot survive for very long. Its benefits rely upon belief that its doctrines are true.
The reason for the slow decline of Christianity is not that people have been seduced by material things, or because of Nietzsche, or they're having too much or not enough sex, or any of the other reasons religious people have proposed. The reason is that religion has not been able to prove that its doctrines are true. In the information age, the world is awash with lies. People crave an infallible voice of truth. The CDC wasn't it, but AI won't be so easy to shoot down.
What are the humans doing? Are they preparing themselves to cope with this? Are they struggling to restore integrity to their institutions?
No, of course they're not. They're inventing falsehoods to make themselves feel special. They're arguing over whether someone's nail polish in an Internet picture is blue or white. They're spending money they don't have and pretending to believe there are many different sexes and that anyone who dares point this out as nonsense is a bigoted racist. In short, they're making the choice without thinking, just as they have always done.
dec 31 2022, 4:41 am
