ChatGPT as a Digital Parrot

I am doing AI since 1998. Those days, when I wrote my first “AI” app in C for my Neuroscience PhD at Medical University - the app which was doing recognition of cells in microscopic images (it was also a “dApp” running on a computer cluster of the University network :) - those days the “AI” was still called “neural networks” and perceptrons. So I have some knowledge of the industry.

It is so much fun seeing people being scared of digital parrots of ChatGTP kind - and governments addressing those “issues” in their usual “COVID dovecot” style (“fly in, shit over everything around and fly out”), that I can’t stop laughing.

Yes, digital parrots will cause many people to lose their jobs - but this fact says much more about people lacking real intelligence than AI possessing any intelligence. Jobs were taken from people many times before - agriculture took the jobs of hunters, engineering took the jobs of slaves, automation took the job of factory workers etc, etc. None of them was a form of “intelligence” - neither ChatGPT is.

Do you know who was fighting these technologies for the “good of people”? Luddites. So congrats to the EU, which is now run by Luddites - and those “advanced opinion leaders and innovators” from the “crypto industry” like Vitalik Buterin welcoming regulations.

“Have fun staying stupid” - probably this should be the new meme for those who’d like to address neo-Luddites.

But what about real GAI? Is it possible? Would it happen one day? So far, even most humans, possessing the same brain, is still not able to develop a proper generic natural intelligence (GNI), so what we can say about GAI? Of course, it is possible, but not with training on human-generated data sets of some large statistical models using differential equations and calling that “neural networks”. These neural networks are like guinea pigs: not a pig - and not from Guinea: they are neither “neural” nor “networks”.

Last updated