As ironic as the situation stands, wherein a human created chatbot ‘still in its infancy’ now threatens human creativity, a latest incidence has jolted academics even more.
Ai finds statistical correlations. It cant find meaning because that requiers a structure wich it is not peogramed to make.
For example the goodreads ai thinks i am japanise always recomending me wierd books in japanise. I dont speak it or even liked any anime or japan related stuff on the site. I moslty revewed campbell era scify, and aparently some of the more obscure srauge de camp stuff is only read by japanise people.
So it cant find the meaninful traits of that type of literature or even that it is cambelian but sees that people who read the queen of zamba also read a lot of things in japanise.
Mbs probably do something similar. Just say buig words in stereotyped paterns without understanding any of it.
The real question is ho do you build these knoledge structures. And its not even about cpu power its about a theoretical framing that most of the people working in ai are too lazy to make. Here has been some work recently in the other direction applying some of chomskys stuff to ai models. But that is not the fancy gpt bulshit.
Ai finds statistical correlations. It cant find meaning because that requiers a structure wich it is not peogramed to make.
For example the goodreads ai thinks i am japanise always recomending me wierd books in japanise. I dont speak it or even liked any anime or japan related stuff on the site. I moslty revewed campbell era scify, and aparently some of the more obscure srauge de camp stuff is only read by japanise people.
So it cant find the meaninful traits of that type of literature or even that it is cambelian but sees that people who read the queen of zamba also read a lot of things in japanise.
Mbs probably do something similar. Just say buig words in stereotyped paterns without understanding any of it.
The real question is ho do you build these knoledge structures. And its not even about cpu power its about a theoretical framing that most of the people working in ai are too lazy to make. Here has been some work recently in the other direction applying some of chomskys stuff to ai models. But that is not the fancy gpt bulshit.