For the past few days, my Twitter feed (yes, it’s still there) has been cluttered with dialogs with the ChatGPT chatbot. Some are on the level of polite and trivial human conversation: the form is correct, but not much fact. Where fact is called for, it is often incorrect, but presented in a correct form.
There are also several art generators available. Neither they nor the chatbots can be called “intelligent.” As far as I can tell (and I’ll be happy to hear if I’ve missed something), they are basically weighted averages of their training sets, which are very large and consist of samples of art or conversation. A part of the program also recognizes, from those training sets, questions from humans and what might be appropriate responses. The programs also add the inputs of people asking them questions to their data bases.
That is a significant achievement, although machine translation, which relies on similar operations, impresses me more, particularly of agglutinative languages.
All this is enabled by large computer memories and fast computation.
Apparently factual matter is too large a database to accommodate along with syntax and form. Yesterday I saw a chemistry-related query that the chatbot responded to with a soup of technical terms, all used incorrectly. Not a word salad, because the syntax was fine, more of a concept salad.
It’s machine learning, the kind of gigantic computation that machines are good at. Intelligence would incorporate fact as well as syntax.
I don’t give examples, because I think reading them is harmful. They do something to my brain that makes it easier for concept salad to slip by. And that, of course, is a major tactic of the gaslighters and liars we are now dealing with politically. I want to remain sharp and ready to eviscerate their lies.
The art generators use the products of many artists, some long out of copyright and others who are current. None seem to have obtained permission of the artists to use their work. So when a weighted average uses an artist’s work, they are neither acknowledged nor compensated.
A further difficulty with these products is that they contaminate the pool from which they draw. At least one coding site has banned the use of ChatGPT for this reason. Cue “The Sorcerer’s Apprentice.”
Cross-posted to Nuclear Diner