Chomsky's views on this topic are kinda outdated and don't go beyond his field of linguistics. Yeah, LLMs is still just a statistical parrots, but with the right tooling around it's a very powerful instrument for analysis for large amounts of data and it can be very helpful in scientific fields.
I agree that LLMs work best the more focussed they are. I think his point is more directed to people who think AI is just like the human mind, that it thinks, learns, corrects, interprets, is creative and even feels or loves.
this is losing focus, the danger of AI isn't about how incapable it is, but it lacks humans' morality and desire, and will always follow orders, ethical or not.
"human brain isnt statistic machine" is also completely wrong in both education and psychology, all children learn by following examples.
It isn't though? GenAI isn't intelligent, hel it doesn't even think; it's a pattern recognition system that assigns the code of words to the code of patterns and colors and spits out a response based on those parameters.
What's factually wrong about it? That's precisely how these machines are made, and work.
You create a data model and have a simplistic program figure out what to do with the data based on it's purpose. There's no intelligence in these machines.
Comments
"human brain isnt statistic machine" is also completely wrong in both education and psychology, all children learn by following examples.
You create a data model and have a simplistic program figure out what to do with the data based on it's purpose. There's no intelligence in these machines.