Chomsky has responded to this in some interviews but he didn't make a good case from what I remember. If I remember, his argument was like 'airplanes fly but we don't learn about bird flight from studying airplanes' and 'submarines go in water but we don't learn about fish swim skills from studying submarines' so therefore 'even if llm is good at language its a computer and we dont learn human language from studying computer'.
OK I found Chomsky's response. I'm not sure it's the best url, it was the first one I found on Google. They refer to the piece we are commenting under directly (modern language models refute chomsky's approach to language) https://mronline.org/2023/04/24/chatgpt-and-human-intelligen...
I should add that I find Chomsky's takes on LLMs to be exceptionally bad. I'm not a Chomsky hater. I just think his LLM takes are super dumb. On the other hand, I'm not sure if his dumb takes on LLMs are relevant to the question of whether LLMs have refuted his 'approach to language.'
OK I found Chomsky's response. I'm not sure it's the best url, it was the first one I found on Google. They refer to the piece we are commenting under directly (modern language models refute chomsky's approach to language) https://mronline.org/2023/04/24/chatgpt-and-human-intelligen...
Also the hacker news discussion of Chomsky's NYT piece https://news.ycombinator.com/item?id=35067619
I should add that I find Chomsky's takes on LLMs to be exceptionally bad. I'm not a Chomsky hater. I just think his LLM takes are super dumb. On the other hand, I'm not sure if his dumb takes on LLMs are relevant to the question of whether LLMs have refuted his 'approach to language.'