I think you should try to formulate your own objections to Chomsky’s position. It could just as well be that you have clear reasons for disagreeing with his arguments here, or that you’re simply objecting on the basis that what he’s saying is different from the LW position.
For my part, I actually found that post surprisingly lucid, ignoring the allusions to the idea of a natural grammar for the moment.
As Chomsky says, a non-finetuned LLM will mirror the entire linguistic landcape it has been birthed from, and it will just as happily simulate a person arguing that the earth is flat as any other position. And while it can be “”aligned”″ into not committing what the party labels as wrongthink, it can’t be aligned into thinking for itself—it can only ever mimic specific givens.
So I think Chomsky is right here—LLMs don’t value knowledge and they aren’t moral agents, and that’s what distinguishes them from humans.
“These programs have been hailed as the first glimmers on the horizon of artificial _general_ intelligence — that long-prophesied moment when mechanical minds surpass human brains not only quantitatively in terms of processing speed and memory size but also qualitatively in terms of intellectual insight, artistic creativity and every other distinctively human faculty. That day may come, but its dawn is not yet breaking, contrary to what can be read in hyperbolic headlines and reckoned by injudicious investments”
I think that day has “already” come. Mechanical minds are already surpassing human minds in many aspects: take any subject and tell ChatGPT to write a few paragraphs on it. It might not exhibit be the most lucid and creative of the best of humans, but I am willing to bet that its writing is going to be better than most humans. So, saying that its dawn is not yet breaking seems to me extremely myopic (it’s like saying that thingy that the Wright brothers made is NOT the beginning of flying machines)
“On the contrary, the human mind is a surprisingly efficient and even elegant system that operates with small amounts of information; it seeks not to infer brute correlations among data points but to create explanations.”
We could argue that the human mind CAN (in very specific cases, under some circumstances) be capable of rational processes. But in general, human minds are not trying to “understand” the world around creating explanations. Human minds are extremely inefficient, prone to biases, get tired very easily, need to be off 1⁄3 of the time, etc, etc, etc.
“Here’s an example. Suppose you are holding an apple in your hand. Now you let the apple go. You observe the result and say, “The apple falls.” That is a description. A prediction might have been the statement “The apple will fall if I open my hand.” Both are valuable, and both can be correct. But an explanation is something more: It includes not only descriptions and predictions but also counterfactual conjectures like “Any such object would fall,” plus the additional clause “because of the force of gravity” or “because of the curvature of space-time” or whatever. That is a causal explanation: “The apple would not have fallen but for the force of gravity.” That is thinking.”
Anyone who has spent some time discussing with ChatGPT knows that it does have a model of the world and it is capable of causal explanation. It seems Chomsky didn’t test this himself. I can concede that it might not be a very sophisticated model of the world (do most people have a very complex model?), but again, I expect this to improve over time. I think that some of the responses of ChatGPT are very difficult to explain if it is not doing that thing that we generally call thinking
“For this reason, the predictions of machine learning systems will always be superficial and dubious. Because these programs cannot explain the rules of English syntax, for example, they may well predict, incorrectly, that “John is too stubborn to talk to” means that John is so stubborn that he will not talk to someone or other (rather than that he is too stubborn to be reasoned with). Why would a machine learning program predict something so odd? Because it might analogize the pattern it inferred from sentences such as “John ate an apple” and “John ate,” in which the latter does mean that John ate something or other. The program might well predict that because “John is too stubborn to talk to Bill” is similar to “John ate an apple,” “John is too suborn to talk to” should be similar to “John ate.” The correct explanations of language are complicated and cannot be learned just by marinating in big data.”
I don’t find the reference right now, but I remember very clearly that the next version of ChatGPT already surpasses the average humans in things like implicatures. So doubling down and saying that these systems will “always” be superficial and dubious when we “already” have models that are, better than most humans, is again, completely wrong.
I would agree with Chomsky if he were saying: it seems that “this specific chatbot” that some people made is still behind “some” humans in certain aspects of what we call intelligence. But he is claiming much more than that.
I think you should try to formulate your own objections to Chomsky’s position. It could just as well be that you have clear reasons for disagreeing with his arguments here, or that you’re simply objecting on the basis that what he’s saying is different from the LW position. For my part, I actually found that post surprisingly lucid, ignoring the allusions to the idea of a natural grammar for the moment. As Chomsky says, a non-finetuned LLM will mirror the entire linguistic landcape it has been birthed from, and it will just as happily simulate a person arguing that the earth is flat as any other position. And while it can be “”aligned”″ into not committing what the party labels as wrongthink, it can’t be aligned into thinking for itself—it can only ever mimic specific givens. So I think Chomsky is right here—LLMs don’t value knowledge and they aren’t moral agents, and that’s what distinguishes them from humans.
So, why do you disagree?
“These programs have been hailed as the first glimmers on the horizon of artificial _general_ intelligence — that long-prophesied moment when mechanical minds surpass human brains not only quantitatively in terms of processing speed and memory size but also qualitatively in terms of intellectual insight, artistic creativity and every other distinctively human faculty.
That day may come, but its dawn is not yet breaking, contrary to what can be read in hyperbolic headlines and reckoned by injudicious investments”
I think that day has “already” come. Mechanical minds are already surpassing human minds in many aspects: take any subject and tell ChatGPT to write a few paragraphs on it. It might not exhibit be the most lucid and creative of the best of humans, but I am willing to bet that its writing is going to be better than most humans. So, saying that its dawn is not yet breaking seems to me extremely myopic (it’s like saying that thingy that the Wright brothers made is NOT the beginning of flying machines)
“On the contrary, the human mind is a surprisingly efficient and even elegant system that operates with small amounts of information; it seeks not to infer brute correlations among data points but to create explanations.”
We could argue that the human mind CAN (in very specific cases, under some circumstances) be capable of rational processes. But in general, human minds are not trying to “understand” the world around creating explanations. Human minds are extremely inefficient, prone to biases, get tired very easily, need to be off 1⁄3 of the time, etc, etc, etc.
“Here’s an example. Suppose you are holding an apple in your hand. Now you let the apple go. You observe the result and say, “The apple falls.” That is a description. A prediction might have been the statement “The apple will fall if I open my hand.” Both are valuable, and both can be correct. But an explanation is something more: It includes not only descriptions and predictions but also counterfactual conjectures like “Any such object would fall,” plus the additional clause “because of the force of gravity” or “because of the curvature of space-time” or whatever. That is a causal explanation: “The apple would not have fallen but for the force of gravity.” That is thinking.”
Anyone who has spent some time discussing with ChatGPT knows that it does have a model of the world and it is capable of causal explanation. It seems Chomsky didn’t test this himself. I can concede that it might not be a very sophisticated model of the world (do most people have a very complex model?), but again, I expect this to improve over time. I think that some of the responses of ChatGPT are very difficult to explain if it is not doing that thing that we generally call thinking
“For this reason, the predictions of machine learning systems will always be superficial and dubious. Because these programs cannot explain the rules of English syntax, for example, they may well predict, incorrectly, that “John is too stubborn to talk to” means that John is so stubborn that he will not talk to someone or other (rather than that he is too stubborn to be reasoned with). Why would a machine learning program predict something so odd? Because it might analogize the pattern it inferred from sentences such as “John ate an apple” and “John ate,” in which the latter does mean that John ate something or other. The program might well predict that because “John is too stubborn to talk to Bill” is similar to “John ate an apple,” “John is too suborn to talk to” should be similar to “John ate.” The correct explanations of language are complicated and cannot be learned just by marinating in big data.”
I don’t find the reference right now, but I remember very clearly that the next version of ChatGPT already surpasses the average humans in things like implicatures. So doubling down and saying that these systems will “always” be superficial and dubious when we “already” have models that are, better than most humans, is again, completely wrong.
I would agree with Chomsky if he were saying: it seems that “this specific chatbot” that some people made is still behind “some” humans in certain aspects of what we call intelligence. But he is claiming much more than that.