The American competitiveness angle is definitely… well an interesting choice; I also was taken aback by that. Ignoring some of the posturing along those lines I find the post itself interesting.
If you think the Chinese room idea is valid we probably just fundamentally disagree and that’s OK.
Fundamentally disagree that an LLM is nothing more and nothing less than a very advanced next-word-prediction algorithm, you know, the scientific basis of an LLM?
I was using “Chinese Room” to indicate that it’s a black box, and even the developers are incapable of pinpointing the statistical basis behind any given response, despite it being a stochastically-generated choice from an ultimately deterministic algorithm.
I.e. are you claiming that LLMs have attained sapience as a purely emergent property?
I claim ignorance as to whether they have or not, but I am fully on board with intelligence as an emergent property and believe that there is no reason a machine can’t think, yes.
Understand: I agree wholeheartedly with the idea of intelligence as an emergent property (or I wouldn’t have brought it up), and don’t think it impossible to achieve AGI. I do contend that this route, of nothing but token input without true correlative context, is a non-starter for such. Making something better and better at predicting what comes next is not going to suddenly create a jump to application in unrelated fields or grant context. Statistical correlation is insufficient to generate actual meaning, no matter how many times you provide feedback on the inference. It doesn’t have any actual connection to anything other than text (or hex codes, or other such tokens). If constant hallucinations (demonstrating the lack of any understanding of context) are insufficient to prove this to you, I suppose there is a fundamental difference in the way you understand how LLMs function.
I understand fairly well how LLMs work, but I don’t think we understand the conditions for the emergence of intelligence or thought enough to conclude that “learning to predict what comes next” is not sufficient to create some form of intelligence, that’s all. [Edit: if we determine that that turns out to be sufficient I will be surprised] That uncertainty is sufficient to make me agnostic on the subject and that stance does have an effect on my actions, for example I’m never mean to an LLM.
I’m only mean to LLMs when they’re being used to pose online as a troll. Given that that’s the only time I ever interact with an LLM anymore, though…
Either way, emerging subintelligence or not, they’re being used in ways such that they are, all of them, a scourge on the planet and on humans’ abilities to think for themselves, and should be eradicated and banned for those reasons alone.
The American competitiveness angle is definitely… well an interesting choice; I also was taken aback by that. Ignoring some of the posturing along those lines I find the post itself interesting.
If you think the Chinese room idea is valid we probably just fundamentally disagree and that’s OK.
Fundamentally disagree that an LLM is nothing more and nothing less than a very advanced next-word-prediction algorithm, you know, the scientific basis of an LLM?
I was using “Chinese Room” to indicate that it’s a black box, and even the developers are incapable of pinpointing the statistical basis behind any given response, despite it being a stochastically-generated choice from an ultimately deterministic algorithm.
I.e. are you claiming that LLMs have attained sapience as a purely emergent property?
I claim ignorance as to whether they have or not, but I am fully on board with intelligence as an emergent property and believe that there is no reason a machine can’t think, yes.
Understand: I agree wholeheartedly with the idea of intelligence as an emergent property (or I wouldn’t have brought it up), and don’t think it impossible to achieve AGI. I do contend that this route, of nothing but token input without true correlative context, is a non-starter for such. Making something better and better at predicting what comes next is not going to suddenly create a jump to application in unrelated fields or grant context. Statistical correlation is insufficient to generate actual meaning, no matter how many times you provide feedback on the inference. It doesn’t have any actual connection to anything other than text (or hex codes, or other such tokens). If constant hallucinations (demonstrating the lack of any understanding of context) are insufficient to prove this to you, I suppose there is a fundamental difference in the way you understand how LLMs function.
I understand fairly well how LLMs work, but I don’t think we understand the conditions for the emergence of intelligence or thought enough to conclude that “learning to predict what comes next” is not sufficient to create some form of intelligence, that’s all. [Edit: if we determine that that turns out to be sufficient I will be surprised] That uncertainty is sufficient to make me agnostic on the subject and that stance does have an effect on my actions, for example I’m never mean to an LLM.
I’m only mean to LLMs when they’re being used to pose online as a troll. Given that that’s the only time I ever interact with an LLM anymore, though…
Either way, emerging subintelligence or not, they’re being used in ways such that they are, all of them, a scourge on the planet and on humans’ abilities to think for themselves, and should be eradicated and banned for those reasons alone.
I agree that they pose serious immediate issues regardless of thinking or not