3 ms·

> because they don’t have knowledge like humans do, they rely on pattern-matching, not understanding. I don't think this comparison makes sense - humans do not

> because they don’t have knowledge like humans do, they rely on pattern-matching, not understanding. I don't think this comparison makes sense - humans do not have knowledge as a unique thing differentiating us from LLMs. "Knowledge" is just the ability to produce relevant/"correct" results in some interaction, and some of us have acquired experiences that involve dealing with a particular subject and other knowledgable people within that subject for years or even decades, granting us (very fallible!) "knowledge" within that specific area. Humans hallucinate answers, even within their area of knowledge. Our memory is imperfect, so we're creating an answer based on bits and pieces, while anything missing is subconsciously glossed over and filled in with something plausible. It's just that the more experiences you have in an area, the better your hallucination. Those weird moments when your parents assure you of some fact that challenge the limits of just how wrong something can be would then be an example of hallucination in the extreme polar opposite end. (Note that I am not implying that the human brain works like an LLM, but rather just challenge the concept of "knowledge" being fundamenetally different from LLM behavior.)

igouy 1y ago
> Humans hallucinate answers Confabulate.
- arghwhat 1y ago
  That does appear to be a better term for both humans and LLMs. Not sure who coined the term "hallucination" in the context of LLMs, which that suggests problems with perception rather than memory...
computerthings 1y ago
[dead]
marginalia_nu 1y ago
If an AI correctly guesses the outcome of a dice roll, it does not mean it knows the outcome in advance. Being correct is a necessary element of knowledge, but far from sufficient. It's probably helpful to reach for epistemology to make sense of knowledge as a concept. While something like "true justified belief" is far from a perfect definition of knowledge, it a least is at least correct to a good approximation.
- arghwhat 1y ago
  > If an AI correctly guesses the outcome of a dice roll, it does not mean it knows the outcome in advance. Being correct is a necessary element of knowledge, but far from sufficient. If we "know in advance", why wouldn't the AI? If the AI doesn't "know in advance", why would we? When I'm asked a question, the response and thoughts around it emerge only at that point, and from thin air. I certainly don't trawl an internal, constant knowledge bank of retrievable facts, and the time of day or preexisting conversation may affect whether or not I successfully "know" something in a particular conversation, despite having learnt it. At the same time, some things I "know" will just be wrong - the memory distorted or mixed with another, the whole thing imagined, or even just a failed recollection. And no, I certainly don't know if it has gone wrong unless it happens to result in e.g., contradictions I notice myself. These kinds of justifications for why AI doesn't "know" always imply that the process of human thought emergence is well established and understood and easily differentiated, which just isn't the case. There most certainly are more differences than similarities between AI and the human mind, but I dont think the concept of knowledge is necessarily one of them.

igouy 1y ago

> Humans hallucinate answers Confabulate.

arghwhat 1y ago
That does appear to be a better term for both humans and LLMs. Not sure who coined the term "hallucination" in the context of LLMs, which that suggests problems with perception rather than memory...

computerthings 1y ago

[dead]

marginalia_nu 1y ago

If an AI correctly guesses the outcome of a dice roll, it does not mean it knows the outcome in advance. Being correct is a necessary element of knowledge, but far from sufficient. It's probably helpful to reach for epistemology to make sense of knowledge as a concept. While something like "true justified belief" is far from a perfect definition of knowledge, it a least is at least correct to a good approximation.

arghwhat 1y ago
> If an AI correctly guesses the outcome of a dice roll, it does not mean it knows the outcome in advance. Being correct is a necessary element of knowledge, but far from sufficient. If we "know in advance", why wouldn't the AI? If the AI doesn't "know in advance", why would we? When I'm asked a question, the response and thoughts around it emerge only at that point, and from thin air. I certainly don't trawl an internal, constant knowledge bank of retrievable facts, and the time of day or preexisting conversation may affect whether or not I successfully "know" something in a particular conversation, despite having learnt it. At the same time, some things I "know" will just be wrong - the memory distorted or mixed with another, the whole thing imagined, or even just a failed recollection. And no, I certainly don't know if it has gone wrong unless it happens to result in e.g., contradictions I notice myself. These kinds of justifications for why AI doesn't "know" always imply that the process of human thought emergence is well established and understood and easily differentiated, which just isn't the case. There most certainly are more differences than similarities between AI and the human mind, but I dont think the concept of knowledge is necessarily one of them.