Comment by DonaldFisk | Hacker News Reader

Comment by DonaldFisk | Hacker News Reader

Parent Full thread

I think that recent LLMs do pass the Turing test, but I don't think it is a good test for intelligence, for two reasons: emergent complexity and pareidolia.

Arguably, ELIZA passed the Turing test ca. 1967: https://archive.computerhistory.org/resources/access/text/20... and not that long ago, ELIZA was still more convincing: https://arstechnica.com/information-technology/2023/12/real-...

The Turing Test can be gamed. This was done in some Loebner Prize entries: see https://www.behavior.org/resources/325.pdf and https://en.wikipedia.org/wiki/Eugene_Goostman

To determine whether a machine is intelligent, I think it's necessary to understand how it works internally and not just take a black-box approach. This was already understood when Alan Turing wrote his 1950 paper. Elmer and Elsie, William Grey Walter's tortoise robots built during the preceding two years, appeared to behave as if they had some degree of self-awareness. But their "brains" consisted of only two thermionic valves, so self-awareness can be ruled out.

I don't know whether LLMs pass the test as Turing originally worded it, but I always interpreted the success condition to be "is indistinguishable from a human" rather than "can trick someone into thinking it's human". Anthropomorphization is a human weakness, and exploiting that doesn't make something intelligent.

I'd only consider an AI to pass the Turing Test if an expert in how the AI works who prepares in advance for the test is unable to tell it apart from a human. The tester should even be allowed to use automated tools as part of the testing process.

If something is indistinguishable from a human in its outward behavior, then I'd be forced to admit that it's as intelligent/conscious/sentient/whatever as a human, regardless of how it works internally. It's not like I make my friends take an MRI before I decide whether they're intelligent.

> If something is indistinguishable from a human in its outward behavior, then I'd be forced to admit that it's as intelligent/conscious/sentient/whatever as a human, regardless of how it works internally. It's not like I make my friends take an MRI before I decide whether they're intelligent.

You can be reasonably sure about your friends, as they're sufficiently like you to leave little doubt. I disagree but can understand you thinking that LLMs are intelligent, but unless by conscious you mean something like being self-aware, I don't understand how you could conclude LLMs are conscious. For example, although LLMs can tell you what colours things are, how the visual system distinguishes colours in humans and other animals, and what wavelength mixtures different colours are, they don't know what it's like to see colours because they cannot see colours. All the input of LLMs is text, and you can't describe, in words, what it's like to see colours, or hear sounds, or feel pain, in a way that someone incapable of these sensations would understand. And to remove any remaining doubt, there was an expert in colour vision (Knut Nordby) who, as he could not see colours because his retinas lacked cones, could not imagine what it's like to see colours, despite all his knowledge.

I have no way of knowing whether other people are conscious except by observing their behavior. Similarly, I have no way of knowing whether a machine is conscious except by observing its behavior. If something is outwardly indistinguishable from an intelligent/conscious/sentient/whatever being and you claim that it's still not i/c/s/w, then you've implicitly accepted the possibility that all humans besides you are not i/c/s/w either.

Your tangent on color vision/hearing/pain is not relevant. None of those things are required for i/c/s/w, nor is it obvious that a machine is incapable of feeling those things.

Also you keep using the term "LLM" a bunch here but I wasn't talking about LLMs. The question of machine intelligence is much broader than whatever the current state of the art happens to be.

I initially responded to CSMastermind whose post explicitly mentioned LLMs and the Turing test, which involves interacting with a computer solely by typing and reading text. You first mentioned consciousness.

We can tell that other people are intelligent and conscious because they're like us in important ways and we are intelligent and conscious. I haven't ruled out intelligent machines in general. My point was about the inadequacy of the Turing Test.

Reply on news.ycombinator.com