Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

When Dawkins met Claude – Could this AI be conscious?

unherd.com

Post reply on HN
Print view View original

371–380 of 450 posts

Re: When Dawkins met Claude – Could this AI be conscious?

#371

Post by gizajob »

threethirtytwo wrote: post #363
gizajob wrote: post #361

Earlier quoted context omitted.

Like I said, copy paste all this into your calendar on 3rd of May 2046 and get back to me.

Why don't you PRESENT your reasoning AND Evidence rather then just tell me to wait for 2046. I don't have the patience to wait that long. If you're a rational person you should have evidence/reasoning for why you feel my point is invalid mr. philosopher.

Feed all this into your LLM. I’m not here to type just for your benefit.

And going back to my first point, you seem to believe you’re at some kind of innovative cutting edge when your comments about the lack of proof in human subjectivity show you’re quite a long way from contemporary currents within epistemology which is why I had such lols from it - you accused Wittgenstein of being outdated while expressing a belief that would have been state of the art in about 1700.

The rest of what you’ve exasperated about is fairly scattergun so would take too much typing to engage with.

Re: When Dawkins met Claude – Could this AI be conscious?

#372

Post by mrandish »

boxed wrote: post #123
tracerbulletx wrote: post #49

We don't even know what the pre-requisites for consciousness are so we have no way of knowing. LLMs have emergent behavior that is reminiscent of language forming brains, but they're also missing a lot of properties that are probably necessary? Mainly continuity over time, more integrated memory, and a better sense of space and time? Brains use the rhythm and timing of neuronal firings, and the length of axons effect…

> We don't even know what the pre-requisites for consciousness are so we have no way of knowing. Imo we don't even have a definition of the word that we agree on.

> we don't even have a definition of the word that we agree on.

Indeed, for any in-depth discussion of LLMs and consciousness to be productive, clearly defining terms and scope is essential. The Stanford Encyclopedia of Philosophy is an excellent resource: https://plato.stanford.edu/entries/consciousness/

Re: When Dawkins met Claude – Could this AI be conscious?

#373

Post by gizajob »

mrandish wrote: post #366
threethirtytwo wrote: post #253

Earlier quoted context omitted.

The problem with your thinking here is that we are creating artificial beings now that display and output the same subjectivity. The argument you present like many arguments breaks down when the topic becomes self referential. It makes sense for other topics as analyzing subjectivity becomes pedantic when asking questions like why is the sky blue. But now subjectivity itself is in question. The argument you present c…

> what makes humans rise to the level where we can assume their subjectivity is true? To dive into this specific question: to me, there's a better reason than the obvious functional utility of not treating other humans like NPCs. It's in three parts. First, is that I subjectively experience a rich and varied internal mental life (aka qualia). So, I have first-hand evidence that N equals (at least) 1 in terms of quali…

Thank you for this clear and correct answer structured for those who need data included from science rather than from philosophy alone.

Re: When Dawkins met Claude – Could this AI be conscious?

#374

Post by qnleigh »

qnleigh wrote: post #155

It's easy, and very tempting to dismiss this sort of thing. But given how little we know about the human brain, let alone consciousness, I don't see how we can be confident that LLMs aren't conscious. I've had a lot of thoughts and conversations over the years that changed my mind on what consciousness likely requires. One was the realization that a purely mechanical computer can, in principle simulate the laws of ph…

A lot of the comments in this thread point out that LLMs could be very good at tricking us into thinking that they are conscious because they operate on language; our brains are primed to empathize with what they say and imagine a being operating behind them. This is a really good point.

I'll even take it a step further; most of an LLM's training is next-token prediction on random internet content. A newly-trained LLM will just continue whatever text appears in its context window, like an extremely capable autocomplete. The illusion of an entity that takes turns in conversation and presents a consistent personality is tacked on at the last minute through RLHF. This was the transition from GPT to ChatGPT.

Any positive evidence of LLM consciousness should probably mostly be taken from the model before post-training, where it displays remarkable capabilities but shows no sign of a consistent personality, and likely no signs of self-awareness or self-understanding.

Re: When Dawkins met Claude – Could this AI be conscious?

#375

Post by qnleigh »

RaftPeople wrote: post #327
qnleigh wrote: post #287

Earlier quoted context omitted.

Well I'm not saying that LLMs are conscious; I'm just saying that I'm not super-confident either way. To flesh this out a bit more, I agree that ability to communicate is not enough (ELIZA probably didn't pass the bar, even if it did kinda pass a Turing test). But that's also not what gives me pause with LLMs. It's how much information processing they seem to be doing under the hood. It's really hard to imagine how n…

> but I find it almost as hard to see why evolution did. Ignoring the concept of consciousness, it seems that self-awareness would be a strong attribute related to survival. It seems like it would help drive or amplify critical emotional states (e.g. my own survival, competition/success, love for self and relatives, etc.) I can't see anywhere in the LLM machinery that would support the notion of self awareness in adv…

Yeah self-awareness is a very different thing, and I agree it's easier to see how evolution would produce this. Many apparent signs of self-awareness in LLMs are probably baked into the models at end via post-training (RLHF), where they learn to behave as conversation agents and maintain a more consistent personality. The raw model probably shows no signs of self-awareness. In fact, I'm pretty sure that LLMs learn that they are LLMs only through post-training.

Re: When Dawkins met Claude – Could this AI be conscious?

#376

Post by mrec »

tovej wrote: post #369
mrec wrote: post #358

Earlier quoted context omitted.

> Biological gender is the inconsistent "sex=gender" conception that cranks and conservative grifters operate under. I don't know about cranks and conservative grifters, but it's definitely not a feature of the "gender critical" position which I thought Dawkins was broadly aligned with. That's more that "sex" is absolutely binary, with your "intersex people" being umambiguously classified through genetics, while "gen…

The rhetorical claim "there are only two sexes" is only ever used to claim that there are two genders. And that is precisely what Dawkins has done. You are wrong about intersex people being genetically classifiable. There is no deterministic causal relationship between genetics and sex characteristics: as an example, a person with XX chromosomes may develop external male genitalia and v.v. for XY chromosomes. But of…

Your first para seems utterly bizarre, to the point of nonsensicality. Maybe it would help if you could say what you think "gender" refers to, but at this point I doubt it.

Your second para's argument would only be valid if you thought that sex is defined by external characteristics. I'm pretty sure you don't think that. And as far as I'm aware, while some DSDs certainly have a gene-expression component, there's no reason to think that they don't all ultimately have a genetic basis. There's a strong whiff here of "it's all terribly complicated so let's just agree that nothing means anything".

Obviously people can disagree about the merits or otherwise of a genetic classification. But it's not straightforwardly wrong or insane, particularly since credible alternatives have been notably lacking.

Re: When Dawkins met Claude – Could this AI be conscious?

#377

Post by kbelder »

jhbadger wrote: post #254
laichzeit0 wrote: post #189

Earlier quoted context omitted.

You could push the analogy even further and run the thought experiment where every forward pass through an LLM could in principle be done on pen and paper, distributed throughout all humanity. Sure it would take a long time, but the output would be exactly the same. We’ve just shifted the implementation from GPU to scribbling things down on paper. If you want to assert that LLMs are “conscious” then you would have to…

When we get complete neuronal connection maps (which we are getting close to for mice and humans will be done within a decade or two), we could in principle simulate a brain on a computer or on paper too. Unless you assert something magical like a "soul", these connections are what determine human consciousness. It is one thing to argue that LLMs don't resemble brains and if they could be "conscious" they wouldn't be…

While I think you're right in principle, there's a lot of reason to think the structure of a brain is more complicated than just the connection map of the neural network. There is a lot of complicated behavior inside each neuron that we don't fully understand yet. They aren't just logic gates.

But, of course, that's just physics. It's not magic, so your point stands.

Re: When Dawkins met Claude – Could this AI be conscious?

#378

Post by qnleigh »

almostjazz wrote: post #257
ganymedes wrote: post #229

To those that genuinely consider LLMs could be conscious... Remember that other mammals do not have language and ability to think in language, but they likely do "think" in visual imagery and are able to navigate their landscape, breed, get food and have varying levels of problem solving intelligence that have been studied in labs. Human brain has evolved on top of that, the language is a layer above that supercharge…

Of course the language abilities of LLMs is not proof of consciousness at all. If some alien entity made a model that was truly just 10^1000 hard-coded if-statements to respond to every possible question, it might seem way better than our best models now but would obviously not be conscious. The problem is just that even in the most lousy, turing test-failing LLM there's no guarantee that not a single subsection of t…

> a model that was truly just 10^1000 hard-coded if-statements to respond to every possible question

That's a really compelling argument against the Turing Test. But in order to build such a machine, you would need an enormous amount of compute to populate the answers. The interesting question is then whether consciousness emerged while doing all that pre-compute.

Re: When Dawkins met Claude – Could this AI be conscious?

#379

Post by xyzsparetimexyz »

energy123 wrote: post #281

Dawkins' point is more subtle than he's getting credit for. > If Claudia is unconscious, her behaviour shows that an unconscious zombie could survive without consciousness. Why wasn’t natural selection content to evolve competent zombies? An excellent and original question. If intelligence is decoupled from consciousness, why did natural selection evolve consciousness?

I believe that consciousness works as a layer of systems. The lowest systems react fastest but to a small, well known set of stimuli. More unexpected stimuli trigger higher layers until really unexpected things trigger the slowest, full conscious layer for deep decision making. Hence why you can feel like you're on autopilot when you're doing something you're experienced in such as driving. It's resllt a system designed to cope with a world that changes too fast to get engrained in our genes.

Also read Blindsight

Re: When Dawkins met Claude – Could this AI be conscious?

#380

Post by mrandish »

amanaplanacanal wrote: post #285
boxed wrote: post #250

Earlier quoted context omitted.

First of all: arguing about the details of a thing that actually exists is an enormous difference from arguing details of a thing that does NOT exist. As for your dig at Dawkins, I just read https://archive.ph/Rq5bw which I assume you're referring to. Notice how he never defined "conscious" and he seems to use it as equivalent to "can process data logically" which is not at all how I would define the word. And if you…

I honestly don't see how Dawkins is so confused. Claude says it can't tell if it has any kind of inner life. Can you imagine a human saying that?

> Claude says it can't tell if it has any kind of inner life.

I don't see how some people apparently believe the text output of an LLM about it's internal mental state is anything other than a plausible fabrication based on what its training data already says about the mental states of LLMs. These are systems specifically designed and iteratively optimized over millions of training generations to generate text output which plausibly simulates what a composite human would say in response to the same input. There is no human-like internal mental state it can reflect on, so all such responses are, by definition, plausible hallucinations based on interpolated training data.

> Can you imagine a human saying that?

Some people do say that: see Aphantasia and, specifically, Anauralia https://en.wikipedia.org/wiki/Aphantasia

Post reply on HN