Live data from Hacker News

When Dawkins met Claude – Could this AI be conscious?

unherd.com

421–430 of 450 posts

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

#421
post #412

Earlier quoted context omitted.

Life can be “switched off” - death is the ultimate off. Some life can be frozen and unfrozen with no ill effects. And life itself doesn’t mean consciousness. And ultimately what is life? Something that has biological processes and reproduces? Why can’t we replace or recreate these processes with manmade equivalents to get the same results?

I don’t know what you are saying here? That a program running on a GPU in a data center is alive because it can predict what words to say next? I find it strange that people are quite often unwilling to see animals as conscious yet here we are discussing if an empheral computer programme is. Have a think on why we don’t eat people but eat chickens - chickens are clearly more conscious than any AI, yet still not enoug…

I don’t think you are having the same conversation as the rest of us here. Think on what the definition of consciousness is and how we will prove that a machine has it or not. I do not think that the current versions of AI are conscious but that doesn’t mean we will never achieve this and face the same questions of how to prove it.

Cannibalism has nothing to do with it as some seemingly conscious humans have shown this behaviour in the past.

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

#422
post #417

Earlier quoted context omitted.

Overall, I like the direction and curiosity. I will answer general points as they occur to me. "Ultimately" is doing a lot of work here. It is hardly a given, but assuming it is true allows you to smuggle a conclusion in. I see what you did:D But lets go with that assumption for rebuttal below. Again.. hardly a given and assumes what it intends to prove. It seems you have a bias for a specific outcome. Not exactly a…

First of all, thanks for responding. >It is hardly a given.. Ok, fair enough. So let us also seed the simulation with all the random events from our universe. That should cover it, right? >A mathematical description of a fire does not burn anything.. This is the self evident dogma that we have to overcome to understand this idea. You say that a mathematical description of a fire does not burn anything. But what if bo…

It is more of an argument. If the mathematical description somehow created fire, it would have been closer to an actual spell, but it doesn't, which would suggest that description is not accurate or the argument is flawed ( edit: or both ). The flaw I noted puts both your and my argument in a difficult place, because it, among other things, exposes your surety about 'ultimately'.

I am engaging with you, because, while I think you are wrong, I don't want to pour cold water on an inquisitive mind ( and some of the thoughts you listed are interesting to explore ) -- I just also happen to think you got too mesmerized by the novelty of the idea.

FWIW, I am just guy on the internet so don't take my word for it.

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

#423
post #156

Earlier quoted context omitted.

LLMs are word prediction engines. They clearly are not conscious, they are just guessing what words should come next.

> They clearly are not conscious Consciousness is emergent. A human is not conscious by our definition until the moment they are. How will we be able to identify the singularity when it comes? I feel like this is what the article is really addressing. > LLMs are word prediction engines Humans can also do this too, so what are the missing parts for consciousness? Close a few loops on learning pipeline and we might be…

[dead]

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

#424

Earlier quoted context omitted.

1. Human consciousness is characterized by intentionality and aboutness . This aboutness has semantic content. We know LLMs lack semantic content, and we know this because computers are purely syntactic simulators. This is definitional. It’s not some mystery. Furthermore, unless you’re a Cartesian dualist, I don’t know of anyone who denies non-human animals consciousness. The consciousness of non-human animals differ…

Thanks for responding to my fuzzy, Sunday morning though-dump. Overall, I'd say that you seem to be leaning pretty heavily on your own definitions. I suspect that other definitions are available. > Human consciousness is characterized by intentionality and aboutness. This aboutness has semantic content. We know LLMs lack semantic content, and we know this because computers are purely syntactic simulators. So tell me…

> Overall, I'd say that you seem to be leaning pretty heavily on your own definitions.

This is just what intelligence involves if intelligence is to be intelligence. The terminology is common in the philosophy of mind. The attack - that these are just my definitions - is also fallacious. It doesn’t respond substantively to them.

> So tell me how human brains create this "aboutness" and "semantic content"?

The burden of proof is not mine. I am pinning down the thing that must be explained. How can you coherently deny aboutness and semantic content as central to intelligence? That is exactly what intelligence entails if it is to be intelligence, which is the capacity to apprehend reality. Otherwise, what are we even debating?

(FWIW, we can strongly argue that brains alone cannot account for concepts, because concepts are universal while matter is concrete and particular. We can have the concept Triangularity and talk about what is true of all triangles by analyzing the concept, but you never encounter Triangularity in the wild, as it were. You only ever witness particular triangles, each determined in ways that Triangularity is not. Triangularity is abstracted from them. So if matter is concrete and determined and can only be concrete and determined, and what a thing is is the substance of the content we apprehend - that must be the case, or else you deny the possibility of knowledge and render your claims incoherent - then these abstracted universal concepts cannot exist as such in matter without losing their universality, destroying knowledge in the process.)

> why can't sufficiently complex logic operations achieve the same outcome?

Why would they? This is magical thinking. These “logic operations” in question are purely syntactic operations - this is just definitional. Computational models are formalizations of effective methods, which are operationally mechanical and syntactic. You cannot pile on more syntax or more fancy looking syntax to get semantics. That makes no sense.

> Using brains to try to understand the emergent properties of brains seeme fraught with conceptual problems

Why? If a “brain” (let’s say mind) can understand reality at all, then why should brains escape the domain of things it can understand? Conversely, if a brain cannot grasp the brain, why should it be able to grasp anything at all? Is not a matter of complexity, because knowledge is about grasping the principles governing a thing, not holding some vast state diagram in your head.

> a potential to over-estimate the centrality of how it subjectively feels to be a brain. In a way it's understandable as, in the end, that feeling is all we have.

This has nothing to do with feelings, though the subjectivity of consciousness (again, intentional in nature) is a fact you cannot ignore and must account for. Otherwise, you are guilty of eliminativism, that is, the denial of the very facts that must be explained, because of a prior commitment to materialistic presuppositions.

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

#425
post #393

Earlier quoted context omitted.

"Seems ridiculous" is a very subjective thing, though, and very dependent on context. It can seem ridiculous that a boxer with male physical advantage since puberty (i.e. 5-ARD) can beat the ** out of a female boxer while the world's media looks on and applauds, but here we are. Personally I'm sympathetic to the idea that CAIS individuals should be a reasonable exception, i.e. they're still biologically male, but in…

The problem with the genetic definition is that it means that nobody knows whether anyone is a man or a woman until they get tested. Nobody pre-genetic testing ever knew. Most people alive today don't know. And that's clearly not how society works, or ever has worked.

I think we're in severe danger of spiralling toward epistemology here, but there's a huge difference between "nobody knows" in the sense you're using it and "nobody has any idea". Society (or more broadly biology) doesn't need us to get it right 100% of the time; these are very, very rare conditions. Going off secondary sexual characteristics is going to get you the right answer as near as dammit every time, especially in premodern contexts, and society has always been happy to work with tiny or even not-so-tiny uncertainties. (Is that kid really mine? Is the accused really guilty?)

The way I see it, the sex binary is fundamentally about reproduction. It's why we can use the same concept for everything from pondweed to platypuses. All across nature, male=small gametes, female=large gametes. In humans that's driven (with the potential for things to go haywire occasionally, sure, but still driven) by the XX/XY system, so that strikes me as a reasonable thing to base a definition on.

Side note re "nobody knows whether anyone is a man or a woman until they get tested" - I'd say that giving birth or fathering a child is a pretty big clue. AFAIK the only cases where that doesn't line up perfectly with genetic sex relate to mosaicism, where I'll freely admit my intuition goes completely kablooie.

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

#426
post #417

Earlier quoted context omitted.

First of all, thanks for responding. >It is hardly a given.. Ok, fair enough. So let us also seed the simulation with all the random events from our universe. That should cover it, right? >A mathematical description of a fire does not burn anything.. This is the self evident dogma that we have to overcome to understand this idea. You say that a mathematical description of a fire does not burn anything. But what if bo…

It is more of an argument. If the mathematical description somehow created fire, it would have been closer to an actual spell, but it doesn't, which would suggest that description is not accurate or the argument is flawed ( edit: or both ). The flaw I noted puts both your and my argument in a difficult place, because it, among other things, exposes your surety about 'ultimately'. I am engaging with you, because, whil…

> I don't want to pour cold water on an inquisitive mind..

Thanks, I appreciate that.

> If the mathematical description somehow created fire..

The mathematical description does not create a fire. It describes a consciousness that is observing a fire.

I am not elaborating so as to not muddle the above point.

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

#427
post #346

Earlier quoted context omitted.

Except that latent space does not change in response to new information, something that thoughts famously do. If you read a book that captures the author's thoughts, disagree, and write an eloquent arguments to the author, you might change the author's mind. But you will not change the "book's thoughts" on the subject. Latent spaces are maps of thoughts other people have had, not the thoughts themselves.

This gets a bit tricky. Over very long task contexts (1M tokens) or with prompt compression (10s of millions of tokens) the model can alter its priors based on updated evidence. This form of knowledge based learning is not necessarily robust, but demonstrably does occur.

"the model can alter its priors"

The model doesn't have high-level priors in the Bayesian sense (though you could have priors about it).

The low-level priors it does have (the weights) are not modified by the context.

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

#428
post #416

Earlier quoted context omitted.

If the notion of consciousness they're referring to doesn't meet the normal philosophical criteria, then they're essentially just wrong. Which is quite possible - many people seem very confused on the subject, which is not too surprising, especially for scientists who essentially reject philosophy, like Dawkins.

Philosophy doesn't own words afaik. Words have different meaning in different domains. > many people seem very confused on the subject, which is not too surprising, especially for scientists who essentially reject philosophy, like Dawkins Or they just use words in a different domain and you didn't notice and now you're angry because what they said didn't make sense. Come now, surely philosophy must handle such trivia…

The philosophical meaning of consciousness is usually what people are referring to when they talk about consciousness. Can you point to some other commonly used "different meaning" that's being used or intended in these discussions?

> I'm gonna have to reject philosophy too

You'll have lots of intellectually stunted company.

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

#429
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?

Sure if that’s what you are brought up to say. If you want a real life example you have kids that were isolated and never taught to speak. They probably worldn’t even understand the question

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

#430
post #349

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…

Maybe Roger Penrose has some meaningful Input https://youtu.be/iTVN6tFknCg regarding …“One was the realization that a purely mechanical computer can, in principle simulate the laws …“ As far as I unterstood,there is no theory of quantum gravity and therefore this is not being simulated on a computer. I think he makes other arguments. So you cannot say for sure that you can simulate a human brain on a computer

This is mainly god of the gaps style argument.
Post reply on HN