A friend of mine drew my attention on a conversation Richard Dawkins has had with ChatGPT on the topic of AI consciousness. — Pierre-Normand

I heard about Dawkins' article through Jerry Coyne's website:

https://whyevolutionistrue.com/2025/02/18/richard-dawkins-chats-with-ai-about-consciousness/

I left a comment there, referring readers of Coyne's website to a couple of your threads here. (No idea whether anyone from there might have followed the links, but Dawkins himself made several comments there.)

You didn't reply to me but since you attacked me and my knowledge then I challenge you! — MoK

I'm not surprised that you interpreted my comments as an attack, but no one can be an expert on everything. So I'd say it is more like I pointed out that you are human and your misconceptions are understandable.

Couldn't you wonder that it could be you who doesn't have the proper knowledge to comprehend the MoK's argument? — MoK

Sure I can wonder, but you demonstrate throughout this thread that you don't have much understanding of phyisical causality. I, on the other hand, am a 62 year old electrical engineer making my living on the basis of my expertise in understanding physical causality.

Can you provide any reason for me to think that your intuitions regarding this topic are better than mine?

The reality is that you simply can't imagine how physicalism could account for awareness and m-experience. You're committing the fallacy argument from incredulity, also referred to as "argument from lack of imagination". — Relativist

It seem worth noting that a scientifically informed physicalism explains MoK's incredulity.

With the understanding that MoK's intuitions are a function of the training of the neural networks in MoK's brain, and that MoK clearly hasn't done any deep investigation into physical causality, it is unsurprising that MoK's intutions result in incredulity as they do.

Humans cannot make objective judgments, and subjective judgements are meaninglesss — RussellA

It seems to me that subjective human judgements can be quite meaningful to humans. For example if someone's society judges them to not be fit to participate in that society and subsequently banishes or imprisons that person, I'd expect that person to find society's judgement to be meaningful.

So I'm not sure what you mean by "meaningless" in the quote above.

P2) Experience is due to the existence of physical and the change in the state of physical is due to the existence of an experience — MoK

Why think that all physical changes are due to experience? Consider the possibility that astronomers today observe a supernova which occurred a billion years ago in a distant galaxy. What role did experience play in causing the supernova?

You bolded that portion yourself in your
, I simply formatted the quote in order to respect that. Because unlike you, I am indeed being charitable towards your intentions. e I will sense the
— previous comment — Arcane Sandwich

I don't know what you are trying to say there, or who you are suggesting that you were quoting.

By assuming ignorance on my part, you're not willing to give me a fair reading as your interlocutor. — Arcane Sandwich

I provided you with an opportunity to show that you weren't ignorant in relevant ways with my first response to you. Unfortunately it seems that you weren't able to take advantage of the opportunity.

If you accuse me of strawmannig, then you're accusing me of charlatanry, hence sophistry, and therefore you are assuming ill intent on my behalf... — Arcane Sandwich

No I'm not assuming ill intent. Ignorance on your part seems a simple enough explanation.

False. You do not sense the force of attraction in that case, you simply feel an increasingly solid sensation, in a tactile sense. — Arcane Sandwich

The bolded portion seems an odd way of expressing whatever you may be trying to express. Have you actually done the experiment?

In any case, yes I have a tactile sensation of the attraction between the magnet and the iron.

It seems to me it would be more productive for you to actually address my points than to whine to the moderators, but whatever floats your boat.

(AE1) If Empiricism is true, then magnetism can be perceived by human beings. — Arcane Sandwich

1. You seem to be attacking an archaic/straw version of empiricism, by stipulating that some sort of 'direct sensing' of properties must be available to humans for empiricism to stand up to scrutiny.

2. I have many ways of detecting the presence of a magnetic field. A simple one is just to hold a magnet near a piece of iron, in which case I will sense the force of attraction between the magnet and the iron.

The thread became entangled in animal intelligence, a garden path, to my eye.
— Banno
Yes. There are those who cannot conceive of a non-human animal that truly shares any concepts with human beings and those who are quite sure that all animals in this world share that world, to a greater or lesser extent. Never the twain shall meet. Looks like two incommensurable conceptual schemes to me. — Ludwig V

To further entangle the thread with animal intelligence...

https://www.sciencedaily.com/releases/2025/02/250203163756.htm

To get treats, apes eagerly pointed them out to humans who didn't know where they were, a seemingly simple experiment that demonstrated for the first time that apes will communicate unknown information in the name of teamwork. The study also provides the clearest evidence to date that apes can intuit another's ignorance, an ability thought to be uniquely human.

It appears bonobos are capable of sharing our ability to conceive of others as knowledgeable or ignorant of some fact.

Thanks for all your posts. Will come back with more of my replies on the rest of your posts in due course. — Corvus

"in due course"?

At a later time?

If any of these conversations were posted 10 years ago, people would have agreed without hesitation that any hypothetical AI which has achieved conversation in this level represented AGI. Now that it has been achieved pop my it seems the goal posts have shifted. — hypericin

:up:

Don't scientists subscribe to a massive metaphysical commitment, that reality can be understood? — Tom Storm

Is that a massive commitment? It seems to me a matter of rather routine observations.

Another critical point of AI's responses is that, they are predictable within the technological limitations and preprogramming specs. To the new users, they may appear to be intelligent and creative, but from the developers point of view, the whole thing is pre-planned and predicted debugging and simulations. — Corvus

Corvus, you are pretending to understand modern AI when you clearly don't.

See here:

Despite trying to expect surprises, I’m surprised at the things these models can do,” said Ethan Dyer, a computer scientist at Google Research who helped organize the test. It’s surprising because these models supposedly have one directive: to accept a string of text as input and predict what comes next, over and over, based purely on statistics. Computer scientists anticipated that scaling up would boost performance on known tasks, but they didn’t expect the models to suddenly handle so many new, unpredictable ones.

Recent investigations like the one Dyer worked on have revealed that LLMs can produce hundreds of “emergent” abilities — tasks that big models can complete that smaller models can’t, many of which seem to have little to do with analyzing text. They range from multiplication to generating executable computer code to, apparently, decoding movies based on emojis. New analyses suggest that for some tasks and some models, there’s a threshold of complexity beyond which the functionality of the model skyrockets. (They also suggest a dark flip side: As they increase in complexity, some models reveal new biases and inaccuracies in their responses.)

2. **Synchronous Gradient Sharing:**
- After all replicas finish processing their respective mini-batches for the current training step, they share their computed gradients with one another.
- These gradients are **averaged (or summed)** across all replicas. This ensures that the weight update reflects the collective learning from all mini-batches processed during that step. — Pierre-Normand

This is a very interesting aspect of the logistics of LLM training that I was unaware of. It suggests that a move from digital to analog artificial neural nets (for reduced power consumption) may not be forthcoming as soon as I was anticipating.

This is because of the need for all replicas to be functionally identical in order to take advantage of this possibility of training multiple replicas in parallel. While analog hardware could be more energy efficient, analog 'replicas' might not be able to be made sufficiently identical to take advantage of such batch learning.

As always, fascinating conversation.

:100: :up:

Taiwan is not important enough to US national interests to risk going to war there. — T Clark

Taiwan produces around 90% of the world's most advanced ICs. It is very much in the iinterest of the US and other countries that such manufacturing capabilities are not taken over by China.

Nature is natural, machines are artificial, and never the twain shall meet — ENOAH

That sounds like dogma. Do you have any reasonining to back it up?

I've never seen my own brain. How do I know that I have one? Maybe there is a machine inside my skull, that has mechanical gears and Steampunk technology in general. — Arcane Sandwich

Well, there are substances you might ingest, which would have results on your thinking which don't seem too consistent with what one would expect the substance to have on a steam and gear mechanism.

I.e. you could conduct experiments.

A.I. changes itself according to principles that we program into it, in relation to norms that belong to us. — Joshs

The same will be true of this new system as the old. It will never be or do anything that exceeds the conceptual limitations of its design. — Joshs

This seems rather naive when it comes to neural net based AI.

Consider this excerpt from a recent Science Daily article:

What is more, the AI behind the new system has produced strange new designs featuring unusual patterns of circuitry. Kaushik Sengupta, the lead researcher, said the designs were unintuitive and unlikely to be developed by a human mind. But they frequently offer marked improvements over even the best standard chips.

"We are coming up with structures that are complex and looks random shaped and when connected with circuits, they create previously unachievable performance. Humans cannot really understand them, but they can work better," said Sengupta, a professor of electrical and computer engineering and co-director of NextG, Princeton's industry partnership program to develop next-generation communications.

Measure the length of a sundial's shadow at noon at different latitudes.

Shoot a laser horizontally across a large still lake to measure the curvature of the lake's surface. There was a documentary on flat earthers (I think on Netflix) where this was done, but the flat earthers dismissed the results falsifying a flat earth.

At what measure of mass density, does the recording device effect that which the device is suppose to record, synonymous with the quantum “observer problem”? — Mww

:up:

At what point in 'sensoring up' a human, have you created something that is no longer a human.

Do you think it's possible to record the individual human experience?

By that I mean, what each of us go through every second of our lives? The inputs to our senses, the thoughts that pass by, the emotions we feel? — Ayush Jain

If you had stopped at "inputs to our senses" you'd have something somewhat technologically feasible (if not very practical). However, with thoughts and emotions you would be talking about highly invasive measurement of a huge amount of activity occurring in people's brains. It's not at all technologically feasible to measure and record the data that would be needed.

C.S. Lewis - The Discarded Image
— Count Timothy von Icarus

I went back and read this section in its entirety. It is an excellent summary of the difference between intellection and ratiocination, as well as the decline of intellection since the modern period. :up: — Leontiskos

It is interesting to consider the relevance of Kahneman's distinction between fast and slow thinking to Lewis' discussion of intellectus and ratiocination.

Kahneman's work suggests that Lewis' claim that intellection (fast thinking) is higher, is rather questionable.

FYI, for those interested in such...

Gene expression in the human brain: cell types become more specialized, not just more numerous

There is an element of human xenophobia that plays a part in war but which doesn't seem particularly relevant to inner conflict.

This is the usual "philosophy is a bridge between the scientific and manifest images" jazz. Though I'm framing metaphysics as an intimate part of the bridge. It's a form of "conceptual engineering", of propagating changes and insights from one to another. — fdrake

:up:

Scientists aren’t paid to think about their work, they’re paid to produce results. — Joshs

:roll:

:up:

They had me at dinobeavers.

I think we can recognize precision and explicitness as thresholds that are negotiated, without idealizing them into unreachable and thus useless perfection. We say enough to be understood, counting on the audience to fill in as much as they need to to get it, and even that can be negotiated.

But that just kills off an unrealistic picture of how conversation works. Even if your speech doesn't have to carry the burden of truth entirely on its own, it has to do its part.

I keep finding myself thinking that the great value of saying something true to someone else is helping them see it ― like when you point out to someone that a photo of the faculty of your department has no women in it. And it's not just a matter of your words being understood and even credited; if I lie to you convincingly, my words hide the world from you, obstruct and undermine your relationship with it, divert your attention into a shadowy fantasy land. But when I tell you the truth, and you see it, my words fall away. — Srap Tasmaner

:100: :up:

The point I am making is not that ink and paper aren't essential to the physical nature of the book but that semantic content exists on a different level from its physical form. Words may be encoded as sounds or written letters in various languages, yet the same information can be encoded in entirely different symbolic systems—whether in different languages, Braille, or even Morse code—and still retain its meaning. This demonstrates that semantic content is independent of the specific physical medium in which it is expressed. — Wayfarer

The thing is, your 'point' is a mystification of what is a relatively simple and clear physical picture.

There is no need for a 'different level' for semantic content to exist on. Semantic content is attributed to linguistic media (letters, Braille, Morse code, etc.) by neural nets which have been trained to attribute semantic content to such media. Such attribution of semantic content to linguistic media is a function of the physical state of systems capable of doing such decoding.

A book 'contains meaning' only insofar as it is read and understood by a subject capable of interpreting its content. Furthermore, different readers may interpret the same information in diverse ways, highlighting the subjective and contextual nature of meaning-making. The meaning is not an inherent property of the physical text itself but arises through the interaction between the symbolic representation and the mind of the reader. — Wayfarer

Right. This is completely consistent with the straightforward physical picture outlined above.

So language has a physical aspect, but it can't be accounted for by physical principles alone. — Wayfarer

And yet you rely on LLMs. :roll:

If you actually understand that language has an aspect that can't be accounted for by physical principles, I'd expect you could come up with a way of falsifying any physicalist account of language. That would be a serious philosophical achievement. Go for it!

I'm not familiar with Alicia Juarrero’s perspective, but what I've gathered from looking at the Amazon page for her book sounds generally compatible with Tse's thinking. FWIW, I searched my Kindle copy of Tse's book for any citation of Alicia Juarrero, and didn't find any. I'm planning to borrow a copy of Juarrero’s book, so perhaps I can let you know more later.

The meaning arises as a brain (containing neural networks trained to recognize the written language the book is written in) detects patterns in the writing which are associated by that brain with the meaning that arises.
— wonderer1

What about this causal relationship is physical? — Wayfarer

Everything. Words are patterns of physical vibrations propagating through the air, or physical text. Neural networks in your brain which recognize words and their semantic associations are physical. The semantic elements in your stream of thought are physically detectable.

How is it explainable in physical or molecular terms? — Wayfarer

There is an enormous amount of science to study to reach a complete account at the molecular level. Can you be more specific about what it is that you don't understand?

How do physical interactions cause or give rise to semiotic processes? — Wayfarer

Reading Peter Tse's Criterial Causation might provide a clue. Before reading Tse, I used an analogy of locks and keys, where in the scenario of reading written language, letters, words, phrases, etc. play the roles of keys, and neural nets trained in written language recognition play the role of locks. Of course I don't expect that to make any sense to anyone so unwilling to consider physicalism charitably as yourself.

Along with the majority of philosophers of mind.

In philosophy, to equate mental with physical is a category error. — Gnomon

Brandolini's law:

Brandolini's law, also known as the bullshit asymmetry principle, is an internet adage coined in 2013 by Alberto Brandolini, an Italian programmer, that emphasizes the effort of debunking misinformation, in comparison to the relative ease of creating it in the first place. The law states:

The amount of energy needed to refute bullshit is an order of magnitude bigger than that needed to produce it.

Philpapers Survey

The meaning arises as a brain (containing neural networks trained to recognize the written language the book is written in) detects patterns in the writing which are associated by that brain with the meaning that arises.
— wonderer1

'Arises' from what, exactly? What is the nature of the causal relationship? — Wayfarer

Arises from interactions within the brain which contains the neural networks trained to process written language, in response to the outputs of those neural networks signaling recognition of linguistic elements in the writing.

[Note 'arises' is the word you chose, and I ran with. Not a word I injected into the discussion.]

The nature of the causal relationship is physical.

If meaning arises purely from physical causation, as described by physical and chemical laws, how to account for the gap between these deterministic processes and the open-ended, adaptive nature of life? Even rudimentary organisms exhibit an agency and intentionality absent in inorganic matter—the ability to heal, reproduce, evolve, and maintain homeostasis. From the moment life begins, biological systems exhibit a kind of semiotic agency that transcends the deterministic causal nexus of physics and chemistry. Life doesn't defy physical laws, but requires principles that can't be reduced to that level of explanation. Recognition of this is one of the drivers behind the emergence of biosemiotics, and of the connection between information and biology, none of which is strictly physicalist, although it falls within the ambit of an evolving naturalism. That's the sense in which biology is evolving beyond physicalism, as physics did with the advent of quantum mechanics. And all the same questions apply to the relatonship of neurobiology and semantics.

refs: From Physical Causes to Organisms of Meaning, Steve Talbott

What is Information?, Marcello Barbieri — Wayfarer

That's an impressive load of red herrings you have there, but how about sticking to this
original question?

All of Greene's books, of which I've read The Fabric of the Universe, consist of paper and ink. Is that all they are? How does the meaning they convey arise from the combination of ink and paper? — Wayfarer

Or did you not actually want people to give serious consideration to the matter?

All of Greene's books, of which I've read The Fabric of the Universe, consist of paper and ink. Is that all they are? How does the meaning they convey arise from the combination of ink and paper? — Wayfarer

The meaning arises as a brain (containing neural networks trained to recognize the written language the book is written in) detects patterns in the writing which are associated by that brain with the meaning that arises.

What alternative explanation would you propose? Or even better, how could you falsify my explanation?

Admittedly, my theory has changed slightly as it is now simpler than the original.

But the original evolutionary story involves random mental state generation and a mind disposed to remember sequences of mental states that seem closely to resemble one another. It experiences A. — Clearbury

The biological theory of evolution is based on all sort of empirical evidence. Is empirical evidence for your theory of evolution even possible in principal?

What's being posited is a mind that is in a mental state - so, whatever total mental mental state you are in now (including all experiential states), just assume the mind is in it. — Clearbury

So whence comes a theory of evolution worth considering in light of this posit?